Datasets:

tomekkorbak

/

pii-pile-chunk3-1800000-1850000

like 0

[ "In the year 2010 I realized a dream. ", "Not any of those unattainable dreams we all have, but a dream nonetheless: I bought my first sports car. ", "Constrained by a student budget and being too sensible, my choice of a car wasn’t any vintage Alfa or Porsche, but a bog-standard early Mazda MX-5. ", "Nevertheless, I felt like a schoolboy on Christmas— which I kind of was, as I was an exchange student at Universität Stuttgart and I chose to buy the car only a few days before the first snow of the season.", "\n\nSlowly the spring arrived. ", "The Mazda started to do my daily driving duties and occasional drives in the local countryside. ", "Yet something more meaningful was needed to bond with the car: a proper road trip featuring many of the greatest roads built by mankind. ", "Lots of Googling, watching Youtube and countless issues of Evo magazines later the final route was chosen. ", "On the 20th of May 2011 my co-driver and I filled up the small trunk of the Mazda with our camera bags only to realize we had to make a decision between having some clothes or a spare wheel. ", "When was the last time you had a puncture? ", "Exactly.", "\n\nAfter a two days’ drive from Stuttgart with a quick overnight stop at Geneva we arrived at Grenoble in France. ", "There lies the northern end of an epic path flowing around the hills all the way from the Mediterranean. ", "It’s officially called the N85, but like most great roads it also has a name: the Route Napoleon. ", "Mr. Bonaparte chose this route on his way back from Elba, but in 1815 he probably couldn’t enjoy the fast nature of this road as much as journalists from Evo (who’ve chosen this particular road to host their many group tests in the past) do. ", "Indeed, with plenty of straights and fast third gear corners it’s probably more suited to the exclusive end of sports cars than an early MX-5, but even with 115 bhp it was damn good fun.", "\n\nIf you’ve got time, there is a really worthwhile stop along the N85. ", "Called Gorges du Verdon, it is a combination of bright blue water, green plants and hard rocks with a tiny road twisting around on both sides of the canyon. ", "It could be from the imagination of a video game designer, but it’s very real. ", "In fact, so real that there is no armco blocking your way to the bottom of the canyon should you make a serious mistake. ", "The road is also full of imperfections, so unless you’re Francois Delecour the road is better taken as a scenic route instead of as driving nirvana.", "\n\nAfter sunrise in a small town called La Palud-sur-Verdon the MX-5’s tiny fuel tank was once again filled before heading back to the last leg on the Route Napoleon. ", "The road itself goes all the way down to the coast in Cannes, but unless you’re absolutely sure you enjoy traffic reminiscent of bumper car rides in amusement parks I’d suggest taking a shortcut. ", "We departed the unforgettable N85 in Grasse and took a random combination of roads to Nice, where the traffic for once had calmed down.", "\n\nOur destination for the night was just a short drive on the Autoroute A8 away, but the day wasn’t over. ", "A sacred piece of road lies just a few hours drive north of Nice. ", "It has some of the most famous hairpin curves in the world and it’s a regular special stage in Monte Carlo Rally: Col de Turini. ", "From a small town of La Bollène-Vésubie to Moulinet it goes like a giant strand of spaghetti thrown from an airplane against the cliffs. ", "I would love to tell you how the road flows through all the bends, but navigating in the mountains and finding the right road without access to the Internet is not an easy task, so I must admit we found our way to Col de Turini at its midway point where road 2566 joins the special stage. ", "It’s a bit of a gutted feeling to realize you were so close yet still missed a legendary road, but to be perfectly honest, most of the mountain passes in the area are all great. ", "Some of the great ones just have a bit more history than the others.", "\n\nA fitting end to a Rally Monte Carlo-themed day was the descent from the mountains to Monaco itself. ", "An obligatory lap around the Grand Prix track got off to an awesome start when a Bugatti Veyron driver waved me to go before him into the famous tunnel. ", "Moments later the W16 engine took a deep breath and the French supercar gained few hundred meters in the blink of an eye. ", "Not my kind of car, but a truly unforgettable experience. ", "After the tunnel the lap was interrupted by construction work around the harbor area. ", "Unsurprising, as the Formula 1 race was only a few days away.", "\n\nThe next day was an exception to our rule of avoiding highways. ", "Our destination was Modena, and it’s a whole 450 km away. ", "For the first half of the day we stayed on the Italian Autostrada all the way to a town called La Spezia, which conveniently lies on the other side of the Apennine mountains than our destination. ", "Time was running short, so we picked the fastest way through the mountains. ", "As the shape of the Apennines is much softer and milder than the Alps, we got to enjoy nice flowing curves and surprisingly decent average speed. ", "Our great plan was to aim for Maranello, as the home of Ferrari lies on a plain literally a few hundred meters away from the hills. ", "Sadly, once again driving took concentration away from navigation and we missed the town by almost twenty kilometers. ", "On the positive side, we avoided all of the Rosso Corsa-colored, Cavallino Rampante-logoed merchandise so inescapable nowadays.", "\n\nThe problem with staying only two days in Modena is that you need to make some really tough decisions. ", "Which supercar factory do you want to visit? ", "An old friend had kindly booked us a visit to Lamborghini, where a few hours were spent around the Aventador assembly line and in the Lamborghini Museum. ", "After admiring the assembly of the all new 6.5 litre V12 it was time to fire up the trusty straight-four under the aluminium bonnet of the Mazda and continue. ", "We had booked a hostel on the northern part of Lake Garda and time was running short.", "\n\nTime was running short on our trip too. ", "As the morning cappuccinos were finished in the sunny town of Riva del Garda, we knew this was going to be the last day of mountain passes. ", "It could, possibly, also be the greatest. ", "The roads on the north of Lake Garda gave no reasons to complain, but our minds were already set on the highlight of the whole trip: SS38, also known as the Stelvio Pass. ", "A stunning and demanding road that takes you through countless hairpin curves up to 2757m (9,050 ft) above the sea level. ", "Even during the hottest summer, the sight of snow banks is common, and sadly this year was no exception. ", "Fifty kilometers before the pass a sign on the side of the road declared current conditions: Closed. ", "Road number SS38 lay under two meters of snow, closed for at least another month.", "\n\nAs accommodation had already been booked in Davos, Switzerland, it was an easy decision to take a detour around the Stelvio Pass via another road number, SS41. ", "From there we took a long tunnel to the tax haven Livigno, where the route via Stelvio Pass would also have taken us. ", "The road we ended up driving was much quicker, but not as memorable as the endless row of hairpins often seen on pictures from Stelvio. ", "Even after the disappointment we still had two roads with the promising prefix “Pass” on their name: Bernina Pass right after leaving Livigno, and Flüela Pass that took us to our destination, Davos.", "\n\nEven though the Berina Pass only rises to 2328m (7,638ft) it was getting rather chilly driving with the top down. ", "On the positive side, the unobstructed view out of a roadster is something you’re really thankful for driving locations like these. ", "Good thing the scenery was breathtaking given that as a driving road, the western side of the Bernina Pass wasn’t really a competitor to the Stelvio.", "\n\nDavos was getting closer, and road number 27 that we took after Bernina Pass to Süs was an ordinary country road full of traffic. ", "The general feeling was one of contentment, even though the main draw, Stelvio, didn’t work out as we had planned. ", "When turning to Flüela Pass we were already talking about memorable events we had during the trip, just as you might do on a return flight after a holiday. ", "I even had second thoughts about overtaking a slowly-driven Škoda that was blocking the road.", "\n\nGood thing I decided to leave the Škoda behind, as the warm weather, slowly setting sun and the nature of the Flüela Pass created a perfect canvas of a surprise. ", "I had actually never given any thoughts to why Top Gear’s “Greatest Driving Road in the World” episode said “From Davos to Stelvio” because I always assumed the fun parts lie on the Italian side of the border. ", "Well, this time the Swiss really didn’t disappoint. ", "After a brief phase through an area full of pine trees we ascended steadily higher and finally left all of the woods behind.", "\n\nJust as the sunlight turned a beautiful golden colour we arrived to a deserted valley in the middle of nowhere. ", "Between two massive slopes there were only two twisty lines through the rocks and green fields: a small river and a black curvy road. ", "Great visibility, impeccably smooth tarmac, and an exciting combination of fast curves created a different rhythm to most other passes around the area. ", "As a driving road it’s more suited for an Italian supercar than an Impreza or an Evo, but it also worked superbly with an MX-5. ", "Quick curves meant we could keep up the speed, only slowing down to admire the views and take some pictures.", "\n\nMaybe the Flüela Pass felt so great because we didn’t have any expectations and great weather helped to create a perfect atmosphere for an unforgettable driving experience. ", "After the Route Napoleon, roads around Gorges du Verdon, the mighty Col de Turini and countless great roads on the Apennine mountains it was the Flüela that stood out. ", "When we finally arrived to the horribly expensive and deserted city of Davos we had totally forgotten the disappointment of missing out on the Stelvio Pass.", "\n\nNext morning, as our minds were still processing all the great moments we had, we woke up to the first rainy day of the trip and drove back to Stuttgart. ", "The little Mazda had taken an awful amount of punishment through very demanding roads and all it ever asked for was more fuel. ", "What an unforgettable way to get to know your own car.", "\n\nImages © Lauri Ahtiainen\n\nSee more of Lauri’s photography here.", "\n\n\n\n\n\nPosted in Grab bag, Roads, Still images\n\nTags: col de turini, flüela pass, mazda, miata, mx-5, mx5, route napoléon, stelvio pass" ]

{ "pile_set_name": "OpenWebText2" }

[ 0, 0, 0.02702702702702703, 0, 0, 0.010416666666666666, 0, 0.009345794392523364, 0.005235602094240838, 0, 0, 0.008849557522123894, 0, 0.02040816326530612, 0.012396694214876033, 0, 0.014084507042253521, 0.006369426751592357, 0, 0, 0, 0.012048192771084338, 0, 0.007407407407407408, 0, 0, 0.007751937984496124, 0, 0.0034602076124567475, 0, 0, 0.009708737864077669, 0.006535947712418301, 0.00819672131147541, 0, 0, 0, 0, 0.017241379310344827, 0, 0, 0, 0.007575757575757576, 0, 0.007874015748031496, 0.009523809523809525, 0, 0.012987012987012988, 0.006289308176100629, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.006172839506172839, 0.00847457627118644, 0.007352941176470588, 0.015151515151515152, 0, 0, 0.006711409395973154, 0.015151515151515152, 0.008695652173913044, 0.00641025641025641, 0.010752688172043012, 0.006097560975609756, 0, 0, 0, 0, 0, 0, 0.0078125, 0.009259259259259259, 0, 0.017857142857142856, 0, 0, 0.007874015748031496, 0, 0.03076923076923077, 0.022388059701492536 ]

[ "/*\n * Copyright (c) 2012 The WebRTC project authors. ", "All Rights Reserved.", "\n *\n * Use of this source code is governed by a BSD-style license\n * that can be found in the LICENSE file in the root of the source\n * tree. ", "An additional intellectual property rights grant can be found\n * in the file PATENTS. ", " All contributing project authors may\n * be found in the AUTHORS file in the root of the source tree.", "\n */\n\n#include \"webrtc/modules/remote_bitrate_estimator/overuse_detector.h\"\n\n#include <math.h>\n#include <stdlib.h>\n\n#include <algorithm>\n#include <sstream>\n#include <string>\n\n#include \"webrtc/modules/remote_bitrate_estimator/include/bwe_defines.h\"\n#include \"webrtc/modules/remote_bitrate_estimator/test/bwe_test_logging.h\"\n#include \"webrtc/modules/rtp_rtcp/source/rtp_utility.h\"\n#include \"webrtc/rtc_base/checks.h\"\n#include \"webrtc/rtc_base/logging.h\"\n#include \"webrtc/rtc_base/safe_minmax.h\"\n#include \"webrtc/system_wrappers/include/field_trial.h\"\n\nnamespace webrtc {\n\nconst char kAdaptiveThresholdExperiment[] = \"WebRTC-AdaptiveBweThreshold\";\nconst char kEnabledPrefix[] = \"Enabled\";\nconst size_t kEnabledPrefixLength = sizeof(kEnabledPrefix) - 1;\nconst char kDisabledPrefix[] = \"Disabled\";\nconst size_t kDisabledPrefixLength = sizeof(kDisabledPrefix) - 1;\n\nconst double kMaxAdaptOffsetMs = 15.0;\nconst double kOverUsingTimeThreshold = 10;\nconst int kMinNumDeltas = 60;\n\nbool AdaptiveThresholdExperimentIsDisabled() {\n std::string experiment_string =\n webrtc::field_trial::FindFullName(kAdaptiveThresholdExperiment);\n const size_t kMinExperimentLength = kDisabledPrefixLength;\n if (experiment_string.length() < kMinExperimentLength)\n return false;\n return experiment_string.substr(0, kDisabledPrefixLength) == kDisabledPrefix;\n}\n\n// Gets thresholds from the experiment name following the format\n// \"WebRTC-AdaptiveBweThreshold/Enabled-0.5,0.002/\".", "\nbool ReadExperimentConstants(double* k_up, double* k_down) {\n std::string experiment_string =\n webrtc::field_trial::FindFullName(kAdaptiveThresholdExperiment);\n const size_t kMinExperimentLength = kEnabledPrefixLength + 3;\n if (experiment_string.length() < kMinExperimentLength ||\n experiment_string.substr(0, kEnabledPrefixLength) !", "= kEnabledPrefix)\n return false;\n return sscanf(experiment_string.substr(kEnabledPrefixLength + 1).c_str(),\n \"%lf,%lf\", k_up, k_down) == 2;\n}\n\nOveruseDetector::OveruseDetector()\n // Experiment is on by default, but can be disabled with finch by setting\n // the field trial string to \"WebRTC-AdaptiveBweThreshold/Disabled/\".", "\n : in_experiment_(!AdaptiveThresholdExperimentIsDisabled()),\n k_up_(0.0087),\n k_down_(0.039),\n overusing_time_threshold_(100),\n threshold_(12.5),\n last_update_ms_(-1),\n prev_offset_(0.0),\n time_over_using_(-1),\n overuse_counter_(0),\n hypothesis_(BandwidthUsage::kBwNormal) {\n if (!", "AdaptiveThresholdExperimentIsDisabled())\n InitializeExperiment();\n}\n\nOveruseDetector::~OveruseDetector() {}\n\nBandwidthUsage OveruseDetector::State() const {\n return hypothesis_;\n}\n\nBandwidthUsage OveruseDetector::Detect(double offset,\n double ts_delta,\n int num_of_deltas,\n int64_t now_ms) {\n if (num_of_deltas < 2) {\n return BandwidthUsage::kBwNormal;\n }\n const double T = std::min(num_of_deltas, kMinNumDeltas) * offset;\n BWE_TEST_LOGGING_PLOT(1, \"T\", now_ms, T);\n BWE_TEST_LOGGING_PLOT(1, \"threshold\", now_ms, threshold_);\n if (T > threshold_) {\n if (time_over_using_ == -1) {\n // Initialize the timer. ", "Assume that we've been\n // over-using half of the time since the previous\n // sample.", "\n time_over_using_ = ts_delta / 2;\n } else {\n // Increment timer\n time_over_using_ += ts_delta;\n }\n overuse_counter_++;\n if (time_over_using_ > overusing_time_threshold_ && overuse_counter_ > 1) {\n if (offset >= prev_offset_) {\n time_over_using_ = 0;\n overuse_counter_ = 0;\n hypothesis_ = BandwidthUsage::kBwOverusing;\n }\n }\n } else if (T < -threshold_) {\n time_over_using_ = -1;\n overuse_counter_ = 0;\n hypothesis_ = BandwidthUsage::kBwUnderusing;\n } else {\n time_over_using_ = -1;\n overuse_counter_ = 0;\n hypothesis_ = BandwidthUsage::kBwNormal;\n }\n prev_offset_ = offset;\n\n UpdateThreshold(T, now_ms);\n\n return hypothesis_;\n}\n\nvoid OveruseDetector::UpdateThreshold(double modified_offset, int64_t now_ms) {\n if (!", "in_experiment_)\n return;\n\n if (last_update_ms_ == -1)\n last_update_ms_ = now_ms;\n\n if (fabs(modified_offset) > threshold_ + kMaxAdaptOffsetMs) {\n // Avoid adapting the threshold to big latency spikes, caused e.g.,\n // by a sudden capacity drop.", "\n last_update_ms_ = now_ms;\n return;\n }\n\n const double k = fabs(modified_offset) < threshold_ ? ", "k_down_ : k_up_;\n const int64_t kMaxTimeDeltaMs = 100;\n int64_t time_delta_ms = std::min(now_ms - last_update_ms_, kMaxTimeDeltaMs);\n threshold_ +=\n k * (fabs(modified_offset) - threshold_) * time_delta_ms;\n threshold_ = rtc::SafeClamp(threshold_, 6.f, 600.f);\n last_update_ms_ = now_ms;\n}\n\nvoid OveruseDetector::InitializeExperiment() {\n RTC_DCHECK(in_experiment_);\n double k_up = 0.0;\n double k_down = 0.0;\n overusing_time_threshold_ = kOverUsingTimeThreshold;\n if (ReadExperimentConstants(&k_up, &k_down)) {\n k_up_ = k_up;\n k_down_ = k_down;\n }\n}\n} // namespace webrtc\n" ]

{ "pile_set_name": "Github" }

[ 0, 0, 0.013793103448275862, 0, 0, 0.002053388090349076, 0.005763688760806916, 0.0028735632183908046, 0, 0.005398110661268556, 0, 0.0024937655860349127, 0.003875968992248062, 0, 0.011764705882352941 ]

[ "Q:\n\nHow to write conditional order by clause criteria in hibernate?", "\n\nI need to convert below query in hibernate criteria.", "\nAny one have idea how to write this order by clause in hibernate?", "\nselect * from StudentMasterTable where SchoolId = 90 and SchoolName IS NOT NULL ORDER BY IIF(StudentType = 'O',StudentName,SchoolName)\n\nA:\n\nfinally i found solution for this.", "\ni created a custom class to override the criteria formula.", "\npublic class customOrderBy extends Order {\n private String sqlFormula;\n\n protected OrderBySqlFormula(String sqlFormula) {\n super(sqlFormula, true);\n this.sqlFormula = sqlFormula;\n }\n\n public String toString() {\n return sqlFormula;\n }\n\n public String toSqlString(Criteria criteria, CriteriaQuery criteriaQuery) throws HibernateException {\n return sqlFormula;\n }\n\n public static Order customOrder(String sqlFormula) {\n return new OrderBySqlFormula(sqlFormula);\n }\n}\n\nAnd i called this as below\ncriteria.addOrder(customOrderBy.customOrder(\"IIF(StudentType = 'O',StudentName,SchoolName)\"));\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0, 0, 0, 0, 0, 0.0015360983102918587 ]

[ "The use of induced sputum in the assessment of pulmonary involvement in Crohn's disease.", "\nOur aim was to evaluate lung involvement in Crohn's disease (CRD) patients by induced sputum (IS). ", "Extraintestinal manifestations are frequent in CRD, but lung involvement is rare. ", "Induced sputum is a reliable noninvasive method of investigating the pathogenesis, pathophysiology, and treatment of lung disease. ", "Twenty-four CRD patients and nine control subjects (all nonsmokers) without respiratory symptoms were tested. ", "Sputum was induced by 20' inhalation of 3.5% saline using ultrasonic nebulizer. ", "Samples were studied by differential counts of 200 cells on cytopreps stained by Giemsa. ", "T-lymphocyte subset analyses were done by FACS using three monoclonal antibodies: CD3 = total T cells, CD4 = T helper cells, and CD8 = T suppressor-cytotoxic cells. ", "CD4/CD8 >2.5 was considered abnormal. ", "Four patients did not produce sputum. ", "Of the remaining 20 patients, seven had active CRD and 13 were in remission. ", "They were divided into two groups: Group A patients had abnormal CD4/CD8 ratio of 6.7 +/- 2.5; Group B (seven patients) had normal CD4/CD8 ratio of 1.7 +/- 0.52 (p = 0.0001). ", "The differential counts of IS samples were similar in both groups, but lymphocyte count was significantly higher in CRD patients than in the control group (13.2 +/- 11.2 vs 4.8 +/- 3.6, p = 0.036). ", "There were no differences in spirometry, treatment, extent, or activity of CRD. ", "Using a simple noninvasive method, we found that among CRD patients without respiratory symptoms there was a high (65%) incidence of lung involvement." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0.011363636363636364, 0.02, 0.012195121951219513, 0, 0, 0, 0.011235955056179775, 0, 0, 0, 0.012987012987012988, 0.017142857142857144, 0.005050505050505051, 0.0125, 0.006666666666666667 ]

[ "That's a Girls Name\n\nThat's a Girls Name is the fourth EP by American rapper DRAM, released on July 18, 2018, via Empire and Atlantic Records. ", "Produced by Josh Abraham and Oligee, it was a surprise release, and consists of three songs that were characterized by Stereogum as \"deeply goofy and fun takes on '80s soul\".", "\n\nMusic\nThe EP's opening track, \"Best Hugs\", is a \"shimmery adult-contempo synth-glide\" that is an \"ode to another person's girlfriend\", while second song \"WWYD?\" ", "contains a \"cocktail bar\" feel with \"canned\" horn accompaniment that features \"amateurish rapping\" from DRAM. ", "The third and final track, \"Sundress\", features a prominent, funky bassline and was considered to be influenced by Prince. ", "The release was called \"inspired by hugs, carwashes, construction workers and Twitter\" with witty lyrics.", "\n\nCritical reception\nStereogum and Paste found the three tracks \"very, very good\", with Paste further calling the EP \"profoundly goofy, in true DRAM style, and it plays on '80s soul vibes\". ", "Clash judged it to be an EP of \"irresistible Summer funk\", while Consequence of Sound stated that its tracks have DRAM \"compellingly embracing elements of soul and hip-hop\". ", "Spin called the EP a \"typically bubbly and optimistic group of poppy hip-hop songs\".", "\n\nTrack listing\n\nReferences\n\nCategory:2018 EPs\nCategory:DRAM (musician) albums" ]

{ "pile_set_name": "Wikipedia (en)" }

[ 0.013986013986013986, 0.017241379310344827, 0.006134969325153374, 0, 0.008130081300813009, 0, 0.015789473684210527, 0.005747126436781609, 0, 0 ]

[ "Q:\n\nPowershell: hashtable content not displayed within a function\n\nI have a function to which I pass a hashtable. ", "Within the function I want to 1) Display text on-screen via Write-Host; 2) display the contents of the hashtable one time -- to provide the usual two-column \"Name\" / \"Value\" hashtable display. ", "3) Have the function return $true or $false.", "\nMyFunction $MyHashTable\n\nWithin the function:\nparam (\n [hashtable]$TheHashTable\n)\n# Sundry things here and then:\nwrite-host \"Some information to display on-screen`n\"\n# and then:\n$TheHashTable\n\nThe expected result of the latter is something like:\nSome information to display on-screen\n\nName Value\n---- -----\na b\nc d\n\nAnd eventually:\nreturn $true # If what I'm doing worked; otherwise, $false\n\nIf I call the function as shown above, I see text displayed via Write-Host on-screen, plus the two-column display of the hashtable's contents -- and the text True or False on-screen, depending on what the function returns.", "\nIf I call it this way:\n$myResult = MyFunction $MyHashTable\n\n... I capture the return value of the function in $myResult -- but the display of the hash table's content is suppressed. ", "It is also suppressed if I do this:\nif ( (MyFunction $MyHashTable) -eq $true ) {\n # do something\n} else {\n # do something different\n}\n\nIs there a way to\n\nEnsure the display of hashtable content, no matter how the function is called;\nIn any case, suppress the on-screen display of True and False when the Return statement is executed?", "\n\nA:\n\nAny output generated by your function will be sent down the pipeline. ", "This is exactly what happens when you write:\n$TheHashTable\n\nIf you want to write this value to the screen instead of the pipeline you should also use Write-Host like you do earlier in the example like so:\nWrite-Host $TheHastTable\n\nHowever using the code above you will probably get something like the following output:\nPS>$table = @{ \"test\"=\"fred\";\"barney\"=\"wilma\"}\nPS> write-host $table\nSystem.", "Collections.", "DictionaryEntry System.", "Collections.", "DictionaryEntry\n\nApparently Write-Host does not apply the formatting you expect, this can be fixed by using Out-String like so:\nPS> $table | Out-String | Write-Host\n\nresulting in:\nName Value\n---- -----\nbarney wilma\ntest fred\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0, 0.0051813471502590676, 0, 0.001564945226917058, 0, 0, 0, 0.002531645569620253, 0, 0.043478260869565216, 0, 0.006097560975609756 ]

[ "In ‘The Rise and Decline of the Medici Bank’, Raymond de Roover wrote that the Medici dynasty created the first instance of “modern capitalism based on private ownership”.", "\n\nLike any modern bank, the Medicis held deposits and profited from loans. ", "Most of theses deposits and loans came from the aristocracy. ", "They were used to fund the Prince’s birthday or military campaigns for example.", "\n\nThe Medici Bank was born in 1397, and since then, the banking system didn’t change much.", "\n\nAccording to McKinsey, 2.5 billions people around the world don’t use any type of financial services to make payments, save, borrow money, or even transfer funds. ", "2.5 billions people have no credit score. ", "They can’t buy a home, start a company or send their kids to college using modern financial products.", "\n\nYes, 2.5 billion people don’t have a saving account or a credit card.", "\n\nOf the 1.2 billion adults who use formal financial services in Africa, Asia, and the Middle East, at least two-thirds, a little more than 800 million, live on less than $5 per day.", "\n\nWorld Bank Group President Jim Yong Kim said that access to financial services is “a bridge out of poverty.”", "\n\nIn this piece, I will write about what has be done so far, what went wrong and how can Bitcoin change everything.", "\n\nBanking the unbanked !", "= financial inclusion.", "\n\nUntil today, financial services institutions, governments and even NGOs are focused on fixing one problem : How to make the poor open a bank account.", "\n\nBy focusing solely on banking the unbanked, we tend to forget about their real needs and how they tend to spend their money.", "\n\nPeople in the developing world tend to spend their money differently. ", "In 2007, Banerjee and Dufloo wrote a report on The Economic Lives of the Poor* where they describe the spending pattern of the poor and the choices they face everyday.", "\n\nLet’s look at the spending of a typical household in Udaipur,India where family members live on $1 per day, per head :\n\n56% — 78% of income on food\n\n5% on alcohol and tobacco\n\nmedian household spends 10% of annual budged spent on festivals (e.g. religious festivities, weddings,…)\n\n11% own a radio, no one owns a TV\n\n99% own some plot of land, but most is dry scrubland which can’t be cultivated for most of the year (small parcels)\n\nless than 1% own a fan, sewing machine or a bullock cart or any motorized vehicle\n\n56 to 78% of the income is spent on calories.", "\n\nY = C\n\nAlmost all their income is spent on food. ", "They don’t save. ", "What is interesting to see in Banerjee and Dufloo’s report is that the reason the poor don’t save is not due to lack of access to credit and insurance or labour market rigidities.", "\n\n“Saving at home is hard: the money may be stolen or simply grabbed by your spouse or your son. ", "Alternatively, you are constantly resisting temptation to spend. ", "The poor seem quite aware of their vulnerability to temptation.”", "\n\nS=0\n\nIn a paper on “Money Management, Financial Inclusion and Banking the Unbanked”, Elisabeth Rhyne wrote that to deal with their low and uneven income streams, poor people engage in many kinds of financial arrangements. “", "They keep savings at home; join savings clubs and savings-and-loan clubs; transact with family, friends and neighbors, and employers; and, where doing so is feasible and attractive, sign on with formal licensed providers.”", "\n\nHow do people keep track of all theses transactions ?", "\n\nIn their heads, small notes or a passbook. “", "They may operate with what behavioral economists call heuristics or rules of thumb — simple strategies such as keeping business and personal money in different pockets, or saving by setting aside a handful of rice every day.”", "\n\nEven after M-Pesa, Cash is still King\n\nA lot of services and institutions have tried to including the 2.5 billion unbanked in a formal financial system, but what has this resulted in?", "\n\nLaunched by Safaricom in 2007, M-Pesa has allowed people to deposit money in accounts and transfer money person-to-person using basic SMS text messages.", "\n\nM-Pesa’s success was huge. ", "In 2014, it reported 19 million registered users, 12.7m of whom are active and over 6 million daily transactions.", "\n\nDespite the huge success of M-Pesa, people still prefer cash. ", "Over 90% of transactions still happen in cash and 2/3 of all transactions using M-Pesa are conducted over the counter, meaning they pay an agent cash in order to make a transaction.", "\n\nSimilarly, according to a survey conduced by the FITS , a lot of Pakistanis aren’t opening an account because they didn’t see the need to open one. ", "Most of their transactions are also conducted over-the-counter.", "\n\nThe poor only make one transaction before cashing out.", "\n\nAccording to the CGAP, “M-Pesa transactions are one loop, meaning that there is only one electronic transaction before money is cashed out.”", "\n\nThe promise of Bitcoin.", "\n\n“The combination of ubiquitous Internet-connected mobile devices and digital currency presents a tremendous opportunity to radically expand access to financial services on a worldwide basis” — Jeremy Allaire, chief executive officer of Circle Internet Financial\n\nBitcoin can allow us to create new type of financial products targeted to those who live in poverty.", "\n\nEvery Penny Counts.", "\n\n“Bitcoin is the first Internetwide payment system where transactions either happen with no fees or very low fees (down to fractions of pennies). ", "Existing payment systems charge fees of about 2 to 3 percent — and that’s in the developed world. ", "In lots of other places, there either are no modern payment systems or the rates are significantly higher.“ — ", "Marc Andreessen.", "\n\nWith Bitcoin and a cheap mobile phone, they can pay anyone in the world any amount of value of bitcoin, without the need of an authorization and almost without fees.", "\n\nWith Bitcoin, the Middleman Is Dead, Long Live the Middleman !", "\n\nThey can exchange money with no pre-existing trust. ", "They don’t have to trust a private bank or the government with their money.", "\n\nTrust is hugely important in the developing countries where corruption is known to be relatively high.", "\n\nBut.. Bitcoin can’t teach people how to read\n\nSo what ? ", "No it can’t and it won’t need to. ", "Making a bitcoin transaction would eventually be as easy as making a phone call. ", "There are 4.88 billion cellphone users in the world and they are doing just fine using their phones. ", "There has never been an easier way to pay than bitcoin. ", "Ok maybe cash, but that’s it.", "\n\nBitcoin, as a money transfer service.", "\n\nAccording to a research by ODI , Africa’s diaspora pays 12% to send $200 back home. ", "And according to the same research, Western Union in Rwanda charges closer to 18–19%.", "\n\nA new innovative Bitcoin startup in Kenya, Bitpesa, is trying to tackle this problem by taking bitcoins and paying out Kenyan shillings.", "\n\n“Bitcoin offers an immediate remedy for cross border payments, remittances, online payments and business-to-business payments because it takes less than an hour to send and receive money, with an added advantage of final settlement” — Michael Kimani, Chairman of the African Digital Currency Association.", "\n\nWhen Muhammad Yunus’ meets Bitcoin\n\nYunus is known as a pioneer of the Microcredit concept, “small loans to very poor people for self-employment projects that generate income, allowing them to care for themselves and their families.”", "\n\nTheses small loans have helped lift millions of families out of poverty in the developing world by helping them purchase a motorbike or start a business. ", "These loans have also allowed for the bulk purchase of food at a lower cost than the ‘normal’ behaviors of buying it on a daily basis. ", "For the poor, small change is big change.", "\n\nMicrofinance has been a revolution of sorts as it was really the first thing that took some power away from the banks. ", "It reversed the power relationship between the borrower and lender.", "\n\nWith Bitcoin, transactions can happen from person to person without the need of a trusted party.", "\n\nNow with Bitcoin, loans can also occur without the need of a middlemen.", "\n\nLoans of the people, by the people, and for the people.", "\n\nPeer-to-peer lending using Bitcoin is already happening with companies like BTCJam, a service that connect individual lenders with borrowers.", "\n\nBut what about their credit scores ?", "\n\nThere is a real possibility of creating a credit score system for the Poor using Bitcoin.", "\n\nUsing Machine learning, BTCJam already built a proprietary credit-scoring system.", "\n\nClosing Thoughts\n\nIn his piece on the state of consumer FinTech, @TanayJ wrote that one of the three way technology is disrupting the consumer finance industry is by “Enabling individuals to borrow, invest and transfer money at lower rates than traditional financial institutions”.", "\n\nUntil now, most of the services offered in the developing region are trying to enable people to transfer money at a lower rate than traditional services.", "\n\nWhen it comes to the Poor, we tend to forget that Y = C + S.\n\nBy focusing on offering them a transactional service and forgetting about improving their saving and borrowing options, we failed to convince them to leave cash and go digital.", "\n\nSurely we could do better. ", "Maybe it’s time to try something else.", "\n\n“You never change things by fighting the existing reality. ", "To change something, build a new model that makes the existing model obsolete.” — ", "Buckminster Fuller\n\nBitcoin still has a long way to go. ", "It is still fairly inaccessible to the developing world.", "\n\n“Bitcoin is for money what TCP/IP is to the Information Age” -Micky Malka\n\nCurrently, Bitcoin is a just like the internet before the browser. ", "It will take time until we build robust and realistic solutions for the developing world.", "\n\nBut it will eventually happen, and it will change everything.", "\n\nWe are just getting started.", "\n\nThanks to Saul Minkoff & Tanay Jaipuria for reviewing a draft of this. ", "If you want to discuss anything on this piece, I’m @mmkhalifaa on Twitter." ]

{ "pile_set_name": "OpenWebText2" }

[ 0.011695906432748537, 0.013333333333333334, 0, 0, 0.011111111111111112, 0.006060606060606061, 0, 0, 0, 0, 0.01818181818181818, 0, 0, 0, 0, 0, 0, 0.005988023952095809, 0, 0, 0, 0.00558659217877095, 0, 0, 0, 0.0044444444444444444, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.007042253521126761, 0.04, 0.005479452054794521, 0, 0, 0, 0, 0.0625, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.011627906976744186, 0.011764705882352941, 0, 0.006535947712418301, 0.01276595744680851, 0, 0, 0, 0, 0, 0, 0, 0, 0.006993006993006993, 0, 0, 0, 0.007067137809187279, 0, 0, 0, 0, 0, 0, 0.017857142857142856, 0, 0.013888888888888888, 0, 0, 0, 0.0136986301369863, 0.013513513513513514 ]

[ "Q:\n\nTemplated requires-expression\n\nI want a Functor concept in C++20.", "\nA functor is a higher-kinded type that can be mapped over. ", "A simple example is std::optional; with a function from type A to type B and a std::optional<A>, you can easily create a std::optional<B> by applying the function to the value if it exists and returning an empty optional otherwise. ", "This operation is called fmap in Haskell.", "\ntemplate<typename A, typename B>\nstd::optional<B> fmap(std::function<B(A)> f, std::optional<A> fa) {\n if (!", "fa) {\n return std::optional<B>{};\n }\n return std::optional<B>(f(*fa));\n}\n\nA concept for all functors is simple enough to write. ", "I've come up with this (using GCC—you'll have to remove bool to get this working in Clang, I think):\ntemplate<template<typename> typename F, typename A, typename B>\nconcept bool Functor = requires(std::function<B(A)> f, F<A> fa) {\n { fmap(f, fa) } -> F<B>;\n};\n\nAnd a simple additional function to make sure this works:\ntemplate<typename A, typename B>\nstd::function<B(A)> constant(B b) {\n return [b](A _) { return b; };\n}\n\ntemplate<template<typename> typename F, typename A, typename B>\nF<B> replace(B b, F<A> fa) requires Functor<F,A,B> {\n return fmap(constant<A,B>(b), fa);\n}\n\nIt works. ", "But it's not pretty. ", "What I want is for the signature of replace to read like so:\ntemplate<Functor F, typename A, typename B>\nF<B> replace(B b, F<A> fa);\n\nNo need for a requires-clause here. ", "Much nicer, don't you agree? ", "To get this to work, however, I'd have to reduce the template on my concept to a single argument. ", "Something like this:\ntemplate<template<typename> typename F>\nconcept bool Functor = requires(function<B(A)> f, F<A> fa) { // Uh-oh\n { fmap(f, fa) } -> F<B>;\n};\n\nProblem is, I've not declared types A and B. As far as I can tell, there's nowhere I can declare them before I have to use them. ", "Can what I want be done, and can it be done simply and elegantly?", "\nOne possible solution that comes to my mind is to make the requires-expression in the concept a template (or at least a template-like thing). ", "I'd then have something like this:\ntemplate<template<typename> typename F>\nconcept bool Functor = requires<typename A, typename B>(function<B(A)> f, F<A> fa) {\n { fmap(f, fa) } -> F<B>;\n};\n\nUnfortunately, this is not valid by the C++20 standard and will not compile with g++-8. ", "Could something like this be viable? ", "Could it make it into the standard?", "\n\nA:\n\nYour \"Functor\" concept represents a complex relationship between three different types: F (a one-parameter template that is the template being projected over), A (the starting object type), and B (the resulting object type). ", "Your concept represents a relationship between 3 parameters, so your concept is going to have to take 3 parameters.", "\nTerse template syntax is for simple cases: constraints relating to a single (type) parameter. ", "Your case is not simple, so you're going to have to spell it out with a requires clause. ", "Anytime you have a concept with multiple parameters like this, you'll have to spell it out.", "\nAs for whether it is \"pretty\" or not, that's a value judgement. ", "But given the complex relationship on display here, seeing it spelled out makes it clear what the relationship between all of these parameters is. ", "And clarity has a beauty all its own.", "\n\nA:\n\nC++ doesn't have parametric polymorphism like this - you can't do things like \"for any type\" in the way you want to, and in the way you can in Haskell. ", "I think that's fundamentally impossible in a world where overloading exists. ", "\nWhat you have is the following (I went ahead and removed the erroneous bool, which isn't part of C++20 concepts, and fixed -> Type, which was also removed):\ntemplate<template<typename> class F, typename A, typename B>\nconcept Functor = requires(std::function<B(A)> f, F<A> fa) {\n { fmap(f, fa) } -> std::same_as<F<B>>;\n};\n\nWhat you want to say is for any types a and b, given an a -> b, you can call this function. ", "We can't do that. ", "But we can just pick arbitrary types ourselves. ", "One way of doing that is to pick secret types that the functor implementation just wouldn't know about:\nnamespace secret {\n struct A { };\n struct B { };\n\n template <typename From, typename To>\n struct F {\n auto operator()(From) const -> To;\n };\n}\n\ntemplate <template <typename> class F>\nconcept Functor = requires(secret::F<secret::A, secret::B> f, F<secret::A> fa) {\n { fmap(f, fa) } -> std::same_as<F<secret::B>>;\n};\n\nThat's probably your best bet. ", "You could even add multiple a/b pairs to make this more likely to be correct. ", "\nRegardless, this:\ntemplate<Functor F, typename A, typename B>\nF<B> replace(B b, F<A> fa);\n\nIsn't going to happen anyway, since we don't have that kind of terse syntax for constrained template template parameters. ", "You'd have to write it this way:\ntemplate <template <typename> class F, typename A, typename B>\n requires Functor<F>\nF<B> replace(B b, F<A> fa);\n\nAs a side note, this is a bad implementation of fmap for optional:\ntemplate<typename A, typename B>\nstd::optional<B> fmap(std::function<B(A)> f, std::optional<A> fa);\n\nTaking a std::function<Sig> means this will only work if you pass in specifically a std::function. ", "Not for lambdas or function pointers or other function objects (like the secret::F I used earlier). ", "And even if it did work, you wouldn't want to do this anyway, since it's needless overhead.", "\nYou want:\ntemplate <typename F, typename A, typename B = std::invoke_result_t<F&, A const&>>\nstd::optional<B> fmap(F f, std::optional<A> fa);\n\nI have a whole post about this exact problem, Declarations Using Concepts. ", "\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0, 0, 0, 0.024390243902439025, 0, 0, 0.0016722408026755853, 0, 0, 0, 0, 0.0033783783783783786, 0, 0, 0, 0, 0, 0, 0, 0.010526315789473684, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.0045662100456621, 0 ]

[ "Ladies of Letters\n\nLadies of Letters is a comedy series that ran for 13 years on BBC Radio 4; based on the series of books of the same name written by Carole Hayman and Lou Wakefield, and starring Patricia Routledge and Prunella Scales. ", "Routledge stars as Vera Small and Scales as Irene Spencer, two elderly ladies who share a sometimes rocky relationship conducted by correspondence.", "\n\nThe success of the first series of Ladies of Letters (broadcast in 1997) spawned adaptations of several sequels: Ladies of More Letters (1999); Ladies of Letters.", "Com (2000); Ladies of Letters Log On (2001); Ladies of Letters Make Mincemeat (2001 Christmas Special); Ladies of Letters Spring Clean (2004); Ladies of Letters Go Global (2006); Ladies of Letters Say No (2007); Ladies of Letters Go Green (2008); Ladies of Letters Crunch Credit (2009) and Ladies of Letters Go Crackers (2010 - this series features Anne Reid instead of Patricia Routledge as Vera Small, see below).", "\n\nEach of the 15-minute episodes come as part of BBC Radio 4's long running magazine programme Woman's Hour. ", "A ten-part television adaptation of the first radio series has been made for ITV3, which began on 3 February 2009.", "\n\nPlot and principals\nVera and Irene reveal their exploits and adventures to each other in their letters and e-mails, but sometimes their correspondence becomes fractious when one accuses the other of being an alcoholic or engages in too much one-upmanship. ", "Nevertheless, when the chips are down and the going gets tough, each is instantly there for the other, like a charge of the cavalry but with a more sarcastic bugle call.", "\n\nOther recurring characters include people we occasionally hear, as well as some we only hear about:\n\nHoward Small (\"Howie\"), Vera's gay son, who runs a sheep farm near the village of Great Shagthorn with his business and personal partner, Anthony Flowers (\"Ants\"). ", "Howie and Ants have a daughter, \"Small-Flowers\" currently known as \"Flo\";\nKaren, Vera's daughter, with whom she has a difficult relationship. ", "Karen's second husband, St. John (pronounced \"sinjun\"), is a veterinarian with whom Vera has an excellent relationship. ", "St. John is also revealed to be the biological father of Baby Small-Flowers. ", "Karen and St. John have two children together: the long-legged Nelson and Millie (so-named after her premature arrival around the time of the new \"Minnellium\" [sic]);\nSabrina Small, Karen's daughter whose father is never mentioned. ", "Karen was married and divorced prior to Sabrina's birth.", "\nLesley, Irene's daughter, who lives in Australia with her second husband Brian and their two children, Cheryl-Marie (from Lesley's first marriage ('Cheryl' pronounced 'CHAIR-ull') and Bubbles (real name: Sarah-Jane). ", "Brian has a squint Lesley finds embarrassing, and a too-obvious hair-transplant;\nChristopher Thoroughgood, Irene's long-lost son, Lesley's elder half-brother. ", "Christopher is the result of an \"illicit and unwanted infringement\" of Irene's maidenhood and was put up for adoption. ", " Christopher's first wife was Margaret and they had two sons, \"Little Christopher\" and Tommy. ", " Christopher's second wife is Michaela, who appears, with their baby daughter Sophie-Irene, out of the blue in Ladies of Letters Crunch Credit;\nYoussou, Vera's \"adoptive grandson\", a former child soldier in Panglawangla, whom she brought back with her from her 'Global' adventure;\nBeryl \"next door\", who was Irene's neighbour and has recurrent medical problems with her bottom;\nNellie Thoroughgood, Christopher's adoptive mother. ", " She is played by Susan Jameson for her sole appearance in Ladies of Letters Go Global, in which she calls herself Nellie Havergood. ", " No reason is given for the difference in surname.", "\n\nVera and Irene always strive to outdo each other, whether with their recipes, holiday destinations or who has the best grandchildren. ", " The humour is mostly derived from the lack of insight of the two main protagonists, but there is also a somewhat melancholic theme that lurks just beneath the surface at the way these two women are taken for granted by their respective families. ", " By the end of 2004's Ladies of Letters Spring Clean it is implied that there has been something of a parting of the ways between the women and their kinfolk. ", "Later series tend to heavily use malapropisms for comic effect.", "\n\nLadies of Letters Go Global (broadcast 2 to 6 January 2006) featured a somewhat bizarre and unlikely world tour undertaken by the two women, in contrast to the more naturalistic tone of previous stories. ", " The BBC issued a CD collection of the first seven series as a special 'biscuit tin' release in October 2006. ", "A short, newly written, episode was featured as part of the Woman's Hour 60th Anniversary concert in October 2006.", "\n\nLadies of Letters Say No was broadcast from 27 to 31 August 2007 and featured a broadly satirical take on the involvement of the private sector in the NHS, concentrating on the underhand dealings of the tycoon Alan Trumper. ", " Ladies of Letters Go Green – the ninth series – was broadcast from 10 to 14 March 2008. ", " Series 10 – Ladies of Letters Crunch Credit – was broadcast from 4 to 8 May 2009.", "\n\nAn eleventh radio series Ladies of Letters Go Crackers was broadcast at Christmas 2010 (27–31 December). ", "Patricia Routledge was not available at the time of recording so Anne Reid, who plays Vera on television, stepped into the role.", "\n\nTelevision series\nA ten-part television adaptation of the first radio series was made by Tiger Aspect for ITV3, which ran from on 3 February 2009. ", "It stars Maureen Lipman as Irene and Anne Reid as Vera. ", "All episodes are on the series one DVD, released in July 2009. ", "The second series was aired in two parts; the first airing in April 2010 and the second aired in October 2010.", "\n\nBook details\nLadies of Letters ()\nMore \"Ladies of Letters\" ()\nLadies of Letters.com ()\nLadies of Letters Log On ()\nLadies of Letters Make Mincemeat ()\nLadies of Letters Spring Clean ()\nLadies of Letters Go Global ()\nLadies of Letters Say No ()\nLadies of Letters Go Green ()\nLadies of Letters Crunch Credit ()\nLadies of Letters Go Crackers ()\n\nReferences\n\nExternal links\n Ladies of Letters official website\n\nCategory:1997 radio programme debuts\nCategory:2000s British comedy television series\nCategory:2009 British television series debuts\nCategory:BBC Radio 4 programmes\nCategory:BBC Radio comedy programmes\nCategory:Television series by Endemol" ]

{ "pile_set_name": "Wikipedia (en)" }

[ 0.02109704641350211, 0.02040816326530612, 0.012195121951219513, 0.02650602409638554, 0.01834862385321101, 0.008771929824561403, 0.003875968992248062, 0, 0.0149812734082397, 0.02097902097902098, 0.016666666666666666, 0, 0.021551724137931036, 0.03571428571428571, 0.03211009174311927, 0.031446540880503145, 0.01680672268907563, 0.0425531914893617, 0.018604651162790697, 0.022556390977443608, 0, 0.007352941176470588, 0, 0.006289308176100629, 0, 0, 0.00909090909090909, 0, 0.008849557522123894, 0, 0, 0.009345794392523364, 0.0234375, 0.013422818791946308, 0.05357142857142857, 0, 0, 0.006172839506172839 ]

[ "---\nabstract: |\n We suggest a general ansatz for the energy-eigenstates when a complex one-dimensional ${\\cal PT}$-symmetric potential possesses real discrete spectrum. ", "Several interesting features of ${\\cal PT}$-symmetric quantum mechanics have been brought out using this ansatz.\\\n PACS No.: ", "03.65.Ge\naddress: |\n Nuclear Physics Division, Bhabha Atomic Research Centre, Bombay 400 085\\\n zahmed@barc.gov.in\nauthor:\n- Zafar Ahmed\ntitle: 'An ansatz for the eigenstates in ${\\cal PT}$-symmetric quantum mechanics'\n---\n\n=10000\n\n.3 in The new option \\[1\\] that the non-Hermitian ${\\cal PT}$-symmetric Hamiltonians may too have real discrete spectrum has given rise to a lot of very interesting investigations. ", "Interestingly, in these developments exactly solvable models \\[2-8\\] have not only shown the ways but also preceded the general proofs so much that several general results are still unproved or unavailable.", "\n\nThe non-orthogonality of the new energy-eigenstate in the usual Hermitian sense was first encountered in exactly solvable models (e.g., \\[4\\]). ", "Then in several independent \\[4,10-13\\] studies a new scalar product (${\\cal PT}$-norm, ${\\cal PT}$-orthogonality) was proposed. ", "Even the indefiniteness of the new norm which is well known now was first displayed only in an exactly solvable model (e.g., \\[9\\])\n\nSo far it is not stated as to when expectation values of various operators $x,f(x),p,p^2,H$ etc. ", "under the new scalar product will be real. ", "In the present work we provide answer to this question by assuming an ansatz for the eigenstates for a complex one-dimensional ${\\cal PT}$-symmetric potential when it possesses real discrete spectrum.", "\n\nA general ${\\cal PT}$-symmetric Hamiltonian can be written as $$H={p^2_x \\over 2m}+V_c(x)=-{d^2 \\over dx^2}+V_0(x)+i\\lambda V_1(x),\n~~\\hbar=1=2m, \\lambda~ real,$$ where real functions $V_0(x)$ and $V_1(x)$ are even and odd respectively. ", "We now propose a general ansatz for the $n^{th}$ eigenstate of the Hamiltonian (1), as $$\\Psi_n(x)=\\psi_{n,0}(x)+i\\psi_{n,1}(x),~~s.t.~\n\\psi_{n,l}(-x)=(-)^{n+l}\\psi_{n,l}(x),$$ where $\\psi_{n,l}(x)$ are real and essentially vanish at $x=\\pm \\infty~ or~ \\pm L$. In the above equation by noticing the novel parity scheme of $\\psi_{0,1}(x)$, one can check that the eigenfunctions of all the exactly solvable models complex one-dimensional ${\\cal PT}$-symmetric potentials conform to this interesting ansatz. ", "In the following the indices (subscripts : $0,1,n$) will be playing a very interesting role while appearing in the integrands of all the integrals in the sequel. ", "If sum of these indices is even the integral will survive and will vanish otherwise.", "\n\nIn one dimension, when the Hamiltonian and eigenfunctions are real we can write $\\cal{H^\\ast}={\\cal}H$, $\\phi^\\ast_n=\\phi_n$ and $({\\cal H}\\phi_n)^\\ast=\n{\\cal H}^\\ast\\phi^\\ast_n={\\cal H}\\phi_n$ as the complex-conjugation operator does not [*transpose*]{}. ", "In the same way, let us notice that $H^{PT}=H$ and $\\Psi^{PT}_n=(-)^n \\Psi_n$. Let us write the eigenvalue equation for Eqs.(1) and (2), assuming that the eigenvalues are complex $E_n=E^r_n+iE^i_n$, $$H\\Psi_n=(E^r_n+iE^i_n)\\Psi_n.$$ Noting that if $c$ is a constant then $(c)^{PT}=c^\\ast$, we do ${\\cal PT}$-operation in Eq. (", "3) to get $$H\\Psi_n=(E^r_n-iE^i_n)\\Psi_n.$$ Subtracting Eq. (", "4) from (3), we find that $E^{i}_n=0$, showing that all the eigenvalues of $H$ are real when the ${\\cal PT}$-symmetry is exact: the energy-eigenstate (2) is also the simultaneous eigenstate of the antiliear operator ${\\cal PT}$.\n\nThe expectation value of an operator $\\hat O$ is usually defined as $$<n|\\hat O|n> ={1\\over N_n}\\int_{-\\infty}^{\\infty} \\Psi^\\ast _n(x)~\n\\hat O~ \\Psi_n(x)~dx.$$ When the boundary is at a finite distance, we will replace the limits by $\\pm L$. Let us also define the new expectation value as $$(n|\\hat O|n)={1 \\over N^\\prime_n}\\int_{-\\infty}^{\\infty} \\Psi_n(x)~\n\\hat O~ \\Psi_n(x)~dx,$$ where the complex conjugation has been discreetly given up. ", "In the following we shall be calculating both $<n|\\hat O|n>$ and $(n|\\hat O|n)$ for various quantal operators to compare and contrast. ", "Since the ${\\cal PT}$-symmetric potentials and their eigenstates corresponding to real part of the spectrum recieve increasing attention in the most recent years, the predictions as to whether the expectation value or the just defined averages of various operators would be real or imaginary, zero or non-zero and positive or negative would be valuable. ", "In the following we will see that the proposed ansatz and the arguments based on parity of various entities would suffice to predict important results.", "\n\nLet us calculate the norm, $N_n=<n|1|n>$ $$N_n=\\int_{-\\infty}^{\\infty} \\Psi^\\ast_m(x) \\Psi_n(x) ~dx=\\int_{-\\infty}^{\\infty}\n[\\psi^2_{n,0}(x)+\\psi^2_{n,1}(x)] ~dx.$$ Let us also calculate a quantity $N^\\prime_n=(n|1|n)$. $$\\begin{aligned}\nN^\\prime_n= \\int_{-\\infty}^{\\infty} \\Psi_n(x) \\Psi_n(x) ~dx\n&=&\\int_{-\\infty}^{\\infty}\n[\\psi^2_{n,0}(x)-\\psi^2_{n,1}(x)+2i\\psi_{n,0}(x)~\\psi_{n,1}(x)] ~dx\n\\nonumber \\\\\n&=&\\int_{-\\infty}^{\\infty} [\\psi^2_{n,0}(x)-\\psi^2_{n,1}(x)] ~dx.\\end{aligned}$$ In view of the interesting odd parity of the integrand (see Eq. (", "2)) the the second integral vanishes. ", "This simply proves that ${\\cal PT}$-norm ($N^\\prime_n$) will be real but indefinite (positive/negative) unlike the Hermitian norm ($N_n$) given in (7) which is positive definite.", "\n\nNow we calculate $<n|x^{2k}|n>, k=1,2...$ by noticing the parity of various integrands in view of Eq.(2), we get $$<n|x^{2k}|n>={1 \\over N_n}\\int_{-\\infty}^{\\infty} x^{2k} [\\psi^2_{n,0}(x)\n+\\psi^2_{n,1}(x)]~dx$$ which is real. ", "Similarly, one can show that $$<n|x^{2k+1}|n>=0$$ On the other hand, using a similar analysis we find that $$(n|x^{2k}|n)={1 \\over N^\\prime_n}\\int_{-\\infty}^{\\infty} x^{2k}\n[\\psi^2_{n,0}(x)-\\psi^2_{n,1}(x)]~dx$$ is real. ", "Further, $$(n|x^{2k+1}|n)={2i\\over N^\\prime_n}\\int_{-\\infty}^{\\infty} x^{2k+1}\n\\psi_{n,0}(x)\\psi_{n,1}(x)~dx,$$ which is imaginary. ", "By noting that differentiation changes the parity of a definite parity function and that $\\psi_{n,l}(\\pm \\infty)=0,$ we find that $$<n|p_x|n>={2\\over N^\\prime_n} \\int_{-\\infty}^{\\infty} \\psi^\\prime_{n,0}(x)\n\\psi_{n,1}(x) ~dx,$$ is real while $\\psi^\\prime_{n,0}(x)={d\\psi_{n,0}(x) \\over dx}$. But on the other hand, we find that $$(n|p_x|n)=0$$ As the double differentiation does not change the parity of a definite parity function we use Eq.(3) to find that $$<n|p^2_x|n>={1 \\over N_n}\\int_{-\\infty}^{\\infty} [E_n-V_0(x)][\\psi^2_{n,0}(x)\n+\\psi^2_{n,1}(x)]~dx$$ is real. ", "Whereas $$(n|p^2_x|n)={1\\over N^\\prime_n}\\int_{-\\infty}^{\\infty} \\{[E_n-V_0(x)]\n[\\psi^2_{n,0}(x)-\\psi^2_{n,1}(x)] +2\\lambda V_1(x)\\psi_{n,0}(x)\\psi_{n,1}(x)\\}\n~dx$$ is also real. ", "Next, we find the expectation value of $V_c(x)=V_0(x)+i\\lambda\nV_1(x)$, we have $$<n|V_c(x)|n>={1\\over N_n}\\int_{-\\infty}^{\\infty} V_0(x)[\\psi^2_{n,0}(x)\n+\\psi^2_{n,1}(x)]~dx$$ and $$(n|V_c(x)|n)={1 \\over N^\\prime_n}\\int_{-\\infty}^{\\infty} \\{V_0(x)\n[\\psi^2_{n,0}(x)-\\psi^2_{n,1}(x)] -2\\lambda V_1(x)\\psi_{n,0}(x)\\psi_{n,1}(x)\\}\n~dx.$$ By adding Eq.(15) with Eq.(17) and Eq.(16) with Eq.(18), we prove that $$<n|H|n>=E_n=(n|H|n) .$$ This is a remarkable result which establishes the equality of expectation values of the Hamiltonian under the old (5) and new (6) definitions. ", "We now take up the issue of the orthogonality of $\\Psi_n(x)$. For two states $m\\ne n$ with distinct eigenvalues $E_m$ and $E_n$ (complex or real), let us write the Schr[ö]{}dinger equation as $$H\\Psi_m(x)=E_m\\Psi_m(x),$$ $$H\\Psi_n(x)=E_n\\Psi_n(x).$$ Let us left-multiply (20) by $\\Psi_n(x)$ and (21) by $\\Psi_m(x)$ and subtract both the equations and integrate $w.r.t.x$ to have $$\\begin{aligned}\n(E_m-E_n)\\int_{-\\infty}^{\\infty} \\Psi_m(x)\\Psi_n(x)~dx \\nonumber\\\\\n=\\int_{-\\infty}^{\\infty}[\\Psi_n(x)H\\Psi_m(x)-\\Psi_m(x)H\\Psi_n(x)]~dx\n\\nonumber \\\\\n=-\\int_{-\\infty}^{\\infty} \\left \\{\\Psi_n(x){d^2 \\over dx^2}\\Psi_m(x)\n-\\Psi_m(x) {d^2 \\over dx^2}\\Psi_n(x)\\right \\}~dx \\nonumber \\\\\n=\\left [\\Psi_n(x){d\\Psi_m(x) \\over dx}-\\Psi_m(x){d\\Psi_n(x) \\over dx}\n\\right ]_{-\\infty}^{\\infty}~=~0.\\end{aligned}$$ The last term vanishes due to the boundary condition i.e., $\\Psi_j(\\pm\\infty)=0$. And so we have proved the orthogonality of two eigenstates corresponding to two distinct eigenvalues for a ${\\cal PT}$-symmetric Hamiltonian in general. ", "Thus, the orthogonality condition for the eigenstates (2) of an arbitrary ${\\cal PT}$-invariant potential can now be stated as $$(E_m-E_n)\\int_{-\\infty}^{\\infty} \\Psi_m (x)~\\Psi_n (x)~dx=0,~m \\ne n,$$ notice the absence of complex conjugation. ", "Using Eqs. (", "20) and (21) and bearing in mind that the complex-conjugation does not [*transpose*]{}, we can derive $$\\begin{aligned}\n(E_n-E_m)\\int_{-\\infty}^{\\infty}\\Psi^\\ast_m(x) \\Psi_n(x)~dx\n=2i\\int_{-\\infty}^{\\infty} \\Psi^\\ast_m(x) \\lambda V_1(x) \\Psi_n(x)~dx\n\\nonumber \\\\\n=2\\int_{-\\infty}^{\\infty}\\lambda V_1(x)[\\psi_{m,1}(x) \\psi_{n,0}(x)\n-\\psi_{m,0}(x)\\psi_{n,1}(x)]~dx, if~ m+n=even, \\nonumber \\\\\n=2i\\int_{-\\infty}^{\\infty} \\lambda V_1(x)[\\psi_{m,0}(x)\n\\psi_{n,0}(x)+\\psi_{m,1}(x) \\psi_{n,1}(x)]~dx, if~ m+n=odd.\\end{aligned}$$ It is helpful to note that whenever $\\psi_{p,q}(x)$ appears in an integral and if sum of all the indices of the integrand is even the integral survives and the integral vanishes otherwise. ", "The Eq.(24) demonstrates that $\\Psi_m(x)$ and $\\Psi_n(x)$ are not orthogonal in the conventional way. ", "For a better insight, let us see the orthogonality in another way by expressing $<m|n>$ and $(m,n)$ as $$\\begin{aligned}\n<m|n>=\\int_{-\\infty}^{\\infty} \\Psi^\\ast_m(x)~\\Psi_n(x)~dx,\n=\\int_{-\\infty}^{\\infty}\\{[\\psi_{m,0}(x)\\psi_{n,0}(x)+\\psi_{m,1}(x)\\psi_{n,1}(x)] \\nonumber \\\\\n+i[\\psi_{m,0}(x)\\psi_{n,1}(x) -\\psi_{m,1}(x)\\psi_{n,0}(x)]\\}~dx,\\end{aligned}$$ $$\\begin{aligned}\n(m|n)=\\int_{-\\infty}^{\\infty} \\Psi_m(x)~\\Psi_n(x)~dx,\n=\\int_{-\\infty}^{\\infty} \\{[\\psi_{m,0}(x)\\psi_{n,0}(x)-\\psi_{m,1}(x)\\psi_{n,1}(x)]\n\\nonumber \\\\\n+i[\\psi_{m,0}(x)\\psi_{n,1}(x)+\\psi_{m,1}(x)\\psi_{n,0}(x)]\\}~dx.\\end{aligned}$$ A comparison of Eq. (", "26) with Eq. (", "23) leads to two interesting additional properties of $\\psi_{p,q}(x)$, viz., ", "$$\\int_{-\\infty}^{\\infty} \\psi_{m,0}(x)\\psi_{n,0}(x)~dx=\\int_{-\\infty}^{\\infty}\n\\psi_{m,1}(x)\\psi_{n,1}(x)~dx, ~m \\ne n,$$ and $$\\int_{-\\infty}^{\\infty} \\psi_{m,0}(x)\\psi_{n,1}(x)~dx=-\\int_{-\\infty}^{\\infty}\n\\psi_{m,1}(x)\\psi_{n,0}(x)~dx, ~m \\ne n.$$ if the ansatz (2) were to represent the eigenstate of a ${\\cal PT}$-symmetric Hamiltonian with real discrete spectrum. ", "More importantly, notice once again that the usual orthogonality condition using the complex-conjugation would not hold as Eq. (", "26) does not vanish in view of properties Eqs.(27) and (28). ", "It is crucial to note that the right hand sides of Eqs. (", "24) and (25) would vanish due to different set of conditions. ", "A common but trivial condition would of course be when $\\lambda=0$ implying that the potential is real. ", "It is also imprtant to notice that $$\\int_{-\\infty}^{\\infty} \\psi^2_{n,0}(x)~dx \\ne \\int_{-\\infty}^{\\infty}\n\\psi^2_{n,1}(x)~dx.$$ from Eq. (", "27). ", "This means that ${\\cal PT}$-norm (8) in case of real discrete spectrum does not vanish.", "\n\nSo far we have worked with the proposed ansatz (2) which represnts only the scenario when the ${\\cal PT}$-symmetry is exact: the energy-eigenstates are also the eigenstaes of the anti-linear operator ${\\cal PT}$ and the discrete energy-eigenvalues are real. ", "We now consider the general case when the ${\\cal PT}$-symmetry could be broken or unbroken. ", "By operating with ${\\cal PT}$ in both the sides and remembering that $(c)^{PT}=c^\\ast$ and the fact that complex-conjugation does not [*transpose*]{}, using Eqs. (", "20,21) we find that $$\\begin{aligned}\n(E^\\ast_m-E_n)\\int_{-\\infty}^{\\infty} \\Psi^{PT}_m(x)\\Psi_n(x)~dx \\nonumber \\\\\n=\\int_{-\\infty}^{\\infty}[\\Psi_n(x)H\\Psi^{PT}_m(x)-\\Psi^{PT}_m(x)H\\Psi_n(x)] ~dx\n\\nonumber \\\\\n=-\\int_{-\\infty}^{\\infty} \\left \\{\\Psi_n(x){d^2 \\over dx^2}\\Psi^{PT}_m(x)\n-\\Psi^{PT}_m(x) {d^2 \\over dx^2}\\Psi_n(x)\\right \\}~dx, \\nonumber \\\\\n=\\left [\\Psi_n(x){d\\Psi^{PT}_m(x) \\over dx}-\\Psi^{PT}_m(x){d\\Psi_n(x) \\over\ndx} \\right ]_{-\\infty}^{\\infty}~=~0.\\end{aligned}$$ It is could be instructive to notice that the just derived orthogonality condition i.e., $$(E^\\ast_m-E_n) \\int_{-\\infty}^{\\infty} \\Psi^{PT}_m(x)~\\Psi_n(x)~dx~=~0.~$$ is more general as it degenerates to (24) when the ${\\cal PT}$-symmetry is unbroken i.e., $\\Psi^{PT}(x)= (-)^n\\Psi(x)$, and the eigenvalues are real, $(E^\\ast_m=E_m)$. When $m=n$ and $\\Psi_m(x)$ are no more eigenstates of ${\\cal PT}$, the ${\\cal PT}$-symmetry is broken and eigenvalues are complex conjugate pairs and consequently according to Eq. (", "31) the ${\\cal PT}$-norm vanishes. ", "Such states are also termed as self-orthogonal states \\[14\\]. ", "Following the suggestions in Refs. ", "\\[4,10-13\\], we could have also defined the expectation value $(n|\\hat O|n)$ (8) alternatively as $$(n|\\hat O|n)={1 \\over N^\\prime_n} \\int_{-\\infty}^{\\infty} \\Psi^{PT}_n(x) \\hat O \\Psi_n(x)~dx,\nN^\\prime_n=\\int_{-\\infty}^{\\infty} \\Psi^{PT}_n(x) \\Psi_n(x)~dx.$$ Since $\\Psi^{PT}_n(x)=(-)^n \\Psi_n(x)$ in the case of the real discrete spectrum, the above results remain unchanged. ", "An interesting account of the definition of the expectation (6) in the light of bi-orthogonality can be found in a very recent work \\[15\\].", "\n\nThe results proved in Eqs. (", "9-12,14,16,18) can be summarized by stating that in ${\\cal PT}$-symmetric quantum mechanics only ${\\cal PT}$-symmetric operators can have real expectation values. ", "This means the expectation values of $ix$ and $e^{ix}$ will be real. ", "This is akin to the conventional Hermitian quantum mechanics wherein only Hermitian operators have real eigenvalues. ", "The result in Eq. (", "19) is interestingly surprising which states that expectation value of the complex ${\\cal PT}$-symmetric Hamiltonian remains equal to the energy eigenvalue irrespective of the definitions of the expectation value given in Eq. (", "5) and (6). ", "This means the variational method \\[10\\] can work for the new Hamiltonians under both old and new definitions. ", "We remark that these results have not appeared before. ", "Eq.(8) proves the indefiniteness of the ${\\cal PT}$-norm. ", "Importantly, Eq. (", "29) proves that in the case of unbroken ${\\cal PT}$-symmetry (real discrete spectrm) the new norm does not vanish.", "\n\nWe would like to re-emphasize that all these results could be proved merely by the parity scheme of the suggested ansatz for the energy-eigenfunction in Eq. (", "2). ", "It may be verified that the energy-eigenfunction of all the exactly solvable one-dimensional models of the complex ${\\cal PT}$-symmetric potentials essentially conform to the suggested ansatz (2).", "\n\nReferences : {#references .unnumbered}\n=============\n\n1. ", " C.M.Bender and S. Boethcher, Phys. ", "Rev. Lett. [**", "80**]{} (1998) 5243.", "\n\n2. ", " M. Znojil, Phys. ", "Lett. ", "A [**259**]{} (1999) 220.", "\n\n3. ", " G. Levai and M. Znojil, J.Phys. ", "A: Gen. Math. [**", "33**]{} (2000) 7165.", "\n\n4. ", " Z. Ahmed, Phys. ", "Lett. ", "A [**282**]{}(2001) 343; [**287**]{} (2001) 295.", "\n\n5. ", " H.F. Jones, Phys. ", "Lett. ", "A [**262**]{} (1999) 242.", "\n\n6. ", " Z. Ahmed, Phys. ", "Lett. ", "A [**286**]{} (2001) 231.", "\n\n7. ", " Jia C-S, Zeng X-L, Sun L-T, Phys. ", "Lett. ", "A [**300**]{} (2002) 78.", "\n\n8. ", " A. Sinha, G. Levai, P. Roy, J. Phys. ", "A: Math. ", "Gen. [**36**]{} (2003) 7611.", "\n\n9. ", " B. Bagchi, C. Quesne, M. Znojil, Mod. ", "pgys. ", "Lett. ", "A [**16**]{} (2001) 2047; G. Levai, F. Cannata and A. Ventura, phys. ", "Lett. ", "A [**300**]{} (2002) 271.", "\n\n10. ", "C.M. Bender, F. Cooper, P.N. Meisinger, V.M. Savage, Phys. ", "Lett. ", "A 259 (1999) 224.", "\n\n11. ", "M. Znojil, math-ph/0104012.", "\n\n12. ", "G.S. Japaridze, J. Phys. ", "A: Math. ", "Gen. [**35**]{} (2003) 1709.", "\n\n13. ", "C.M. Bender, D.C. Brody and H.F. Jones, Phys. ", "Rev. Lett. [**", "89**]{} (2002) 270401.", "\n\n14. ", "E. Narevicius, P. Serra and N. Moiseyev, Eur. ", "Phys. ", "Lett. [**", "62**]{} (2003) 789.", "\n\n15. ", "A.V. Sokolov, A.A. Andrianov, F. Cannata, quant-ph/0602207.", "\n" ]

{ "pile_set_name": "ArXiv" }

[ 0.005813953488372093, 0.0078125, 0.014354066985645933, 0, 0, 0.015503875968992248, 0.004347826086956522, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.002962962962962963, 0.007407407407407408, 0, 0, 0.0036101083032490976, 0, 0, 0.004366812227074236, 0, 0, 0, 0, 0.0017391304347826088, 0.002909796314258002, 0.00819672131147541, 0.08333333333333333, 0.009845288326300985, 0, 0.0015723270440251573, 0.07142857142857142, 0, 0.005405405405405406, 0.0078125, 0, 0.017543859649122806, 0, 0, 0.014285714285714285, 0, 0, 0.0038461538461538464, 0, 0.006134969325153374, 0.005025125628140704, 0.02857142857142857, 0, 0, 0.005291005291005291, 0, 0, 0.006134969325153374, 0, 0, 0.05263157894736842, 0, 0, 0, 0, 0.017241379310344827, 0.05555555555555555, 0.008771929824561403, 0.00625, 0, 0, 0, 0.027777777777777776, 0.07142857142857142, 0, 0, 0, 0.16666666666666666, 0, 0, 0.06060606060606061, 0, 0, 0, 0.058823529411764705, 0.16666666666666666, 0, 0, 0, 0.16666666666666666, 0, 0, 0.058823529411764705, 0.16666666666666666, 0, 0, 0.05714285714285714, 0.16666666666666666, 0, 0, 0.07894736842105263, 0, 0, 0, 0.05128205128205128, 0, 0.16666666666666666, 0.043478260869565216, 0.16666666666666666, 0, 0, 0.05084745762711865, 0.16666666666666666, 0, 0, 0.037037037037037035, 0, 0.08, 0, 0, 0, 0.06521739130434782, 0.07142857142857142, 0, 0, 0.08695652173913043, 0, 0.1111111111111111, 0, 0, 0.03389830508474576, 0 ]

[ "We were delighted with all of the team events that you have presented, and we know that they have left a lasting impression on everyone in the team. ", "A huge thank you for your professionalism and creativity.", "\n\n—Rene W. Davis\n\nSunny's Surplus" ]

{ "pile_set_name": "OpenWebText2" }

[ 0, 0, 0.030303030303030304 ]

[ "Eco-Fuel Africa Works on Creating Sustainable Energy Technologies\n\nMoses Sanga's Eco-Fuel Africa is a social enterprise supporting sustainable energy technologies through the creation of green coal.", "\n\nOn the Eco-Fuel website, the team explains and shows some statistics of deforestation in Africa: \"In sub-Saharan Africa, 80 percent of the people cook their meals over wood fires. ", "The very poorest individuals cut down trees for cooking fuel; those slightly less poor buy charcoal made from wood in those same forests. ", "Every year Africa loses forest cover equal to the size of Switzerland. ", "At this rate Africa will have no forests left by 2050.\"", "\n\nThe featured image in the media gallery shows the business model and its interconnections in a graph, that is, how it melds micro-franchisees, charcoal powder, technology training, women retailers, farmers with kilns and final consumers." ]

{ "pile_set_name": "Pile-CC" }

[ 0.005050505050505051, 0.005494505494505495, 0, 0, 0, 0 ]

[ "A controlled trial of selective decontamination of the digestive tract in intensive care and its effect on nosocomial infection.", "\nNosocomial infection is a major problem in intensive therapy units (ITUs) and a significant cause of mortality. ", "Selective decontamination of the digestive tract (SDD) has been advocated as a means to reduce ITU morbidity and mortality. ", "Ninety-one patients in a general ITU underwent SDD, consisting of topical polymyxin E, tobramycin and amphotericin B administered throughout the unit stay together with parenteral ceftazidime for the first three days, and were compared with 84 historical and 92 contemporaneous control patients who were treated conventionally. ", "Twenty-seven historical and 32 contemporaneous control patients developed unit-acquired infections, in comparison with only three patients in the SDD group (P less than 0.01). ", "Mortality in the SDD group (36%) was not significantly different from that in the other two groups (historical control 40%, contemporaneous control 43%). ", "Screening specimens revealed a significantly higher rate of colonization with resistant Acinetobacter spp. ", "in the contemporaneous control than in the other two groups of patients; infection caused by resistant bacteria did not occur. ", "SDD did not lead to a significant reduction in the use of systemically-administered antibiotics when compared with either control group. ", "SDD may be used selectively in an ITU without ill effects on those patients not receiving SDD; nevertheless, microbiological monitoring is needed to detect emergence of resistant bacteria in the unit." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0.016129032258064516, 0.006097560975609756, 0.005681818181818182, 0.006493506493506494, 0.009345794392523364, 0, 0, 0.015 ]

[ "INDEX_CONTENT_REDIS_KEY=INDEX_CONTENT_REDIS_KEY\n\n#商品信息\nREDIS_ITEM_KEY=REDIS_ITEM_KEY\n#过期时间,默认为1天\nREDIS_ITEM_EXPIRE=86400\n\n\n#列表\nINDEX_CATEGORY_REDIS_KEY=INDEX_CATEGORY_REDIS_KEY\n" ]

{ "pile_set_name": "Github" }

[ 0.005649717514124294 ]

[ "The popliteal lymph node assay (PLNA) was proposed for the preclinical prediction of xenobiotics-induced autoimmune reactions in humans. ", "Among the substances so far tested in this model, procainamide and isoniazid gave negative PLNA responses despite reports of lupus syndromes in man. ", "To confirm the hypothesis that a metabolite instead of the parent molecule is involved, rats were pretreated with phenobarbital or beta-naphthoflavone, then injected with procainamide or isoniazid. ", "In additional groups of animals, procainamide or isoniazid were injected together with S9-mix following various incubation times. ", "Pretreated rats had a positive PLNA response when injected with procainamide, whereas preincubation with S9-mix resulted in a positive response to isoniazid. ", "These results further support the validity of the PLNA." ]

{ "pile_set_name": "Pile-CC" }

[ 0.0072992700729927005, 0.006711409395973154, 0, 0, 0, 0.01818181818181818 ]

[ "Intelligence and educational level in relation to body mass index of adult males.", "\nIn this study we explore the relationships between intelligence test score, educational level, and degree of body fatness. ", "We collected data on intelligence test score, educational level, height, and weight from a sample of 26,274 young Danish men and computed body mass index (weight/height2) as a measure of fatness. ", "Both intelligence test score and educational level had maximum values below the median for body mass index and declined monotonically thereafter. ", "The decline was equal for test score and educational level and above 32 kg/m2 corresponded to almost half a standard deviation. ", "These results are not readily explained by stigmatization of frank obesity, and other mechanisms, possibly genetic, may be responsible." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0, 0, 0, 0 ]

[ "Q:\n\nUIImage drawInRect doesn't draw with scale\n\nI'm trying to scale down a 2x image and then draw a UIImage in a rect, but somehow that has no effect. ", "Even though the property .scale changes, the images appears to be of the same scale.", "\nHere's the implementation for my drawRect()\n override func drawRect(rect: CGRect) {\n // Load the artwork\n let areaRect = CGRect(x: 0, y: 0, width: rect.width, height: rect.height)\n var art = UIImage(named: \"art-1-square-640\")\n // Make the circular mask\n let maskRef = UIImage(named: \"art-circular-mask\").CGImage\n let maskCircle = CGImageMaskCreate(\n UInt(rect.width),\n UInt(rect.height),\n CGImageGetBitsPerComponent(maskRef),\n CGImageGetBitsPerPixel(maskRef),\n CGImageGetBytesPerRow(maskRef),\n CGImageGetDataProvider(maskRef),\n nil,\n false)\n let artMaskedRef = CGImageCreateWithMask(art.", "CGImage, maskCircle)\n var artMasked = UIImage(CGImage: artMaskedRef, scale: 0.5, orientation: UIImageOrientation.", "Up);\n artMasked.drawInRect(areaRect)\n }\n\nA:\n\nThe scale:0.5 says that that there are two points per pixel in the image instead of the normal (for a @2x image) two pixels per point. ", "That would make the image 4x bigger. ", "However, drawInRect doesn't care, it just scales the image back down to fit the rectangle.", "\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0.013245033112582781, 0, 0, 0.016666666666666666, 0.005263157894736842, 0, 0, 0 ]

[ "Ultrasound contrast imaging research.", "\nThis article is a review of the research on ultrasound contrast agents in general imaging. ", "While general imaging contrast agent applications are still undergoing investigation and waiting FDA approval in the United States, they are approved for clinical use in Europe and other countries. ", "The contrast microbubble properties are described, including their nonlinear behavior and destruction properties. ", "Imaging techniques like harmonic imaging, pulse inversion, power pulse inversion, agent detection imaging, microvascular imaging, and flash contrast imaging are explained. ", "A connection is made between the aforementioned imaging techniques and the different contrast agents available. ", "The blood flow appearance of different liver tumors in the presence of contrast agents is demonstrated with examples." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0.005050505050505051, 0, 0, 0, 0 ]

[ "Parameter passing between molecular dynamics and continuum models for droplets on solid substrates: the static case.", "\nWe study equilibrium properties of polymer films and droplets on a solid substrate employing particle-based simulation techniques (molecular dynamics) and a continuum description. ", "Parameter-passing techniques are explored that facilitate a detailed comparison of the two models. ", "In particular, the liquid-vapor, solid-liquid, and solid-vapor interface tensions, and the Derjaguin or disjoining pressure are determined by molecular dynamics simulations. ", "This information is then introduced into continuum descriptions accounting for (i) the full curvature and (ii) a long-wave approximation of the curvature (thin film model). ", "A comparison of the dependence of the contact angle on droplet size indicates that the theories agree well if the contact angles are defined in a compatible manner." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0, 0.005747126436781609, 0, 0 ]

[ "High myopia causing bilateral abduction deficiency.", "\nWe present two cases of degenerative myopia with abduction deficiency. ", "Three mechanisms can explain the defect in the abduction: (a) the size of the long globe filling the space of the orbits, (b) the tightness of the medial recti due to long axis of the globe, and (c) longstanding esotropia becoming decompensated later in life. ", "We believe that high myopia is not a well-known cause of abduction deficiency, and it should be considered in the differential diagnosis." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0, 0 ]

[ "Rehabilitation of Combat-Related Injuries in the Veterans Administration: A Web of Support.", "\nThe Department of Veterans Affairs provides acute, subacute, and continuing rehabilitation for veterans using a hub-and-spoke system of hospitals and outpatient facilities. ", "Using traumatic brain injury as an example, this commentary illustrates how this system provides interdisciplinary rehabilitative care to veterans throughout North Carolina." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0.01098901098901099, 0.005747126436781609, 0 ]

[ "This invention relates to apparatus and methods for chemical mechanical polishing and cleaning a substrate in an integrated system.", "\nAn integrated circuit is typically formed on a substrate by the sequential deposition of conductive, semiconductive or insulative layers on a silicon wafer. ", "After each layer is deposited, it is etched to create circuitry features. ", "As a series of layers are sequentially deposited and etched, the outer or uppermost surface of the substrate, i.e., the exposed surface of the substrate, becomes increasingly non-planar. ", "This non-planar surface presents problems in the photolithographic steps of the integrated circuit fabrication process. ", "Therefore, there is a need to periodically planarize the substrate surface.", "\nChemical mechanical polishing (CMP) is one accepted method of planarization. ", "This planarization method typically requires that the substrate be mounted on a carrier or polishing head. ", "The exposed surface of the substrate is placed against a rotating polishing pad. ", "The polishing pad may be either a “standard” pad or a fixed-abrasive pad. ", "A standard pad has a durable roughened surface, whereas a fixed-abrasive pad has abrasive particles held in a containment media. ", "The carrier head provides a controllable load, i.e., pressure, on the substrate to push it against the polishing pad. ", "A polishing slurry, including at least one chemically-reactive agent, and abrasive particles if a standard pad is used, is supplied to the surface of the polishing pad.", "\nAfter the substrate has been polished, it is typically cleaned, e.g., by a brush scrubber or megasonic cleaner, to remove excess slurry, polishing chemistry and other debris from the polishing process. ", "After cleaning, the substrate is dried, e.g., by a spin-rinse drier, for return to the clean room. ", "In some systems, e.g., systems that use the Marangoni effect, the cleaning and drying functions can be combined. ", "Unfortunately, even after cleaning and drying, particles and other defects may remain on the substrate surface.", "\nThe chemical mechanical polishing and cleaning devices can be constructed as an integrated system with a single front-end automation interface between the clean room and the polishing and cleaning systems. ", "An example of such an integrated system is the Mirra Mesa™ chemical mechanical polishing system from Applied Materials, Inc. of Santa Clara, Calif. Such an integrated system can include multiple ports to receive multiple cassettes at the interface with the clean room. ", "Unfortunately, transport of the substrates between the cassettes and the polisher and cleaner can occasionally become backlogged, leading to a decrease in throughput." ]

{ "pile_set_name": "USPTO Backgrounds" }

[ 0, 0, 0, 0, 0, 0, 0.01282051282051282, 0, 0, 0, 0, 0, 0, 0, 0, 0.008849557522123894, 0, 0, 0.0037174721189591076, 0 ]

[ "Malignant glioma is very aggressive and notoriously difficult to treat. ", "With an annual incidence of around 5 cases per 100 000 it is relatively uncommon, but still comprises approximately 70% of malignant brain tumours ([@bib27]). ", "Owing to their tendency to infiltrate brain tissue, gliomas are very difficult to treat surgically, and the vast majority recur after apparent 'complete\\' resection ([@bib27]). ", "One of the first clinical trials ([@bib26]) showed that glioma patients receiving 'optimal conventional care\\' (no radiotherapy or chemotherapy) had a median survival of 3 months and a 1-year survival of 3%. ", "Various trials between 1970 and the 1990s supported the use of radiotherapy, but existing chemotherapies were of limited use. ", "Despite showing some promise in animal research ([@bib1]), the nitrosoureas BNCU and CCNU had little clinical efficacy. ", "Indeed, it took a large meta-analysis including \\>3000 patients ([@bib20]) to show a small increase in survival between radiotherapy with adjuvant chemotherapy (46% at 1 year, 20% at 2 years) and radiotherapy alone (40% and 15%, respectively).", "\n\nTemozolomide is an oral alkylating agent, a prodrug for 3-methyl-(triazen-1-yl)imidazole-4-carboxamide, whose anticancer activity was first described in 1987 ([@bib19]). ", "In 2005, a landmark phase III trial ([@bib21]) described the efficacy the new therapeutic agent temozolomide: when given together with radiotherapy and as adjuvant therapy there were significant increases in median survival (14 *vs* 12 months, *P*\\<0.001), 2-year survival and progression-free survival compared with radiotherapy alone. ", "Since then, temozolomide has emerged as the first-line chemotherapeutic agent for the treatment of malignant glioma ([@bib23]).", "\n\nAlthough temozolomide was being used clinically for melanoma in the mid-1990s, supported by substantial experimental data ([@bib2]), there were no such animal data on temozolomide for brain tumours. ", "The efficacy of temozolomide in glioma was actually first reported in patients somewhat anecdotally ([@bib12]), with a more structured study by the same group published in 1996 ([@bib11]). ", "The first animal work in the area was published in 1994 ([@bib15]), and by 1998 and 2000 the first patients were recruited for phase II ([@bib24]) and III ([@bib21]) trials, respectively. ", "A preliminary literature search of PubMed suggests that by the end of 2000, only seven relevant experimental studies had been published. ", "The basis for the decision to proceed to clinical trial is therefore not clear: possible factors include those experimental data, which did exist; the non-randomised clinical evidence of O\\'Reilly ([@bib12]) and Newlands ([@bib11]); and evidence from phase II studies testing temozolomide with other cancer types ([@bib2]; [@bib29]).", "\n\nSystematic review and meta-analysis of the efficacy of nitrosoureas ([@bib1]) and gene therapy (Conlin, unpublished observations) for experimental glioma models have shown some efficacy, but neither led to substantial improvement in outcomes in human clinical trials ([@bib20]; [@bib16]). ", "Identification of differences in the experimental data for nitrosourea, gene therapy and temozolomide studies might therefore provide insight into translational challenges in neuro-oncology.", "\n\nHere, we use systematic review and meta-analysis of experimental animal research to describe the evidence supporting the application of temozolomide in human glioma. ", "We also compare the evidence available before and after the publication of Stupp\\'s phase III trial ([@bib21]). ", "We were particularly interested in evidence within the experimental temozolomide data that was predictive of its successful translation into clinical use. ", "Secondary aims were (i) to compare temozolomide data from animal experiments before and after 2005 (the publication of the phase III trial); (ii) to seek evidence of the use of measures to avoid bias, and of publication bias; and (iii) to compare the efficacy of temozolomide in animal glioma models with that reported for nitrosoureas ([@bib1]) and gene therapy (Conlin, unpublished observations). ", "Our hypotheses were that temozolomide would significantly improve outcome in animal models of glioma; and second that publication and expectation biases would lead to significantly higher estimates of efficacy in studies published after the Stupp trial.", "\n\nMaterials and methods\n=====================\n\nInformation sources\n-------------------\n\nWe searched PubMed, EMBASE (from 1980) and Medline (from 1950) on the 4 August 2011, using these search terms: Temozolomide AND (glioma OR glioblastoma OR astrocytoma OR ependymoma OR oligodendroglioma OR 'brain tumour\\'). ", "For EMBASE and Medline the search was limited to animals, and for PubMed we used the Hooijmans ([@bib9]) animal search filter.", "\n\nStudy selection\n---------------\n\nAll abstracts and titles were screened by two authors (TCH and KJE) to include papers addressing temozolomide therapy in *in vivo* xenograft studies. ", "All publications in languages other than English were translated before screening. ", "We included studies that met the following inclusion criteria: (1) temozolomide used as monotherapy; (2) animal model of glioma used; (3) intracranial or subcutaneous implantation of tumour cells; (4) outcome reported as either median survival or tumour volume; and (5) the number of animals per group was stated within the publication.", "\n\nData collection\n---------------\n\nWe extracted data relating to experimental design and for each comparison, we recoded median survival or tumour volume in treated and control groups. ", "When data were not clearly or fully described we contacted authors seeking the data; if no reply was received after 1 week these were excluded. ", "Data were entered to the Collaborative Approach to Meta-Analysis and Review of Animal Data from Experimental Studies (CAMARADES) data manager application in Microsoft Access 2003.", "\n\nFor experimental design we extracted data for species, strain and sex of animals; glioma cell type, tumour implantation site, tumour implantation method, number or volume of implanted tumour cells and method of tumour volume measurement (where relevant). ", "Glioma models were stratified into groups ([Table 1](#tbl1){ref-type=\"table\"}). ", "In addition, we recorded the total dose of temozolomide (in mg kg^−1^), delay to treatment (in days) and route of administration. ", "When dose was given in mg m^−2^, we converted it to mg kg^−1^ using Freirich\\'s ([@bib8]) conversion factor; where animals received continuing treatment courses, we calculated the total dose received by the time of median survival or the last time at which tumour volume was measured. ", "Experimental starting point (i.e., day 0) was defined as the day of tumour implantation. ", "When the delay from implantation to treatment was not given, we contacted authors seeking this information.", "\n\nFor each comparison, we extracted data for the number of animals in each group and either median survival or mean tumour volume, relative tumour volume or percentage change in tumour volume. ", "For measures of tumour volume, we also recorded the s.d. ", "or s.e.m. ", "If any of these measures were missing, the publication was excluded. ", "When outcomes were not quantified in the text, we measured outcome values from Figures using AVPSoft 'Universal Desktop Ruler\\'.", "\n\nQuality assessment\n------------------\n\nStudy quality was evaluated for each publication using a modified 12-point checklist ([@bib1]) with one point allocated to each reported item; (1) peer-reviewed publication, (2) sample size calculation, (3) random allocation to groups, (4) blinded assessment of outcome, (5) compliance with animal welfare regulations, (6) statement of potential conflict of interests, (7) consistent volume or number of cells inoculated, (8) consistent site of tumour implantation, (9) reported number of animals in which the xenograft did not grow, (10) number of excluded animals stated, and reasons for exclusion given, (11) explanation of tumour model used, or multiple glioma models used and (12) presentation of evidence that temozolomide acts directly against tumour.", "\n\nData analysis\n-------------\n\nFor tumour volume data, efficacy was quantified using the normalised mean difference summary statistic ([@bib17]). ", "These effect sizes were then pooled using DerSimonian and Laird random effects meta-analysis ([@bib4]) to provide summary estimates of efficacy and to explore differences between groups of experiments.", "\n\nTo summarise median survival data we used the median survival ratio, MSR (treated survival divided by control survival) as an approach consistent with the gold standard hazard ratio method ([@bib10]; [@bib22]). ", "Median survival ratio was log-transformed to give a normal distribution ([@bib18]), and then log-MSRs were pooled using a modified DerSimonian and Laird random effects model. ", "As no measure of variance was available for MSRs, we weighted studies according to the number of animals used (the number of treated animals plus the number of control animals divided by the number of treatment groups per control group). ", "As measures of variance were not present in the extracted data, we estimated the standard error of summary estimates from the inter-study variance: random effects-weighted s.d. ", "of log-MSR from the random effects -- effect size was divided by the square root of the number of comparisons. ", "We used random effects-weighted s.e. ", "because the resulting estimate of variance is broader and therefore more conservative than unweighted s.e. ", "95% confidence intervals about the random effects mean log-MSR were calculated and the exponential of this log-MSR generated summary MSR data.", "\n\nFor stratified meta-analysis, we used the χ^2^ statistic, with n-1 degrees of freedom, to determine the extent to which stratification accounted for the observed heterogeneity. ", "We used Bonferroni correction to account for the number of stratifications -- with 12 strata for survival and 13 for volume data (1 extra stratifications for the method of volume measurement) giving critical values for significance of *P*=0.0043 and *P*=0.0039, respectively. ", "We used funnel plots, Egger regression ([@bib6]) and Trim and Fill analysis ([@bib5]) to seek evidence of publication bias; for survival data, we used the number of animals as the measure of precision ([@bib14]).", "\n\nResults\n=======\n\nElectronic searching identified 298 publications, of which 84 satisfied the inclusion criteria; 24 were excluded at data extraction ([Figure 1](#fig1){ref-type=\"fig\"}). ", "Four papers in foreign languages were identified of which one met our inclusion criteria. ", "From the remaining 60 publications, 123 and 26 individual comparisons were identified for median survival and tumour volume outcomes respectively, representing data from 2044 and 399 animals. ", "Temozolomide treatment led to significant improvements in survival (MSR: mean 1.88, 95% CI 1.74--2.03; χ^2^=388, *df*=122; *P*\\<0.0043) and reductions in tumour volume (50.4% (41.8--58.9); χ^2^=172, *df*=25; *P*\\<0.0039).", "\n\nThe median quality score was 6 out of 12 (IQR 5--7, range 3--8). ", "Of the 60 publications, none reported sample size calculations or allocation concealment; 32 reported random group allocation and 5 reported blinding of outcome assessment. ", "All studies used either mice or rats; mice were used more frequently for both survival and volume outcomes (89 *vs* 34; 18 *vs* 8), and 41 different glioma models were tested. ", "Information regarding delay to treatment was not available for seven publications.", "\n\nStudy characteristics\n---------------------\n\nRandomised treatment allocation was associated with an increase in the MSR (2.03 (1.83--2.26), *n*=66 *vs* 1.71 (1.53--1.90), *n*=57; *χ*^2^=13.2, *df*=1; *P*\\<0.0043) and a reduction in tumour volume (50.1% (40.1--60.0), *n*=19% *vs* 51.1% (33.7--68.4), *n*=7; *χ*^2^=18.7, *df*=1; *P*\\<0.0039). ", "Studies that blinded the assessment of outcome reported a greater reduction in tumour volume (64.4% (9.57--119), *n*=2% *vs* 48.8% (40.3--57.4), *n*=24; *χ*^2^=9.70, *df*=1; *P*\\<0.0039), but no association with MSR was observed. ", "Total quality score had a significant impact on tumour volume (*χ*^2^=44.5, *df*=5; *P*\\<0.0039 ([Figure 2B](#fig2){ref-type=\"fig\"}) but not survival ([Figure 2A](#fig2){ref-type=\"fig\"})). ", "Method of volume quantification (stratified into excision and histology, MRI, external calliper and unknown) was also associated with differences in reported volume reduction. ", "The single study measuring tumour volume by MRI reported greater temozolomide efficacy compared with those measuring volume by either excision and histology or external caliper (61.0% (50.5--71.6), *n*=1% *vs* 52.6% (38.9--66.3), *n*=9 and 52.1% (38.4--65.7), *n*=12, respectively: *χ*^2^=23.3, *df*=3; *P*\\<0.0039).", "\n\nAnimal and tumour models\n------------------------\n\nThere were no significant differences in temozolomide efficacy between rats and mice for either survival or reduction in tumour volume. ", "Athymic animals were commonly used in survival studies (*n*=50 publications) and occasionally to study effects on tumour volume (*n*=3). ", "Other comorbidities used with survival were severe compromised immunodeficient (*n*=1) and unspecified immunocompromised (*n*=11). ", "Comorbidity accounted for a significant proportion of between-study heterogeneity for survival (*χ*^2^=40.1, *df*=3; *P*\\<0.0043, [Figure 3A](#fig3){ref-type=\"fig\"}), with efficacy being significantly greater in animals with comorbidities, but not for tumour volume ([Figure 3b](#fig3){ref-type=\"fig\"}). ", "Almost all experiments reporting median survival used intracranial glioma models, the remainder using subcutaneous glioma (121 *vs* 2). ", "For tumour volume experiments, 12 comparisons used intracranial and 14 used subcutaneous models. ", "For both survival and volume experiments, the reporting of the amount of tumour cells injected was good -- most reporting either a number of cells (124 out of 146 comparisons, ranging from 1000 to 1 000 000 implanted cells, although only 11 comparisons were made with models using fewer than implanted 100 000 cells) or a set volume of implanted cells (14 out of 146, 1 mm^3^ in all cases).", "\n\nNext, we examined the impact of the glioma cell species of origin and the glioma cell type used ([Figure 4](#fig4){ref-type=\"fig\"}). ", "Human- and rat-derived cells predominated, with human glioma lines being more commonly used for both outcomes (MSR: *n*=87, volume: *n*=15). ", "The species of origin did not account for a significant proportion of the observed heterogeneity for MSR but temozolomide therapy was associated with a larger volume reduction in human-derived glioma cell lines (χ^2^=43.9, *df*=1; *P*\\<0.0039). ", "In both models, there was significant variation of efficacy between glioma models (MSR: χ^2^=53.6, *df*=12; *P*\\<0.0043; volume: χ^2^=66.7, *df*=6; *P*\\<0.0039), with GBM and 9L lines appearing to be most sensitive to temozolomide (for specifications of glioma lines, see [Table 1](#tbl1){ref-type=\"table\"} and [Supplementary Figure 2](#sup1){ref-type=\"supplementary-material\"}). ", "O-6-methylguanine-DNA methyltransferase (MGMT) status of the glioma models used was investigated and reported in only 11 of 60 publications.", "\n\nTemozolomide dosing\n-------------------\n\nFor temozolomide dosing regimens, we analysed data for total temozolomide dose ([Figures 5A and B](#fig5){ref-type=\"fig\"}), delay to treatment ([Figures 5C and D](#fig5){ref-type=\"fig\"}) and route of delivery ([Figures 5E and F](#fig5){ref-type=\"fig\"}); for treatment duration, number of cycles and other aspects of the dosing regime there was such diversity of approach that we did not consider a stratified analysis likely to be informative. ", "Both outcomes showed significant variability between dose stratified into groups of \\<50, 51--100, 101--500, 501--1000 and \\>1000 mg kg^−1^. The dose of temozolomide explained a significant proportion of the observed heterogeneity, with higher doses generally associated with greater efficacy (MSR: *χ*^2^=17.2, *df*=4; *P*\\<0.0043; volume: *χ*^2^=91.7, *df*=3; *P*\\<0.0039), although there may have been a decline in efficacy at total doses above 1 g kg^−1^ ([Figures 5A and B](#fig5){ref-type=\"fig\"}).", "\n\nDelay to treatment was stratified into groups of 0, 1--10, 11--20 and \\>20 days post-inoculation, and this accounted for a significant proportion of between-study heterogeneity for both MSR and volume reduction (MSR: *χ*^2^=47.0, *df*=4; *P*\\<0.0043; volume: *χ*^2^=45.9, *df*=3; *P*\\<0.0039). ", "For survival, efficacy was greatest the earlier temozolomide treatment was initiated. ", "However, we saw the converse for tumour volume: the longer the delay to treatment the more effective temozolomide was in reducing tumour volume. ", "Those studies in which treatment was initiated \\>20 days after tumour inoculation were associated with a greater reduction in tumour volume than those where the delay was 1--10 or 11--20 days.", "\n\nWhen we examined the effect of the route of drug administration. ", "We found that those treated with local temozolomide survived longer than those treated systemically (*χ*^2^=17.6, *df*=2; *P*\\<0.0043), but there was no difference in tumour volume reduction. ", "However, when we looked in more detail at systemic delivery routes for tumour volume reduction we found significant differences (*χ*^2^=75.7, *df*=3; *P*\\<0.0039), oral delivery providing the best survival benefit and intraperitoneal delivery having the largest effect on tumour volume.", "\n\nPublication bias\n----------------\n\nFor both MSR and tumour volume, there was no difference in reported efficacy before and after the Stupp publication ([Figures 6A and B](#fig6){ref-type=\"fig\"}). ", "For both survival and volume data, the intercept of Egger regression was positive ([Figures 6C--F](#fig6){ref-type=\"fig\"}) indicating an excess of small imprecise studies overstating efficacy. ", "Trim and Fill analysis supported asymmetry of survival data ([Figure 6G](#fig6){ref-type=\"fig\"}) -- corrected estimate of efficacy was reduced to 1.56 (1.42--1.70), compared with an unadjusted estimate of 1.88 (1.74--2.03), after the addition of 24 'missing\\' studies.", "\n\nDiscussion\n==========\n\nIn this systematic review and meta-analysis of 60 publications involving 2377 animals, we found that overall temozolomide therapy almost doubled survival and halved tumour volume. ", "Furthermore, observed efficacy seemed to be influenced by randomisation, blinding, total quality score, animal comorbidity, glioma model (including species), total dose, delay to treatment and route of temozolomide delivery.", "\n\nWe identified 17 papers published before 2005 describing the efficacy of temozolomide. ", "Although this is a reasonable basis on which to proceed to clinical trials it is a cause of some concern that many of these early publications, which were of most interest to us, did not present control data or measures of variance, and indeed nine did not present sufficient data to allow them to be included in this meta-analysis.", "\n\nOverall study quality was modest; the median number of study quality checklist items scored was 6 (of a possible 12) and no study scored higher than 8. ", "Studies that did not report simple measures to reduce bias such as randomisation and blinding of outcome assessment gave significantly higher estimates of efficacy, confirming an important role in experimental design in glioma similar to that established in other neurological disease models ([@bib7]; [@bib25]).", "\n\nPrevious meta-analyses in the animal modelling of glioma have been conducted using median survival as the primary outcome, but there is as yet no consensus about the most appropriate analysis of such data. ", "We used MSRs as an approximation of the gold standard hazard ratio approach ([@bib10]; [@bib22]). ", "By estimating standard error from the size of each study, we were able to combine data using a random effects approach. ", "As weights and the heterogeneity statistics were therefore calculated using study size to estimate standard error it is possible that this might have confounded the relative weighting given to individual studies. ", "However, given the high heterogeneity observed, in practice this approach has trended towards a simple average, with studies being given roughly equal weight. ", "As MSR analysis does not provide a measure of within-study variance, we calculated s.e. ", "and therefore 95% confidence intervals using inter-study variance to provide the most reliable measure of overall variance from the available data. ", "A major limitation of our analysis is, because it is univariate, it does not provide insight into how different variables interact with each other. ", "An example of this is the dose regimen and the interaction between different aspects of this; however, there were insufficient data to allow multivariate analysis to help disentangle the impact of multiple factors.", "\n\nDifferences between different experimental glioma models have previously been discussed ([@bib28]; [@bib1]). ", "There is a large amount of variation in neuropathology, natural history and drug sensitivity ([@bib28]; [@bib3]; [@bib30]) and this underpins the importance of glioma model selection during experimental design. ", "Of the 60 studies, 18 justified their choice of glioma model, or used multiple models for comparison. ", "Our data show a significant variation in temozolomide sensitivity between tumour models. ", "Previous evidence suggests that human-derived tumours are more sensitive to chemotherapy than those originating in rodents ([@bib1]): temozolomide also appeared to be more effective at reducing tumour volume in human-derived tumours, however, there was no difference in survival data and there may be an artefact of between-model heterogeneity. ", "Nevertheless, it remains clear that tumours show a great degree of variability in their sensitivity to treatment. ", "This may reflect either the degree of differentiation within the tumour or temozolomide resistance through a variety of mechanisms ([@bib30]), including MGMT expression. ", "Therefore, genomic analysis of specific tumour models and analysis of drug sensitivities and resistances may give insights into the future development of tailored chemotherapy.", "\n\nTemozolomide is clinically effective in malignant glioma ([@bib21]; [@bib23]) and this meta-analysis shows temozolomide also improves survival and reduces tumour volume in experimental glioma; whereas a substantial proportion of this evidence comes after the publication of the pivotal clinical trial this remains an example of concordance between experimental and clinical findings ([@bib13]; [@bib17]). ", "There was no significant change in reported efficacy for temozolomide after the publication of the phase III temozolomide randomised control trial, but there was evidence for a significant publication bias.", "\n\nThe nitrosoureas previously used in glioma such as carmustine (BCNU) and lomustine (CCNU) had limited efficacy ([@bib20]). ", "In a meta-analysis of nitrosoureas in experimental glioma efficacy was less than reported here for temozolomide with greater variability, and survival was not significantly increased ([@bib1]). ", "Similarly, gene therapy has been studied in glioma models, but has not translated successfully into clinical use. [", "@bib16]) concluded that clinical efficacy of gene therapy in several small RCTs was variable, and overall efficacy remains questionable. ", "Data from a meta-analysis of gene therapy in experimental glioma models (Conlin, under review) came to a similar conclusion; gene therapy does prolong survival in experimental glioma, but there was substantial variability, that is to say inconsistency of effect. ", "This contrasts with the findings reported here for temozolomide, and suggests that consistency of effect across a range of circumstances may be important for successful translation.", "\n\nIn conclusion, temozolomide is effective in experimental glioma models, although this is qualified by concerns of the internal validity (randomisation, blinding) and of publication bias. ", "Temozolomide appears to be more consistent in its efficacy than either nitrosoureas or gene therapy, and this observation may help guide future translational research in neuro-oncology.", "\n\nThe authors acknowledge financial support from MRC Trials Methodology Hub (KJE, MRM) CCBS Studentship (HMV), the MRC (IRW), CRUK (IRW) and the Melville Trust (IRW).", "\n\n[Supplementary Information](#sup1){ref-type=\"supplementary-material\"} accompanies this paper on British Journal of Cancer website (http://www.nature.com/bjc)\n\nThis work is published under the standard license to publish agreement. ", "After 12 months the work will become freely available and the license terms will switch to a Creative Commons Attribution-NonCommercial-Share Alike 3.0 Unported License.", "\n\nThe authors declare no conflict of interest.", "\n\nSupplementary Material {#sup1}\n======================\n\n###### \n\nClick here for additional data file.", "\n\n###### \n\nClick here for additional data file.", "\n\n###### \n\nClick here for additional data file.", "\n\n{#fig1}\n\n{#fig2}\n\n{#fig3}\n\n{ref-type=\"table\"}). ", "Plots represent mean±95% CI and diamond size a measure of number of comparisons within each stratum. ", "The solid line in **A** represents the level of neutral treatment effect.](bjc2012504f4){#fig4}\n\n{#fig5}\n\n{#fig6}\n\n###### Grouping of glioma models\n\n **Group** **Model**\n ----------- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------\n 9L 9L\n A172 A172\n C6 C6, C6-Fluc\n D54 D54, D54-Hif, D54MG\n DF-1 DF-1\n F98 F98\n GBM GBM6, GBM8, GBM10, GBM12, GBM12 TMZ resistant, GBM14, GBM14 TMZ resistant, GBM22, GBM22 TMZ resistant, GBM26, GBM34, GBM36, GBM39, GBM39 TMZ resistant, GBM43, GBM44, SJGBM2\n GL26 GL26\n GS GS2, GS22, GS28\n Hs683 Hs683\n RG2 RG2\n T98 T98\n U251 U251, U251SP\n U373 U373\n U87 U87, U87MG, U87-luc, U87-Fluc, U87MG.delta2-7\n" ]

{ "pile_set_name": "PubMed Central" }

[ 0, 0.006289308176100629, 0.005649717514124294, 0.009615384615384616, 0, 0.016666666666666666, 0.00411522633744856, 0.005813953488372093, 0.005934718100890208, 0.007874015748031496, 0.004975124378109453, 0.010582010582010581, 0.015957446808510637, 0, 0.015015015015015015, 0.013745704467353952, 0, 0, 0.017857142857142856, 0, 0.007518796992481203, 0.003952569169960474, 0.003215434083601286, 0.015873015873015872, 0.016216216216216217, 0, 0, 0, 0, 0.0223463687150838, 0, 0, 0, 0.0035087719298245615, 0, 0, 0, 0, 0, 0, 0.0078125, 0.0012515644555694619, 0.00684931506849315, 0.014925373134328358, 0.014084507042253521, 0.011428571428571429, 0, 0, 0, 0, 0, 0.007042253521126761, 0, 0.0036231884057971015, 0.014150943396226415, 0, 0, 0, 0.00904977375565611, 0, 0, 0, 0, 0.0029069767441860465, 0.004347826086956522, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.004081632653061225, 0.010526315789473684, 0.007142857142857143, 0.004106776180698152, 0.003976143141153081, 0.010135135135135136, 0, 0, 0, 0, 0.005208333333333333, 0.0034965034965034965, 0.005050505050505051, 0, 0, 0, 0, 0, 0, 0, 0.00641025641025641, 0, 0.02040816326530612, 0, 0, 0, 0, 0, 0, 0, 0.018018018018018018, 0.014218009478672985, 0, 0, 0.002898550724637681, 0, 0.011764705882352941, 0, 0.009828009828009828, 0.0048543689320388345, 0.016, 0.005154639175257732, 0, 0.0072992700729927005, 0.0038022813688212928, 0, 0, 0, 0.04216867469879518, 0.008583690987124463, 0.005917159763313609, 0, 0, 0, 0, 0, 0, 0.003257328990228013, 0, 0.0034129692832764505, 0, 0, 0.009009009009009009, 0.006211180124223602, 0, 0, 0, 0.007352941176470588, 0.005714285714285714, 0.009900990099009901, 0, 0.0196078431372549, 0.012578616352201259, 0.012658227848101266, 0, 0.006993006993006993, 0.00847457627118644, 0.00392156862745098, 0, 0, 0.009345794392523364, 0.006097560975609756, 0, 0, 0, 0, 0.00909090909090909, 0, 0, 0, 0.017957351290684626 ]

[ "import xmlrpc.client\n\nfrom .call import RPCCall\nfrom .method import RPCMethod\nfrom .result import RPCResult\n\n\nclass RPCCaller(object):\n def __init__(self, context):\n self.context = context\n self.calls = []\n self.available_methods = None\n\n def add(self, *args):\n if isinstance(args[0], RPCCall):\n call = args[0]\n elif isinstance(args[0], RPCMethod):\n call = RPCCall(args[0], *args[1:])\n elif isinstance(args[0], str):\n call = RPCCall(RPCMethod(args[0]), *args[1:])\n elif hasattr(args[0], '__self__'):\n call = args[0].__self__.rpc_call(args[0].__name__, *args[1:])\n else:\n raise RuntimeError(\"Unexpected args[0]: {0}\".format(args[0]))\n\n self.calls.append(call)\n return self\n\n def call(self):\n multi_call = xmlrpc.client.", "MultiCall(self.context.get_conn())\n for rpc_call in self.calls:\n method_name = self._get_method_name(rpc_call.get_method())\n rpc_call.do_pre_processing()\n getattr(multi_call, method_name)(*rpc_call.get_args())\n\n results = []\n for rpc_call, result in zip(self.calls, multi_call()):\n print(rpc_call.get_method().get_method_names())\n result = rpc_call.do_post_processing(result)\n results.append(result)\n\n return RPCResult(self.calls, results)\n\n def _get_method_name(self, rpc_method: RPCMethod):\n if self.available_methods is None:\n self.available_methods = self.context.get_available_rpc_methods()\n\n method_name = rpc_method.get_available_method_name(\n self.available_methods)\n\n if method_name is None:\n # TODO: Use a different Error subclass\n raise AttributeError(\"No matches found for {0}\".format(\n rpc_method.get_method_names()))\n\n return method_name\n" ]

{ "pile_set_name": "Github" }

[ 0.0011641443538998836, 0.003861003861003861 ]

[ "The best restaurants in the Camino de Santiago\n\nBehind every pilgrim a personal motivation to start their particular Camino de Santiago lies, and the truth is that there are few who find in enjoying the cuisine as rich and varied offering their routes. ", "Treat yourself to the stomach on the Road is a debt that the pilgrim must pay himself and opportunities you will not miss, choose the route you choose.", "\n\nThe Camino de Santiago passes through many regions and autonomous communities of Spain, without forgetting its way through Portugal, if we choose to do the Portuguese Way or the Pyrenean France, if we decided to choose the French Way. ", "Therefore the variety of dishes, desserts and wines is beyond question, the pilgrim can try some of the most popular specialties of each area restaurants, inns, taverns or brasseries.", "\n\nIt is widespread that that \"in Spain eat well anywhere,\" but if we talk about the Camino de Santiago, the expression is greater meaning. ", "It may be that many of its routes, such as the Camino Frances -the busiest of all-touring the northern peninsula from east to west across regions of ancient culinary tradition as Aragon, Navarra, La Rioja, Basque Country, Castile and Leon , Asturias, Cantabria and Galicia. ", "You should not miss on the table Pilgrim great wine, a good roast lamb, the best tapas or local seafood at the end of our trip.", "\n\nVive Camino approach we constantly expand a selection of some of the best restaurants you'll find on your journey through the Camino de Santiago, along with indications about their prices or offer menus. ", "You will find modern and sophisticated cuisine, but also the biggest gastronomic tradition and homemade touch that we cherish pilgrims on the plate at the end of each long day on the road." ]

{ "pile_set_name": "Pile-CC" }

[ 0.003952569169960474, 0, 0, 0, 0.007194244604316547, 0.014598540145985401, 0.007874015748031496, 0.0048543689320388345, 0 ]

[ "Geoffrey Walsh\n\nLieutenant-General Geoffrey Walsh, CBE, DSO, CD (19 August 1909 – 3 April 1999) was a Canadian soldier and Chief of the General Staff, the head of the Canadian Army from 1961 – 1964; Walsh was the last officer to hold this appointment as it was eliminated in 1964 as part of the reorganization of Canada's military in the lead-up to the 1968 unification of the Canadian Forces. ", "The most senior army appointment after unification, the Commander of Mobile Command, had a much-reduced scope of authority.", "\n\nMilitary career\nEducated at St Catherine's Collegiate School, Walsh was commissioned into the Royal Canadian Engineers in 1930.", "\n\nHe served in World War II with the Canadian Army Service Force and took part in the Spitsbergen Raid in 1941. ", "In 1942 he transferred to the 1st Canadian Division and fought in Sicily and Italy. ", "In 1944 he was made Commander Royal Engineers for 4th Canadian (Armoured) Division.", "\n\nAfter the war he was appointed Commander, Eastern Ontario Area, followed by Commander, 27th Brigade before his appointment as Director-General of Military Training. ", "In 1955 he was appointed Quartermaster-General of the Canadian Army and in 1959 he became General Officer Commanding, Western Command. ", "In 1961 he was made Chief of the General Staff and in 1964, became Vice Chief of the Defence Staff.", "\n\nFamily\nIn 1935, he married Gwynn Abigail Currie with whom he had one son.", "\n\nReferences\n\nCategory:1909 births\nCategory:1999 deaths\nCategory:Canadian generals\nCategory:Canadian Commanders of the Order of the British Empire\nCategory:Canadian Companions of the Distinguished Service Order\nCategory:Vice Chiefs of the Defence Staff (Canada)" ]

{ "pile_set_name": "Wikipedia (en)" }

[ 0.02030456852791878, 0.008130081300813009, 0.023255813953488372, 0.008928571428571428, 0.011904761904761904, 0.012048192771084338, 0.017964071856287425, 0.022222222222222223, 0, 0.013333333333333334, 0.0038314176245210726 ]

[ "Watch great Bundesliga soccer on the big screen, and enjoy a FREE breakfast, FREE round and FREE swag on Saturday, May 11 at Endeavor Brewing Company in Columbus. ", "Legendary German soccer team FC Schalke 04 takes on rivals Bayer Leverkusen at 9:30am.", "\n\n\n\n\n\nHosted by Schalke and fan club \"Six-One Schalke 04,\" this is a free event open to all ages. ", "Attendees receive a free German-style breakfast compliments of Schmidt’s Sausage Haus, a free Schalke scarf, and free round of Endeavor brew when Schalke scores (21+). ", "Also get ready for selfies with Schalke’s famous mascot, Erwin, who will be joining us from Germany.", "\n\n\n\n\n\nBreakfast will be served at 9:15am, with kick-off at 9:30am . ", "Fun for fans, friends and family. ", "We hope you can join us for a great morning!", "\n\n\n\n" ]

{ "pile_set_name": "OpenWebText2" }

[ 0.024539877300613498, 0.023255813953488372, 0, 0.023809523809523808, 0.02, 0, 0, 0, 0 ]

[ "---\nabstract: |\n We consider the problem of testing hypotheses on the copula density from $n$ bi-dimensional observations. ", "We wish to test the null hypothesis characterized by a parametric class against a composite nonparametric alternative. ", "Each density under the alternative is separated in the $L_2$-norm from any density lying in the null hypothesis. ", "The copula densities under consideration are supposed to belong to a range of Besov balls. ", "According to the minimax approach, the testing problem is solved in an adaptive framework: it leads to a $\\log\\log$ term loss in the minimax rate of testing in comparison with the non-adaptive case. ", "A smoothness-free test statistic that achieves the minimax rate is proposed. ", "The lower bound is also proved. ", "Besides, the empirical performance of the test procedure is demonstrated with both simulated and real data.\\\n\n [**[Index Terms]{}**]{} — Adaptation, Copula Density, Minimax Theory of Test, Goodness Test of Fit.\\\n [**[AMS Subject Classification]{}**]{} — 62G10, 62G20, 62G30.", "\nbibliography:\n- 'CEK.bib'\n---\n\n[****]{}\n\nGhislaine GAYRAUD $^*$ and Karine TRIBOULEY $^\\flat$\n\n$^*$ CREST and LMRS-UMR 6085\\\nUniversité de Rouen\\\nAvenue de l’Université, BP.12\\\n76801 Saint-Étienne-du-Rouvray, FRANCE\\\n\\\n\n$^\\flat$ LPMA and ModalX\\\nUniversite Paris 10\\\nBatiment G\\\n200 avenue de la République\\\n92 001 Nanterre Cedex, FRANCE\\\n\nIntroduction\n============\n\nCopulas became a very popular and attractive tool in the recent literature for modeling multivariate observations. ", "The nice feature of copulas is that they capture the structure dependence among the components of a multivariate observation without requiring the study of the univariate margins. ", "More precisely, Sklar’s Theorem ensures that any $d-$varied distribution function $H$ may be expressed as $$H(x^1,\\ldots,x^d)=C\\left(F^1(x^1),\\ldots,F^d(x^d)\\right),$$ where the $F^p$’s are the margins and $C$ is called the copula function. [", "@Sklar:1959] states the existence and the uniqueness of $C$ as soon as the random variables with joint law $H$ are continuous.", "\n\nModeling the dependence is a great challenge in statistics, specially in finance or assurance where (for instance) the identification of the dependence structure between assets is essential. ", "Many authors proposed parametrical families of copulas $\\{C_\\lambda,\\lambda\\in \\Lambda\\}$, each of them being available to capture different dependence behavior. ", "The elliptic family contains the Gaussian copulas and the Student copula which are often used in finance. ", "For insurance purposes, heavy tails are needed and copulas coming from the archimedian family are used. ", "Among others, the more common are the Gumbel copula, the Clayton copula or the Frank copula. ", "In view to illustrate the different behaviours of the tails of several copula densities, some graphs corresponding to the models cited above are presented below. ", "The parameters are chosen such a way that the associated Kendall’s tau (i.e. the indicator of concordance/discordance) is identical in all illustrations.", "\n\n![", "Kendall’s tau$=0.25$. Left: Bi-dimensional Gaussian copula density with parameter $\\rho=0.4$. Right: Bi-dimensional Student copula density with parameter $(\\rho,\\nu)=(0.4,1)$. []{data-label=\"fig1\"}](gaus04.eps \"fig:\"){width=\"4cm\" height=\"4cm\"} ![", "Kendall’s tau$=0.25$. Left: Bi-dimensional Gaussian copula density with parameter $\\rho=0.4$. Right: Bi-dimensional Student copula density with parameter $(\\rho,\\nu)=(0.4,1)$. []{data-label=\"fig1\"}](T04.eps \"fig:\"){width=\"4cm\" height=\"4cm\"}\n\n![", "Kendall’s tau$=0.25$. Left: Bi-dimensional Frank copula density with parameter $\\theta=2.5$. Center: Bi-dimensional Gumbel copula density with parameter $\\theta=1.33$. Right: Bidimensional Clayton copula density with parameter $\\theta=0.66$. []{data-label=\"fig2\"}](frank25.eps \"fig:\"){width=\"4cm\" height=\"4cm\"} ![", "Kendall’s tau$=0.25$. Left: Bi-dimensional Frank copula density with parameter $\\theta=2.5$. Center: Bi-dimensional Gumbel copula density with parameter $\\theta=1.33$. Right: Bidimensional Clayton copula density with parameter $\\theta=0.66$. []{data-label=\"fig2\"}](gumbel133.eps \"fig:\"){width=\"4cm\" height=\"4cm\"} ![", "Kendall’s tau$=0.25$. Left: Bi-dimensional Frank copula density with parameter $\\theta=2.5$. Center: Bi-dimensional Gumbel copula density with parameter $\\theta=1.33$. Right: Bidimensional Clayton copula density with parameter $\\theta=0.66$. []{data-label=\"fig2\"}](clayton066.eps \"fig:\"){width=\"4cm\" height=\"4cm\"}\n\nSince many parametric copula models are now available, the crucial choice for the practitioner is to identify the model which is well-adapted to data at hand. ", "Many goodness-of-fit tests are proposed in the literature. [", "@Genest/Remillard/Beaudoin:2008] give an excellent review and propose a detailed empirical study for different tests: we refer to this paper for any supplementary references. ", "Roughly speaking, they study procedures based on empirical processes. ", "Among others, they deal with rank-based versions of the Cramér-von-Mises and Kolmogorov-Smirnov statistics. ", "They also consider test based on Kendall’s transform. ", "Basically, they restrict themselves to test statistics built from empirical distributions (empirical copula or transform of this latter). ", "On a theoretical point of view, the asymptotic law under the null of the test statistic is stated in a number of papers (see by instance [@Deheuvels:1979], [@Deheuvels:1981a] and [@Deheuvels:1981b]). ", "It allows in particular to derive the critical value but generally the alternative is unspecified and the properties on the power are empirically given from simulations.", "\n\nIn our paper, it is supposed that the copula $C$ admits a density copula $c$ with respect to the Lebesgue measure. ", "To our knowledge, [@Fermanian:2005] was the first author to propose a goodness-of-fit test based on nonparametric kernel estimations of the density copula. ", "In the same spirit as the papers cited above, he derived the asymptotic law of the test statistic under the null. ", "His results are valid for bandwidths greater than $n^{-2/(8+d)}$ which correspond to enough smooth copula densities.", "\n\nHere, we focus on the minimax theory framework: we define the test problem as initiated by [@Ingster:1982]. ", "One of the advantages of this point of view is to precisely define the alternative: it is then possible to quantify the risk associated with the test problem as the sum of the first type error and the second type of error. ", "Since this risk measure provides a quality criterion, it is then possible to compare the test procedures. ", "Indeed, the alternative $H_1(v_n)$ is defined from a positive quantity $v_n$ measuring the distance between the null and the latter. ", "Obviously, the larger is this separating distance, the easier is the decision. ", "The aim of the minimax theory is to determine the larger alternative for which the decision remains feasible. ", "Solving [**the lower bound problem**]{} is equivalent to exhibit the faster separating rate $v_n$ such that the risk is bounded from below by a given positive constant $\\alpha$: this rate is called [**the minimax rate of testing**]{}. ", "Next, [**the upper bound problem**]{} has to be solved exhibiting a test procedure whose risk is bounded from above by a given $\\alpha$, that is, the statistic test allows to distinguish the null from $H_1(v_n)$, where $v_n$ is the minimax rate.", "\n\nIn the white noise model or in the density model, the goodness-of-fit problem (stands as explained above) was solved for different regularity classes (Hölder or Sobolev or Besov) associated with various geometries: pointwise, quadratic and supremum norm. ", "For fixed smoothness of the unknown density ([**minimax context**]{}), there is a rich literature summed-up in [@Ingster:1993] and in [@Ingster/Suslina:2002]. ", "Optimal test procedures include orthogonal projections, kernel estimates or $\\chi^2$ procedures. ", "Goodness-of-fit tests with alternatives of variable smoothness into some given interval ([**adaptive context**]{}) were introduced by [@Spokoiny:1996] for the $L_2$-norm in the Gaussian white noise model and generalized by [@Spokoiny:1998] to $L_p$-norms. [", "@Ingster:2000] proved that a collection of $\\chi^2$ tests attains the adaptive rates of goodness-of-fit tests in $L_2$-norm as well as for the density model.", "\n\nFor sake of simplicity, we restrict ourselves to bi-dimensional data but there is no theoretical obstacle to generalize our results to higher dimensions. ", "Suppose that we observe $n$ i.i.d. ", "copies $(X_i,Y_i)_{i\\in {\\cal I}}$ where ${\\cal I}=\\{ 1,\\ldots,n\\}$ of $(X,Y)$. The random vector $(X,Y)$ is drawn from the distribution function $H$ expressed through the copula $C$. ", "Moreover, it is assumed that $C$ has a copula density $c$ with respect to the Lebesgue measure and $F$ and $G$ stand for the cdf’s of $X$ and $Y$ respectively. ", "From $(X_i,Y_i)_{i\\in {\\cal I}}$, we are interested in studying the goodness-of-fit problem when the null is a composite hypothesis $H_0:c\\in {\\cal C}_{\\Lambda}$ for a general class ${\\cal C}_{\\Lambda}$ of parametrical copula densities. ", "Since the alternative is defined from the quadratic distance, we propose a goodness-of-fit test based on wavelet estimation of an integrated functional of the copula density. ", "Indeed, [@Genest/Masiello/Tribouley:2008] and [@Autin/LePennec/Tribouley:2008] show that the wavelet methods are an efficient tool to estimate the copula densities since these latter have very specifics behaviors. ", "Unfortunately no direct observations $(F(X_i),G(Y_i))$ for $i \\in {\\cal I}$ are available since $F$ and $G$ are unknown, the test statistic is then built with pseudo-observations $(\\widehat{F}(X_i),\\widehat{G}(Y_i))_{i\\in {\\cal\nI}}$: as usual in the copula context, the quantities of interest are rank-based statistics. ", "We provide an auto-driven test procedure and we produce its rate when the alternative contains a regular constraint: since the procedure is based on wavelet methods, the linked functional classes are the Besov classes $B_{s,p,q}$. We give results for $p\\geq 2$ ([**dense case**]{}) and $s\\geq 1/2$. The constraint $s\\geq 1/2$ is due to the fact that pseudo-data are used and then a minimal regularity is required in order to pay no attention to substitute the direct data with the ranked data. ", "Observe that [@Kerkyacharian/Picard:2004] have the same constraint in the univariate regression model when the design is random with unknown distribution. ", "Next, we prove that our procedure is minimax (and adaptive) optimal by exhibiting the minimax adaptive rate. ", "This one looks like the minimax rate but an extra $\\log\\log$ term appears: we prove that this loss is the price to paid for adaptivity. ", "To our knowledge, the proof of the adaptive lower bound in the multivariate density model when the null is composite has never been clearly written.", "\n\nNext, we allocate a part to empirical studies. ", "Simulation allows us to show that, when the theoretical framework is respected, the power qualities of our test procedures are good. ", "We choose to make simulations starting from the parametrical copula families presented at the beginning of the introduction and which are the more common for applications. ", "We compare our simulation results with those of [@Genest/Remillard/Beaudoin:2008]. ", "Then, we study a very well known sample of real life data of [@Frees/Valdez:1998] consisting of the indemnity payment (LOSS) and the allocated loss adjustment expense (ALAE) for 1500 general liability claims. ", "The most popular model for the copula is a Gumbel copula model with parameter $\\theta=1.45$ (which may be estimated by inverting the Kendall’s tau) given in Figure \\[fig3\\]. ", "Among other results, it is empirically shown that the Gumbel and the Gaussian copula models are acceptable while Student, Clayton or Frank models are rejected. ", "Figure \\[fig3\\] gives a wavelet estimator of the copula density of $(LOSS,ALEA)$ by the method explained in [@Autin/LePennec/Tribouley:2008]. ", "Visually, fitting the unknown copula with the Gumbel model seems indeed to be the most appropriated.", "\n\n![", "Left: Thresholded wavelet estimator for the copula density of $(LOSS,ALEA)$ as given in [@Autin/LePennec/Tribouley:2008]. ", "Center: Gumbel copula density with parameter $\\theta=1.45$. Right: Gaussien copula density with parameter $\\rho=0.48$. []{data-label=\"fig3\"}](frees.eps \"fig:\"){width=\"4cm\" height=\"4cm\"} ![", "Left: Thresholded wavelet estimator for the copula density of $(LOSS,ALEA)$ as given in [@Autin/LePennec/Tribouley:2008]. ", "Center: Gumbel copula density with parameter $\\theta=1.45$. Right: Gaussien copula density with parameter $\\rho=0.48$. []{data-label=\"fig3\"}](gumbel145.eps \"fig:\"){width=\"4cm\" height=\"4cm\"} ![", "Left: Thresholded wavelet estimator for the copula density of $(LOSS,ALEA)$ as given in [@Autin/LePennec/Tribouley:2008]. ", "Center: Gumbel copula density with parameter $\\theta=1.45$. Right: Gaussien copula density with parameter $\\rho=0.48$. []{data-label=\"fig3\"}](gaus048.eps \"fig:\"){width=\"4cm\" height=\"4cm\"}\n\nThe paper is organized as follows. ", "In Section 2, we first provide a general description of orthonormal wavelet bases, focusing on the mathematical properties that are essential to the construction of the statistics that we consider. ", "In Section 3, we provide the inference procedures: first, we explain how to estimate the square $L_2$-norm of the copula density and next we derive the procedure of goodness-of-fit. ", "The theoretical part is exposed in Section 4: first, we state very precisely the test problem under consideration; we define the criterion allowing to measure the quality of test procedures and define the separating minimax rate. ", "In Section 5, the main results are stated: our test procedure is shown to be optimal in the sense defined in the previous section. ", "Section 6 is devoted to practical results with both simulated and real data. ", "We conclude these parts with a discussion in Section 7. ", "The proof of the upper bound is given in Section 8 while the proof of the lower bound is given in Section 9. ", "Finally, all technical or computational lemmas which are not essential to understand the main proofs, are postponed in appendices.", "\n\nWavelet Setting\n===============\n\nWavelet expansion\n-----------------\n\nIn the univariate case, we consider a wavelet basis of $L_2([0,1])$ (see [@Cohen/Daubechies/Vial:1993]). ", "Let $\\phi$ be the scaling function and let $\\psi$ be the same notation for the associated wavelet function and its usual modifications near the frontiers $0$ and $1$. They are chosen compactly supported on $[0,L]$, $L>0$. Let $j$ in $\\BBn$, $k_1$ in $\\BBz$ and for any univariate function $\\Phi$, set $\\Phi_{j,k_1}(\\cdot)=2^{j/2}\\Phi(2^j\\cdot-k_1)$. In the sequel, we use wavelet expansions for bivariate functions and we keep the same notation as for the univariate case. ", "Then, a bivariate wavelet basis is built as follows: $$\\begin{aligned}\n\\phi_{j,k}(x,y)=\\phi_{j,k_1}(x)\\phi_{j,k_2}(y),&\\quad&\n\\psi_{j,k}^{(1)}(x,y)=\\phi_{j,k_1}(x)\\psi_{j,k_2}(y), \\\\\n\\psi_{j,k}^{(2)}(x,y)=\\psi_{j,k_1}(x)\\phi_{j,k_2}(y),&\\quad&\n\\psi_{j,k}^{(3)}(x,y)=\\psi_{j,k_1}(x)\\psi_{j,k_2}(y),\\end{aligned}$$ where the subscript $k=(k_1,k_2)$ indicates the number of components of the functions $\\phi_{j,k}$ and $\\psi_{j,k}$. For a given $j\\in\n\\BBn$, the set $$\\{\\phi _{j,k},\\psi _{\\ell,k^\\prime}^\\epsilon, \\;\n\\ell\\geq j,(k,k^\\prime)\\in \\BBz^{2}\\times \\BBz^2,\\epsilon =1,2,3\\}$$ is an orthonormal basis of $L^2([0,1]^2)$ and the expansion of any real bivariate function $\\Phi$ in $L^2([0,1]^2)$ is given by: $$\\Phi(x,y)=\\sum_{k\\in \\BBz^2}A _{j,k}\\phi _{j,k}(x,y)+\\sum_{\\ell=\nj}^\\infty\\sum_{k\\in \\BBz^2}\\sum_{ \\epsilon =1,2,3}B\n_{\\ell,k}^\\epsilon \\psi _{\\ell,k}^\\epsilon(x,y),$$ where the scaling coefficients and the wavelet coefficients are $$\\forall j\\in \\BBn,\\forall k\\in \\BBz^2,\\quad\nA_{j,k}=\\int_{[0,1]^2}\\Phi \\phi_{j,k},\\qquad\nB_{j,k}^\\epsilon=\\int_{[0,1]^2}\\Phi \\psi_{j,k}^\\epsilon.$$ The Parseval Equality immediately leads to the expansion of the square $L_2$-norm of the function $\\Phi$: $$\\begin{aligned}\n\\label{decomposition}\n\\int \\Phi^2&=&T_{j}+B_{j},\\end{aligned}$$ where the trend and the detail terms are respectively: $$\\begin{aligned}\n\\label{TB}\nT_{j}=\\sum_{k\\in \\BBz^2}\\left( A_{j,k}\\right)^2 \\mbox{ {\\rm and} }\nB_{j}=\\sum_{\\ell=j}^\\infty\\;\\sum_{k\\in \\BBz^2}\\;\\sum_{\\epsilon=1}^3(\nB_{\\ell,k}^\\epsilon)^2.\\end{aligned}$$ Notice that, since the support of $\\Phi$ is $[0,1]^2$, the sum over the indices $k$ is finite: there are no more than $(2^j+L)^2$ terms in the sum (recall that $L$ is the length of the support of $\\phi$). ", "In order to simplify the notations, the bounds of variation of $k$ and $\\epsilon$ in expansion of any $\\Phi$, are omitted in the sequel.", "\n\nBesov Bodies and Besov spaces\n-----------------------------\n\nDealing with wavelet expansions, it is natural to consider Besov bodies as functional spaces since they are characterized in term of wavelet coefficients as follows.", "\n\n\\[BesovBody\\] For any $s>0, \\; p\\geq 1$ and any radius $M>0$, a $d-$varied function $\\Phi$ belongs to the ball $b_{s,p,\\infty}(M)$ of the Besov body $b_{s,p,\\infty}$ if and only if its sequence of wavelet coefficients $B_{j,k}^\\epsilon$ satisfies $$\\begin{aligned}\n\\forall j\\in \\BBn,\\;\\sum_{k\\in\n\\BBz^2}\\sum_{\\epsilon=1}^3|B_{j,k}^\\epsilon|^p <M2^{-j(s+d/2-d/p)p}.\\end{aligned}$$\n\nThe Besov body $b_{s,p,\\infty}$ coincides with the more standard Besov space $\\mathcal{B}_{s,p,\\infty}$ when there exists an integer $N$ strictly larger than $s$ and such that the $q-$th moment of the wavelet $\\psi$ vanishes for any $q=0,\\ldots, N-1$. It is possible to build univariate wavelets whose support is included in $[0, 2N -1]$ satisfying this property for any choice of $N$ (see the Daubechies wavelets).", "\n\nIn the sequel, we need to bound the detail term $B_j$ defined in (\\[TB\\]). ", "We use the following inequality $$\\begin{aligned}\n\\forall j\\in \\BBn,\\quad B_j\\leq\n\\sum_{\\ell=j}^\\infty\\;\\left(\\sum_{k\\in\n{\\ensuremath{\\mathbb{Z}}}^2}\\;\\sum_{\\epsilon=1}^3\\left|B_{\\ell,k}^\\epsilon\\right|^p\\right)^{2/p}\\left(K\\,2^{2j}\\right)^{1-2/p},\\end{aligned}$$ where $K$ is a positive constant depending on the supports of $\\Phi$ and $\\psi$. Assuming that the function $\\Phi$ belongs to $b_{s,p,\\infty}(M)$ with $s,p$ and $M$ as in Definition \\[BesovBody\\], the following inequality holds $$\\begin{aligned}\n\\label{reg}\n\\forall j\\in \\BBn, \\quad B_j&\\leq& {\\tilde K}\\;2^{-2js},\\end{aligned}$$ where ${\\tilde K}$ is a positive constant depending on the supports of $\\Phi$, $\\psi$ and on the radius $M$. When $\\Phi$ is a copula density, ${\\tilde K}=M^{2/p}\\left( 3(L+1)^2 \\right)^{1-2/p}$.\n\nStatistical Procedures {#procedure}\n======================\n\nAssuming that the copula density $c$ belongs to $L_2([0,1]^2)$, we first explain the procedure to estimate the square $L_2-$norm of $c$ $$\\theta=\\,\\|c\\|^2:=\\,\\int_{[0,1]^2} c^2,$$ which is used to define the alternative of the goodness-of-fit test. ", "The statistical methods depend on parameters (the level $j$ for the estimation procedure and $j$ and the critical value $t_j$ for the test procedure) which are discussed and determined in an optimal way in Section \\[mainR\\].", "\n\nIt is fundamental to Notice that, for any bivariate function $\\Phi$, one has $$\\begin{aligned}\n\\label{passage_c_h}\n\\E_c\\left[\\Phi(U,V)\\right]&= &\\E_h\\left[\\Phi(F(X),G(Y))\\right],\\end{aligned}$$ where $h$ stands for the joint density of $(X,Y)$. This means in particular that the wavelet coefficients $\n\\{c_{j,k},c_{\\ell,k}^\\epsilon, \\; \\ell\\geq j,k\\in\n\\BBz^2,\\epsilon=1,2,3\\} $ of the copula density $c$ on the wavelet basis $$\\{\\phi _{j,k},\\psi_{\\ell,k}^\\epsilon, \\; \\ell\\geq j,k\\in\n\\BBz^2,\\epsilon=1,2,3\\}$$ are equal to the coefficients of the joint density $h$ on the warped wavelet family $$\\{\\phi_{j,k}(F(\\cdot), G(\\cdot)),\\psi\n_{\\ell,k}^\\epsilon(F(\\cdot), G(\\cdot)), \\; \\ell\\geq j,k\\in\n\\BBz^2,\\epsilon =1,2,3\\}.$$ The statistical procedures are based on the wavelet expansion of the copula density $c$, for which the wavelet coefficients have to be estimated.", "\n\nProcedures to estimate $\\theta$ {#ESTprocedure}\n-------------------------------\n\nLet $J$ be a subset of $\\BBn$ and consider a given $j$ in $J$. Motivated by the wavelet expansion (\\[decomposition\\]), we propose to estimate $\\theta$ with an estimator of the trend $T_{j}$ omitting the detail term $B_j$. Using the orthonormality property of the wavelet basis, it leads to estimate the square of the coefficients of the copula density on the scaling function. ", "As usual, a $U-$statistic associated with the empirical coefficients is used in order to remove the bias terms. ", "Due to (\\[passage\\_c\\_h\\]), we first consider the following family of statistics $\\{\\widehat{\nT_j},j\\in J\\}$ defined by $$\\begin{aligned}\n \\label{That}\n\\widehat{ T_j}&=&\\sum_{k} \\widehat{ \\theta_{j,k}},\\end{aligned}$$ where $\\widehat{ \\theta_{j,k}}$ is the following $U-$statistic $$\\begin{aligned}\n \\widehat{ \\theta_{j,k}} &=&\n\\frac{1}{n(n-1)} \\mathop{\\sum^n_{i_1, i_2 =1}}_{ i_1 \\neq i_2\n}\\phi_{j,k}\\left(F(X_{i_1}), G(Y_{i_1})\\right) \\phi_{j,k}\\left(\nF(X_{i_2}), G(Y_{i_2})\\right).\\end{aligned}$$ Since no direct observation $(F(X_i),G(Y_i))$ is usually available, it is replaced in $\\widehat{ \\theta_{j,k}}$ by the pseudo observation $(\\widehat {F}(X_i), \\widehat {G}(Y_i))$, where $\\widehat {F}, \\widehat {G}$ denote some estimator of the margins. ", "To preserve the independence given by the observations, we split the initial sample $(X_i,Y_i)_{i \\in {\\cal I}}$ into disjoint samples $(X_i,Y_i)_{i \\in {\\cal I}_1}$ and $(X_i,Y_i)_{i \\in {\\cal I}_2}$ with ${\\cal I}_2 \\cup {\\cal I}_1={\\cal I}, \\; {\\cal I}_2 \\cap {\\cal\nI}_1=\\emptyset$, and whose size is $n_1$ and $n_2$ respectively. ", "The sub-sample with indices in ${\\cal I}_1$ is used to estimate the marginal distributions and the second one with indices in ${\\cal\nI}_2$ is devoted to the computation of the $U$-statistic. ", "We consider the usual empirical distribution functions: $$\\begin{aligned}\n\\widehat{F}(x)=\\frac{1}{n_1} \\mathop{\\sum_{i \\in {\\cal I}_1}} \\I_{\\{X_i \\leq x\\}}&\\mbox{\nand }&\n\\widehat{G}(y)=\\frac{1}{n_1}\\mathop{\\sum_{i \\in {\\cal I}_1}}\n\\I_{\\{Y_i \\leq y\\}} .\\end{aligned}$$ It leads to the family $\\{\\widetilde{T_j},j\\in J\\}$ of estimators of $\\theta$ $$\\begin{aligned}\n\\label{Ttilde}\n\\widetilde{T_{j}}&=&\\sum_{k}\\widetilde{\\theta_{j,k}},\\end{aligned}$$ with $$\\begin{aligned}\n\\widetilde{\\theta_{j,k}}&=&\\frac{1}{n_2(n_2-1)}\\mathop{\\sum_{i_1,\ni_2\\in {\\cal I}_2}}_{ i_1 \\neq i_2 } \\phi_{j,k}\\left(\\frac{R_{i_1}}{n_1}\n,\\frac{S_{i_1}}{n_1} \\right) \\phi_{j,k}\\left(\\frac{R_{i_2}}{n_1},\n\\frac{S_{i_2}}{n_1} \\right),\\nonumber\\end{aligned}$$ where $R_p=n_1 \\widehat{F}(X_p)$ and $S_p = n_1 \\widehat{G}(Y_p)$, $p \\in {\\cal I}_1$, could be viewed as estimates of the rank statistics of $X_p$ and $Y_p$ respectively.", "\n\nTest Procedures {#TESTprocedure}\n---------------\n\nIn this part, we consider a family of known bivariate copula densities ${\\cal\nC}_\\Lambda=\\left\\{c_\\lambda,\\lambda\\in\\Lambda\\right\\}$ indexed by a parameter $\\lambda$ varying in a given set $\\Lambda \\subset\n\\BBr^{d_{\\Lambda}}$, $d_{\\Lambda} \\in \\BBn^*$. From the observations $(X_i,Y_i)_{i \\in {\\cal I}}$, our aim is to test the goodness-of-fit between any $c_\\lambda$ and a copula density $c$, which is [*enough distant*]{} in the $L_2$-norm, from the parametric family ${\\cal\nC}_\\Lambda$. Acting as in paragraph \\[ESTprocedure\\], we estimate the square $L_2$-norm between $c$ and a fixed element $c_\\lambda$ lying in the family ${\\cal C}_\\Lambda$ by $$\\begin{aligned}\n\\label{Ttilde}\n\\widetilde {T_j}(\\lambda)&=&\\sum_{k}\\widetilde\n{\\theta_{j,k}}(\\lambda),\\end{aligned}$$ for $$\\begin{aligned}\n\\widetilde\n{\\theta_{j,k}}(\\lambda)&=&\\frac{1}{n_2(n_2-1)}\\mathop{\\sum_{i_1,\ni_2\\in {\\cal I}_2}}_{ i_1 \\neq i_2 } \\left(\\phi_{j,k}\\left(\\frac{R_{i_1}}{n_1}\n,\\frac{S_{i_1}}{n_1} \\right)- c_{j,k}(\\lambda)\n\\right)\\\\&&\\hspace{3cm} \\times\n\\left(\\phi_{j,k}\\left(\\frac{R_{i_2}}{n_1}, \\frac{S_{i_2}}{n_1}\n\\right)- c_{j,k}(\\lambda) \\right)\\nonumber,\\end{aligned}$$ where $ \\{c_{j,k}(\\lambda),k\\in\\BBz^2,j\\in \\BBn\\}$ denote the known scaling coefficients of the target copula density $c_\\lambda$. Notice that, if direct observations $(F(X_i),G(Y_i))_{i \\in\n{\\cal I}}$ would be available, the appropriate test statistic $\\widehat{\nT}_j (\\lambda)$ would be $$\\begin{aligned}\n \\label{ThatLambda}\n\\widehat{ T}_j (\\lambda)& =& \\sum_k \\widehat{\\theta}_{j,k}(\\lambda),\\end{aligned}$$ where $$\\begin{aligned}\n\\widehat\n{\\theta_{j,k}}(\\lambda)&=&\\frac{1}{n_2(n_2-1)}\\mathop{\\sum_{i_1,\ni_2\\in {\\cal I}_2}}_{ i_1 \\neq i_2 }\n\\left(\\phi_{j,k}\\left(F(X_{i_1}),G(Y_{i_1}) \\right)-\nc_{j,k}(\\lambda) \\right)\\\\&&\\hspace{3cm} \\times\n\\left(\\phi_{j,k}\\left(F(X_{i_2}),G(Y_{i_2}) \\right)-\nc_{j,k}(\\lambda) \\right)\\nonumber.\\end{aligned}$$ Now we are ready to build the test procedures. ", "Let us give a set of indices $J$ and a set of critical values $\\{t_j, \\; j \\in\nJ\\}$ and define $\\{D_j^\\Lambda,j\\in J\\}$, the family of test statistics $$\\begin{aligned}\n\\label{testopti}\nD_j^\\Lambda&=& \\I_{ \\displaystyle{\\inf_{\\lambda \\in \\Lambda}}\n\\widetilde {T_j}(\\lambda) > t_j} \\quad ,\\end{aligned}$$ allowing to test if $c$ belongs to the parametric family ${\\cal\nC}_\\Lambda=\\{c_\\lambda,\\lambda\\in\\Lambda\\}$. Notice that $\\Lambda=\\{\\lambda_0\\}$ leads to the single null hypothesis $H_0:c=c_{\\lambda_0}$. We are also interested in building auto-driven procedures by considering all the tests in the family $$\\begin{aligned}\nD_\\Lambda&=& \\max_{j\\in J}D_j^\\Lambda= \\I_{ \\displaystyle{ \\max_{j\\in J}\n(\\inf_{\\lambda \\in \\Lambda} \\widetilde {T_j}(\\lambda)- t_j )}>\n0}\\;. ", "\\label{testAdap}\\end{aligned}$$ The sequence of parameters $t_j$ of the method are determined in an optimal way in Section 5. ", "We explain in Section 4 what “optimal way” means in giving a presentation of the minimax theory for our framework.", "\n\nMinimax Theory\n==============\n\nWe adopt the minimax point of view to solve the problem of hypothesis testing, initiated by [@Ingster:1982] in Gaussian white noise. ", "A review of results obtained in problems of minimax hypothesis testing is available in [@Ingster:1993] and [@Ingster/Suslina:2002]. ", "Let us describe this approach.", "\n\nMinimax hypothesis testing Problem\n----------------------------------\n\nAs in the previous section, we consider ${\\mathcal\nC}_{\\Lambda}=\\{ c_\\lambda, \\lambda \\in \\Lambda\\}$ a given functional class of copula densities. ", "For any given $\\tau=(s,p,M)$, with $s>0,p\\geq 1,M>0$, the following statistical problem of hypothesis testing is considered,\n\n$$\\begin{aligned}\nH_0 : c=c_\\lambda \\in {\\cal C}_{\\Lambda}\\quad &\\mbox{ {\\rm against}\n}&\\quad H_1 : c \\in \\Gamma(v_n(\\tau)),\\label{pbtest}\\end{aligned}$$\n\nwith $$\\begin{aligned}\n\\Gamma(v_n(\\tau))= b_{s,p,\\infty}(M)\\cap \\left\\{c : \\inf_{c_\\lambda\n\\in {\\cal C}_{\\Lambda}}\\|c -c_\\lambda\\| \\geq v_n(\\tau) \\right\\},\\end{aligned}$$ where $b_{s,p,\\infty}(M)$ is the ball of radius $M$ of the Besov body $b_{s,p,\\infty}$ defined in Definition \\[BesovBody\\] and $v_n(\\tau)$ is a sequence of positive numbers, depending on $\\tau$ and decreasing to zero as $n$ goes to infinity. ", "Recall that $\\|g\\|$ denotes the $L_2$-norm of any function $g $ in $L_2([0,1]^2)$. Observe that the functional class $\\Gamma(v_n(\\tau))$, which determines the alternative $H_1$, is characterized by three parameters: the regularity class $b_{s,p,\\infty}$ where the copula density is supposed to belong, the $L_2$-norm which is the geometrical tool measuring the distance between both hypotheses, and the sequence $v_n(\\tau)$.\n\nAccording to the principle of the minimaxity, the regularity space and the loss function are chosen by the statistician. ", "Notice that the parameter $\\tau$ could be known or unknown. ", "Obviously, our aim is to consider tests which are able to detect alternatives defined with sequences $v_n(\\tau)$ as small as possible. ", "It can be shown ([@Ingster:1993]) that $v_n(\\tau)$ cannot be chosen in an arbitrary way: indeed, if $v_n(\\tau)$ is too small, then $H_0$ and $H_1$ cannot be distinguished with a given error $\\alpha \\in (0,1)$. Therefore, solving hypothesis testing problems via the minimax approach consists in determining the smallest sequence $v_n(\\tau)$ for which such a test is still possible and to indicate the corresponding test functions. ", "The smallest sequence $v_n(\\tau)$ is called the minimax rate of testing. ", "Let $D_n$ be a [*test statistic*]{} i.e. an arbitrary function with possible values $0,1$, measurable with respect to $(X_i,Y_i)_{i \\in {\\cal I}}$ and such that we accept $H_0$ if $D _n=0 $ and we reject it if $D_n=1$.\n\nAssuming $\\tau$ to be known, the sequence $v_n(\\tau)$ is the minimax rate of testing $H_0$ versus $H_1$ if relations (\\[Bi\\]) and (\\[BS1\\]) are fulfilled:\n\n- for any given $\\alpha_1 \\in (0,1)$, there exists $a>0$ such that $$\\begin{aligned}\n && \\lim_{n\\rightarrow +\\infty}\\inf_{{D}_n} \\left(\n \\sup_{c_\\lambda \\in {\\cal C}_{\\Lambda}} \\P_{\\lambda} ({D}_n=1)\n +\\sup_{c \\in \\Gamma(a\\;v_n(\\tau))} \\P_c ({D}_n=0) \\right) \\geq \\alpha_1 , \\!\\!\\!\\! ", "\\label{Bi}\\end{aligned}$$ where the infimum is taken over any test statistic $D_n$,\n\n- there exists a sequence of test statistics $(D_n^\\star)_n$ for which for any given $\\alpha_2 $ in $(0,1)$, it exists $A>0$ such that $$\\begin{aligned}\n &&\n \\lim_{n\\rightarrow +\\infty} \\left(\\sup_{c_\\lambda \\in {\\cal C}_{\\Lambda}} \\P_{\\lambda} (D_n^\\star=1)+\\sup_{c \\in \\Gamma(A\\;v_n(\\tau))} \\P_c\n (D_n^\\star=0) \\right)\\leq \\alpha_2,\n \\label{BS1}\\end{aligned}$$\n\nwhere $\\P_c$, respectively $\\P_{\\lambda}$ denotes the distribution function associated with the copula density $c$, respectively with $c_\\lambda$.\n\nAdaptation\n----------\n\nNevertheless, since the copula function itself is unknown, the a priori knowledge on $\\tau$ could appear unrealistic. ", "Therefore, the purpose of this paper is to solve the previous problem of test in an adaptive framework i.e. in supposing that $\\tau=(s,p,M)$ is unknown but varying in a known set ${\\cal S}$. Comparing the adaptive case with the non-adaptive case, it has been proved in different frameworks that a loss of efficiency in the rate of testing is unavoidable (see for instance [@Spokoiny:1996], [@Gayraud/Pouet:2005]). ", "This loss is expressed as $t_n$, a positive constant or a sequence of positive numbers increasing to infinity with $n$ (as slow as possible), which appears in the rate of testing $v_{nt_n^{-1}}(\\tau)$. Similarly to the minimax rate of testing, we define the adaptive minimax rate of testing as follows.", "\n\nThe sequence $v_{nt_n^{-1}}(\\tau)$ is the adaptive minimax rate of testing if relations (\\[borneInf\\]) and (\\[borneSup\\]) are satisfied\n\n- for any given $\\alpha_1 \\in (0,1)$, there exists $ a>0$ such that $$\\begin{aligned}\n &&\n \\lim_{n\\rightarrow +\\infty}\\inf_{{D}_n} \\left(\n \\sup_{c_\\lambda \\in {\\cal C}_{\\Lambda}} \\P_{\\lambda} ({D}_n=1) +\n \\sup_{\\tau \\in {\\cal S}} \\sup_{c \\in \\Gamma(a\\;\n v_{nt_n^{-1}}(\\tau))} \\P_c ( { D}_n =0) \\right)\\geq \\alpha_1 ,\n \\hspace{1cm} \\label{borneInf}\\end{aligned}$$ where the infimum is taken over any test statistic $D_n$,\n\n- there exists a sequence of universal test statistics $D_n^{\\star}$ (free of $\\tau$) such that, for any given $\\alpha_2$ in $(0,1)$, there exists $A>0$ such that $$\\begin{aligned}\n &&\n \\lim_{n\\rightarrow +\\infty} \\left(\n \\sup_{c_\\lambda \\in {\\cal C}_{\\Lambda}} \\P_{\\lambda} (D_n^{\\star}=1)+\\sup_{\\tau \\in {\\cal\n S} } \\sup_{c \\in \\Gamma(A\\;v_{nt_n^{-1}}(\\tau))} \\P_c (\n D_n^{\\star} =0)\\right) \\leq \\alpha_2 \\hspace{1cm} \\label{borneSup}\\end{aligned}$$ where $t_n$ is either a positive constant or a sequence of positive numbers increasing to infinity with $n$ as slow as possible.", "\n\nNotice that relations (\\[borneInf\\]) and (\\[borneSup\\]) (instead of relations (\\[Bi\\]) and (\\[BS1\\])) mean that the minimax rate of testing $v_n(\\tau)$ is contaminated by the term $t_n$ in the adaptive setting. ", "Observe that the same phenomenon is observed in the estimation problem where an extra logarithm term ${\\tilde t}_n=\\log(n)$ has often (but not always) to be paid for the adaptation.", "\n\nMain results {#mainR}\n============\n\nIn this section, we focus on test problems for which the parametric family ${\\cal C}_\\Lambda$ is included in some $b_{s_\\Lambda,p_\\Lambda,\\infty}(M_\\Lambda)$ where $s_\\Lambda>0$, $p_\\Lambda\\geq 1$ and $M_\\Lambda >0$ are known.", "\n\nOur theoretical results concern the minimax resolution of the problem of hypothesis testing defined in (\\[pbtest\\]) in an adaptive framework. ", "Theorem \\[BInf\\] states the result of the lower bound (see relation (\\[borneInf\\])). ", "Then, Theorem \\[BSup\\] exhibits the rate achieved by the test procedure proposed in Section 3 (see relation (\\[borneSup\\])). ", "Comparing the rate of our procedure with the fastest rate given in Theorem \\[BInf\\] leads to Theorem \\[opt\\] establishing the optimality of our procedure.\\\nFirst, let us state the assumption which gives a control of the complexity of ${\\cal C}_\\Lambda$.\n\n- [**A0:**]{} the set $\\Lambda$ is compact in $\\BBr^{d_{\\Lambda}}$ and $$\\sup_{(x,y) \\in [0,1]^2} |c_\\lambda(x,y)-c_{\\lambda'}(x,y) | \\leq Q \\|\\lambda-\\lambda'\\|_{\\BBr^{d_{\\Lambda}}}^\\nu, \\; \\forall \\lambda, \\; \\lambda' \\in \\Lambda,$$ where $\\nu$ is a positive real, $Q$ is a positive constant and $\\| \\cdot \\|_{\\BBr^{d_{\\Lambda}}}$ denotes the Euclidean norm in $\\BBr^{d_{\\Lambda}}$.\n\nLower Bound {#LB}\n-----------\n\nAs it is usual for composite null hypotheses, the result of the lower bound requires the existence of a particular density $c_{\\lambda_0} \\in {\\cal C}_\\Lambda$ (see assumption [**AInf**]{} below) in order to construct a randomized class of functions which must be included in the alternatives.", "\n\n- [**AInf:** ]{} there exists a parameter $\\lambda_0$ in $\\Lambda$ such that $$\\forall (u,v)\\in[0,1]^2, \\quad\n c_{\\lambda_0}(u,v)\n >m, \\; \\mbox{ {\\rm with }} m >0.$$\n\n\\[BInf\\] Suppose that ${\\cal S}$ defined by $$\\begin{aligned}\n\\label{S}\n&&{\\cal S}=\\{\\tau=(s,p,M) , s \\geq 1/2, p \\geq 2,M >0: \\; s-2/p\\leq\ns_\\Lambda-2/p_\\Lambda, M_\\Lambda \\leq M\\}\\hspace{1.2 cm}\\end{aligned}$$ is nontrivial (see [@Spokoiny:1996]), which means that there exist $p\\geq 2$, $M>0$ and $0<s_{\\mbox{min}} <\ns_{\\mbox{max}}$ such that $$\\forall s \\in [s_{\\mbox{min}}, s_{\\mbox{max}}],\\quad (s,p,M) \\in\n{\\cal S}$$ and assume that [**A0**]{} and [**AInf**]{} hold. ", "Set $$v_{nt_n^{-1}}(\\tau)=(nt_{n}^{-1})^{-2s/(4s+2)} \\mbox{ {\\rm with} } \\;\nt_{n}=\\sqrt{\\log (\\log(n))}.$$ Then, it exists a positive constant $a$ such that $$\\begin{aligned}\n&&\\lim_{n\\rightarrow +\\infty}\\left( \\inf_{D_n}\\{ \\sup_{\\lambda \\in\n\\Lambda} \\P_\\lambda( D_n=1) + \\sup_{\\tau \\in {\\cal S}} \\sup_{c \\in\n\\Gamma(a\\; v_{nt_n^{-1}}(\\tau))} \\P_c ( { D}_n =0) \\}\\right)= 1,\n\\label{BI}\\end{aligned}$$ where the infimum is taken over any test function $D_n$.\n\nUpper Bound {#UB}\n-----------\n\nTheorem \\[BSup\\] deals with relation (\\[borneSup\\]) which holds for the test statistic $D_{\\Lambda}$ defined by relation (\\[testAdap\\]) as soon as the parameters of the methods are chosen as follows. ", "The set $J=\\{\\lfloor j_0\\rfloor,\\ldots,\\lfloor j_\\infty \\rfloor\\}$ is determined by $$\\begin{aligned}\n\\label{j0jinfty}\n&&2^{j_0} = \\log (n_2)\\log (n_1),\\quad\n2^{j_\\infty}=\\left(\\frac{n_2}{\\log(n_2)}\\right)^{1/2}\\wedge\n\\left(\\frac{n_1}{\\log(n_1)}\\right)^{1/2-1/2q},\\end{aligned}$$ where $q$ is the order of differentiability of the scaling function $\\phi$. The critical values satisfy $$\\begin{aligned}\n&&\n\\forall j\\in J,\\quad t_j=3 \\mu\\;\\frac{2^j}{n_2}\\sqrt{\\log\\log(n_2)},\n\\label{seuil}\\end{aligned}$$ where $\\mu$ is a positive constant such that $\\mu > \\sqrt{2 K_g K_1}$, and $K_g$ and $K_1$ are positive constants depending on $\\|\\phi\\|_\\infty$, $\\|c\\|_\\infty$, $\\|c_\\lambda\\|_\\infty$ and the length of the support of $\\phi$ (see Lemma \\[LDUstat\\]).", "\n\n\\[BSup\\] Let us choose $n_1=\\pi\\,\\,n$ and $n_2=(1-\\pi)n$ for some $\\pi$ in $(0,1)$. Assume that the scaling function $\\phi$ is continuously $q-$differentiable for $$q\\geq\n\\left[1-\\frac{\\log\\left(\\frac{n_2}{\\log(n_2)}\\right)}\n{\\log\\left(\\frac{n_1}{\\log(n_1)}\\right)}\\right]^{-1}.$$ Moreover assume that any density $c$ under the alternatives or any $c_\\lambda$ under the null are uniformly bounded. ", "Then, the test statistic $D_\\Lambda$ defined by (\\[testAdap\\]) is such that $$\\begin{aligned}\n \\lim_{n_1\\wedge n_2 \\rightarrow ¨+ \\infty} \\sup_{c_\\lambda \\in {\\cal\nC}_\\Lambda} \\P_{\\lambda} ( D_\\Lambda =1) = 0 \\label{PremE}.\\end{aligned}$$ Assume that [**A0**]{} holds, then there exists a positive constant $A$ such that $$\\begin{aligned}\n\\lim_{n_1\\wedge n_2 \\rightarrow +\\infty} \\sup_{\\tau \\in {\\cal S}}\n\\sup_{c \\in \\Gamma (A v_{nt_n^{-1}}(\\tau))} \\P_{c} ( D_\\Lambda =0)\n&= &0, \\label{SecondE}\\end{aligned}$$ where $$v_{nt_n^{-1}}(\\tau)=(n_2t_{n_2}^{-1})^{-2s/(4s+2)}\n\\mbox{ {\\rm and} } \\; t_{n_2}=\\sqrt{\\log (\\log(n_2))}.$$\n\nRelation (\\[borneSup\\]) of the upper bound holds since both relations (\\[PremE\\]) and (\\[SecondE\\]) are satisfied. ", "Notice also that relation (\\[PremE\\]) indicates that the test statistic $D_\\Lambda$ is asymptotically of any level in $(0,1)$.\n\nOptimality\n----------\n\nAs a corollary of Theorem \\[BInf\\] and Theorem \\[BSup\\], we obtain\n\n\\[opt\\] Under the assumptions of Theorem \\[BInf\\] and Theorem \\[BSup\\], our test procedure defined by Relation (\\[testAdap\\]) is adaptive optimal over the range of parameters $\\tau \\in {\\cal S}$ where ${\\cal S}$ is defined by equation (\\[S\\]).", "\n\nPractical results\n=================\n\nThe purpose of this section is to provide several examples to investigate the performances of the test procedure presented in Section \\[procedure\\]. ", "This part is not exactly an illustration of the theoretical part since it does not focus on the separating rate between the alternative and the null hypothesis, but it is devoted to the study of our test procedure from a risk point of view. ", "Note also that we do not use exactly the theoretical procedure described in the previous section. ", "As usual for practical purpose, we replace theoretical quantities by more adapted quantities obtained with resampling methods. ", "In the first part, we fix the test level $\\alpha=5\\%$ and we study the empirical power function. ", "In the second part, we present an application to some economical series.", "\n\nMethodology\n-----------\n\nOn the contrary to the estimation problem, a smooth wavelet is not needed. ", "The test statistic is then computed with the Haar wavelet since it has a small support and then it leads to a fast computation time. ", "The critical value of the test is determined with bootstrap methods: the standard deviation of the test statistic is computed thanks to $N_{boot}=20$ resampling. ", "The size of the simulated samples is $n=2048$ which is reasonable for bi-dimensional problems in an asymptotic context. ", "For the real life data example, the number of data is around $n=4000$. For the simulation part, the empirical level of the test is derived from $N_{MC}=500$ replications for each test problem.", "\n\nSimulations\n-----------\n\nThe setup of our simulations is closely related to the work of [@Genest/Remillard/Beaudoin:2008], except that they consider small samples (of size $150$) since their test procedures are based on the empirical copula distribution (and thus generate parametrical rates). ", "To explore various degrees of dependance, three values of Kendal’s tau are considered, namely $\\tau = 0.25, 0.50, 0.75$ for the following copula families: Clayton, Gumbel, Frank, Normal and Student with four degrees of freedom (df). ", "Calculations are made with the MatLab Sofware. ", "The results of the simulations are presented in Table \\[outside\\]. ", "For an easier reading, the estimated standard errors of the empirical powers are presented in italics. ", "Furthermore, for each testing problem we highlighted the estimated errors of the first type (estimators of $\\alpha=0.05$) using bold characters. ", "In brackets, we give the results obtained by [@Genest/Remillard/Beaudoin:2008] with their test procedures, denoted CvM and built on rank-based versions of the familiar Cramér-von Mises statistics. ", "It would be also possible, if one is interested in, to compare with the different test procedures (based on the empirical copula distribution) proposed also by [@Genest/Remillard/Beaudoin:2008].", "\n\nLet us now summarize the conclusions made from the simulation results.", "\n\n- Our test is degenerated: we almost always accept $H_0$ (when $H_0$ is true) while the procedure of Genest et al. [", "@Genest/Remillard/Beaudoin:2008] produces an excellent estimation of the prescribed level $\\alpha$. It is a characteristic of the adaptive minimax procedures.", "\n\n- For small level of dependence $\\tau=0.25$, our procedure is very competitive and produces (almost) always a better empirical power than the CvM test. ", "The results are spectacular when the fit $c_{\\lambda_0}$ is a Student(4).", "\n\n- When a large Kendal’s tau is considered, our procedure fails when the data are issued from a Clayton copula density. ", "The procedure is not available to recognize a structure of dependence modeled with a Clayton.", "\n\n- The improvment of our results with respect to the CvM test is decreasing with the Kendal’s tau. ", "The CvM test becomes better when the tau is increasing whereas for us it is the opposite.", "\n\nIn conclusion, we recommend the use of our test procedures when the Kendal’s tau is not too large since it seems to outperform the existing procedures based on the copula distribution. ", "This situation corresponds to our theoretical setup related to the functional spaces in which the unknown copula density is supposed to live. ", "Unfortunately, the practical results do not give hope for using this procedure when the copula densities present high peaks (as it is the case for the Clayton copula density with a large tau).", "\n\nReal data\n---------\n\nWe present now an application to real data of our test procedure. ", "The level of each test (with simple null hypothesis or multivariate null hypothesis) is $\\alpha=5\\%$. To obtain the empirical level, $N=50$ replications of our procedure computed with the half of the available data (chosen randomly) is used. ", "Table \\[test\\_frees\\] gives the empirical probability to reject the null hypothesis and the final decision. “", "Yes” means that we accept that the structure of dependence belongs to the considered family and “No” that we reject the fitting.", "\n\nWe consider the data of [@Frees/Valdez:1998], which were also analyzed by [@Genest/Ghoudi/Rivest:1998], [@Klugman/Parsa:1999], [@Chen/Fan:2005] and [@Genest/Quessy/Remillard:2006], among others. ", "The data consist of the indemnity payment (LOSS) and the allocated loss adjustment expense (ALAE) for 1466 general liability claims.", "\n\nWe consider the following test problems: $$H_0: c\\in {\\cal{C}}_\\Lambda$$ where the parametrical family ${\\cal{C}}_\\Lambda$ is described in Table \\[test\\_frees\\]. ", "Since the Kendall’s tau computed with the sample is $\\tau=0.31$, we choose an adapted grid of parameters for each parametrical family of copula densities. ", "Next, assuming that the density copula of the data belongs to a fixed parametric family, we estimate the parameter $ \\lambda$\n\n- by $\\hat \\lambda$ in inverting the Kendall’s tau (third part of Table \\[test\\_frees\\] where $H_0: c=c_{\\widehat{\\lambda}}$).", "\n\n- by $\\tilde \\lambda$ in minimizing the average square error (ASE) computed thanks to the benchmark given in Figure \\[fig3\\] (fourth part of Table \\[test\\_frees\\] where $H_0: c=c_{\\widetilde {\\lambda}}$). ", "For information, we give the relative $ASE$ computed with $c_{\\widetilde {\\lambda}}$ into brackets.", "\n\nThe various authors who analyzed this data set concluded that the Gumbel copula provides an adequate representation of the underlying dependence structure. ", "The Gumbel parametric family of extreme-value copulas captures the fact that almost all large indemnity payments generate important adjustment expenses (e.g., investigation and legal costs) while the effort invested in the treatment of a small claim is more variable. ", "Accordingly, the copula exhibits positive but asymmetric dependence. ", "Confirming this result, the adaptive method of estimation proposed by [@Autin/LePennec/Tribouley:2008] provides a benchmark (see Figure \\[fig3\\]) for the copula density associated with the data.", "\n\nDiscussion\n==========\n\nThe paper is mainly devoted to construct an optimal procedure for solving a general nonparametric problem of test: both hypotheses are composite, very general parametric family could be considered under the null. ", "Our procedure is proved asymptotically to be adaptive minimax and the minimax separating rate is exhibited over a range of Besov balls.", "\n\nThanks to the simulations and a application to real data, our procedure seems to be competitive on the power point of view even if the setting of test under consideration is, in the simulation study, clearly parametric.", "\n\nIt is worthwhile to point out that the copula model requires more regularity (than the usual density model) since the approximation due to the rank-based statistics needs to be accurate enough (see Lemma \\[nombredei\\]).", "\n\nOne must notice that only copulas densities belonging to [**dense Besov spaces**]{} (i.e. defined with a parameter $p$ larger than 2) are under consideration in this paper although several copula densities with a strong dependence structure belong to [**sparse Besov spaces**]{} (i.e. defined with a parameter $p$ smaller than 2). ", "As it is illustrated in the simulation study, our test procedure fails for the Clayton copula density with large parameters. ", "This density is suspected to belong to a sparse Besov ball. ", "The study of sparse Besov balls would require the determination of a new test strategy which would lead to another minimax rate of testing: these objectives are beyond those of the present paper and will be explored in a further work since the set of copulas densities contains a number of sparse functions. ", "For sparse Besov balls and in the white noise model for testing the existence of the signal, [@Lepskii/Spokoiny:1999] proved that the minimax testing rate in the sparse and the dense cases is different. ", "They also proved that it is possible to built an adaptive minimax (non linear) procedure of test for the sparse case.", "\n\nA very close problem is the sample comparison test (problem with two samples). ", "It could be interesting to test if the structure of dependence between a couple of variables $V_1=(X,Y)$ is the same as for another couple $V_2=(Z,T)$. This problem of tests could be stated as follows:\n\n$$\\begin{aligned}\nH_0 : c_{V_1}=c_{V_2} \\quad &\\mbox{ {\\rm against} }&\n \\quad H_1:(c_{V_1},c_{V_2}) \\in \\Gamma(v_n(\\tau)),\\end{aligned}$$\n\nwith $$\\begin{aligned}\n\\Gamma(v_n)=\\left \\{ c_{V_1}\\in b_{s_1,p_1,\\infty}(M_1)\\right\\} \\cap\n\\left\\{c_{V_2}\\in\nb_{s_2,p_2,\\infty}(M_2)\\right\\}\\\\\n\\cap \\left\\{(c_{V_1},c_{V_2}): \\|c_{V_1}-c_{V_2}\\| \\geq v_n.", "\n\\right\\}\\end{aligned}$$ where $v_n$ is the separating rate of both hypotheses. ", "In an analogous way as in Section 3, the rule for the comparison test would be $$\\begin{aligned}\n D&=& \\I_{ \\displaystyle{ \\max_{j\\in J}\n(\\sum_k \\widetilde {\\theta_{j,k}}- t_j )}> 0}\n \\end{aligned}$$ with $$\\begin{aligned}\n \\widetilde {\\theta_{j,k}}\n &=&\n \\frac{1}{n_2(n_2-1)}\\mathop{\\sum_{i_1,\ni_2\\in {\\cal I}_2}}_{ i_1 \\neq i_2 }\n\\left(\\phi_{j,k}\\left(\\frac{R^X_{i_1}}{n_1} ,\\frac{R^Y_{i_1}}{n_1}\n\\right)- \\phi_{j,k}\\left(\\frac{R^Z_{i_1}}{n_1}\n,\\frac{R^T_{i_1}}{n_1} \\right) \\right)\\\\&&\\hspace{3cm} \\times\n\\left(\\phi_{j,k}\\left(\\frac{R^X_{i_2}}{n_1}, \\frac{R^Y_{i_2}}{n_1}\n\\right)- \\phi_{j,k}\\left(\\frac{R^Z_{i_2}}{n_1},\n\\frac{R^T_{i_2}}{n_1} \\right)\\right),\\end{aligned}$$ where $R^X,R^Y,R^Z,R^T$ are the rank statistics associated with $X,Y,Z,T$. Using the same tools as in [@Butucea/Tribouley:2006], in which the homogeneity in law of the both samples is studied, it is possible to prove that this test is adaptive optimal and that the minimax rate of testing is $$v_n=\\left(\\frac{n}{\\sqrt{\\log(\\log(n_2))}}\\right)^{-2(s_1\\wedge\ns_2)/(4(s_1\\wedge s_2)+2)}.$$ Obviously, all these test procedures could be used in the multivariate framework ($d >2$), but as usual in the nonparametric context, it will provide slower minimax rates of testing.", "\n\nProof of Theorem \\[BSup\\]\n=========================\n\nRecall that for any given $\\lambda\\in \\Lambda$, $\\P_{\\lambda}$ (respectively $\\P_{c}$) denote the distribution associated with density $c_\\lambda$, respectively with $c$. In the same spirit, denote also $\\E_{\\lambda}$ and $\\V_{\\lambda}$ (respectively $\\E_{c}$ and $\\V_{c}$) the expectation and the variance with respect to $\\P_{\\lambda}$, respectively to $\\P_{c}$. When no index appears in $\\E$ or in $\\P$ it means that the underlying distribution is either $\\P_{c}$ or $\\P_{\\lambda}$.\n\nExpansion of the statistics of interest\n---------------------------------------\n\nFix a level $j$ in $J$. For the test problem, the statistic of interest $\\widetilde{ T_j}(\\lambda)$ (for $\\lambda\\in\\Lambda$) defined in (\\[Ttilde\\]) is an estimate of $$\\begin{aligned}\n\\label{expansionVRAIE}\nT_j(\\lambda) &=&\\sum_k \\theta_{j,k}(\\lambda)\n=\\sum_k\\left(c_{j,k}-c_{j,k}(\\lambda)\\right)^2,\\end{aligned}$$ which is the quantity that we need to detect under the alternative. ", "It would be useful to expand the statistic $\\widetilde{T_j}(\\lambda)$ as follows $$\\begin{aligned}\n&&\\widetilde{T_j}(\\lambda)= 2 T_j^\\diamond (\\lambda)\n+{T_j}^\\heartsuit +T_j^\\spadesuit +2 T_j^\\clubsuit\n(\\lambda)+T_j(\\lambda) \\label{expansion}\n\\\\&&\\quad \\quad =2\\sum_k {\\theta}_{j,k}^\\diamond(\\lambda)+\\sum_k\n\\theta_{j,k}^\\heartsuit+\\sum_k {\\theta}_{j,k}^\\spadesuit +2\\sum_k\n{\\theta}_{j,k}^\\clubsuit (\\lambda)+\\sum_k \\theta_{j,k}(\\lambda),\n\\nonumber\\end{aligned}$$ where $$\\begin{aligned}\n{\\theta}_{j,k}^\\heartsuit & =&\n\\frac{1}{n_2(n_2-1)}\\mathop{\\sum_{i_1, i_2 \\in {\\cal I}_2}}_{ i_1\n\\neq i_2 }\\left(\\phi_{j,k}(F(X_{i_1}), G(Y_{i_1}))-c_{j,k}\\right)\\\\\n&&\\times (\\phi_{j,k}\\left( F(X_{i_2}), G(Y_{i_2}))-c_{j,k}\\right)\n\\\\\n\\theta_{j,k}^\\spadesuit\n &=&\\frac{1}{n_2(n_2-1)}\\mathop{\\sum_{i_1,\ni_2 \\in {\\cal I}_2}}_{ i_1 \\neq i_2} \\left(\\phi_{j,k}\\left(\\frac{R_{i_1}}{n_1},\n\\frac{S_{i_1}}{n_1} \\right)-\n\\phi_{j,k}\\left(F(X_{i_1}),G(Y_{i_1})\\right)\n\\right)\\,\\\\&&\\times\\left(\\phi_{j,k}\\left(\\frac{R_{i_2}}{n_1},\n\\frac{S_{i_2}}{n_1}\n\\right)-\\phi_{j,k}\\left(F(X_{i_2}),G(Y_{i_2})\\right) \\right)\n\\\\\n\\theta_{j,k}^\\clubsuit (\\lambda)\n &=&\\frac{1}{n_2(n_2-1)}\\mathop{\\sum_{i_1,\ni_2 \\in {\\cal I}_2}}_{ i_1 \\neq i_2\n}\\left(\\phi_{j,k}\\left(\\frac{R_{i_1}}{n_1}, \\frac{S_{i_1}}{n_1}\n\\right)-\\phi_{j,k}\\left(F(X_{i_1}),G(Y_{i_1})\\right)\n\\right)\\\\&&\\times\\left(\\phi_{j,k}\\left(F(X_{i_2}),G(Y_{i_2})\n\\right)- c_{j,k}(\\lambda) \\right)\\\\\n{\\theta}_{j,k}^\\diamond (\\lambda) &=& \\frac{1}{n_2}\\sum_{i_1\\in\n{\\cal I}_2} \\left(\\phi_{j,k}(F(X_{i_1}), G(Y_{i_1}))-c_{j,k}\\right)\n(c_{j,k}-c_{j,k}(\\lambda)).\\end{aligned}$$ The sequence $\\{c_{j,k}\\}_{j,k}$ denotes the unknown scaling coefficients of the unknown copula density $c$. Recall that $$\\widehat{T_j}(\\lambda)=\\sum_k \\widehat{\\theta_{j,k}}(\\lambda),$$ with $$\\begin{aligned}\n\\widehat{\\theta_{j,k}}(\\lambda) &=&\n\\frac{1}{n_2(n_2-1)}\\mathop{\\sum_{i_1, i_2 \\in {\\cal I}_2}}_{ i_1\n\\neq i_2 }\\left(\\phi_{j,k}(F(X_{i_1}),\nG(Y_{i_1}))-c_{j,k}(\\lambda)\\right) \\\\\n&&\\times \\left(\\phi_{j,k}( F(X_{i_2}),\nG(Y_{i_2}))-c_{j,k}(\\lambda)\\right).\\end{aligned}$$ The following lemma gives some evaluation for the first moments of each statistic of interest.", "\n\n\\[etude-stat\\] Let $q$ be a positive integer and assume that $\\phi$ is continuously $q-$differentiable. ", "Let $j$ be a level smaller than $ j_\\infty$ defined in (\\[j0jinfty\\]). ", "Then, it exists some positive constant $\\kappa$ which may depend on $\\phi$, $\\|c\\|_\\infty$, $\\|c_\\lambda\\|_{\\infty}$ and $M$ such that $$\\begin{aligned}\n\\E \\widehat{T_{j}}(\\lambda)&=&T_j(\\lambda)\\quad\\mbox{ {\\rm and} }\\quad \\V\n\\widehat{T_{j}}(\\lambda)\\leq\n\\kappa \\left(\\left(\\frac{2^{j}}{n_2}\\right)^2+\n\\left(\\frac{2^{j}}{n_2}\\right)\\, T_j(\\lambda)\\right) \\\\\n\\E |T_{j}^\\spadesuit |&\\leq& \\kappa\\;\\frac{\\log(n_1)}{n_1}\\\\\n \\E_c\n|T_{j}^\\clubsuit (\\lambda)|&\\leq& \\kappa\\,\n\\left(\\frac{\\log(n_1)}{n_1}T_j(\\lambda)\\right)^{1/2}\\;\\mbox{ {\\rm\nand} }\\; \\E_\\lambda( T_{j}^\\clubsuit (\\lambda))^2\\leq \\kappa\n\\,2^{j}\\left(\\frac{\\log(n_1)}{n_2n_1}\\right).\\end{aligned}$$\n\nUsing the Bernstein Inequality, we establish the following bound for the deviation of the statistic $T_{j}^\\diamond(\\lambda)$ under the alternative. ", "The proof is postponed to Appendix B.\n\n\\[Bernstein\\] For any level $j$, for all $x>0$ $$\\begin{aligned}\n\\P_c\\left(\\,|T_{j}^\\diamond(\\lambda)|\\,\\geq x\\right)\\leq\n\\exp\\left(-K (\\frac{n_2^2x^2}{n_2 T_j(\\lambda)+n_2x2^{j}T_j(\\lambda)^{1/2}})\\right),\\end{aligned}$$ where $K$ is a positive constant depending on $L,\\|\\phi\\|_\\infty $ and $\\|c\\|_\\infty$.\n\nUsing a result from [@Gine/Latala/Zinn:2000], we establish the following bound for the deviation of the $U$-statistics $\\widehat{T_j}(\\lambda)$ and $T_j^\\heartsuit$. The proof is postponed to Appendix C.\n\n\\[LDUstat\\] For any level $j$, as soon as $ x\\geq 2^j\\,n_2^{-1}\\sqrt{\\log(\\log(n_2))}$, for all $\\mu>0$, $$\\begin{aligned}\n\\P_\\lambda\\left(|\\widehat{T_j}(\\lambda)|> \\mu x \\right)+\\P_c\\left(|\nT_j^\\heartsuit|> \\mu x \\right)&\\leq& K_g(\\log(n_2))^{-\\delta}\\end{aligned}$$ for any positive $\\delta \\leq \\mu^2(K_g K_1)^{-1}$, where $K_g$ is an universal positive constant given in [@Gine/Latala/Zinn:2000] and $K_1$ is a positive constant depending on $L,\\|\\phi\\|_\\infty $ and either $\\|c_{\\lambda}\\|_\\infty$ or $\\|c\\|_\\infty$ depending on the underlying distribution i.e. either $\\P_\\lambda$ or $\\P_c$.\n\nProof of Relation (\\[PremE\\]) (First type error)\n------------------------------------------------\n\nLet us fix $\\lambda\\in\\Lambda$ and set $$p_\\lambda = \\P_{\\lambda}\n\\left(\\max_{j \\in J} \\left[\\inf_{\\lambda' \\in \\Lambda}\n\\widetilde{T_j}(\\lambda') - t_j\\right]\n>0 \\right).$$ Notice that under the null $$T_j^\\diamond (\\lambda)\n =T_j(\\lambda)=0 \\mbox{ and }\n {T_j}^\\heartsuit=\\widehat{T_j}(\\lambda).$$ Using expansion (\\[expansion\\]), we get $$\\begin{aligned}\np_{\\lambda}& \\leq & \\sum_{j \\in J} \\P_{\\lambda}\\left(\n\\inf_{\\lambda^\\prime \\in \\Lambda} \\widetilde{T_j}(\\lambda^\\prime) > t_j\\right)\\nonumber \\\\\n& \\leq & \\sum_{j \\in J} \\P_{\\lambda}\\left(\n \\widetilde{T_j}(\\lambda) > t_j\\right)\\nonumber \\\\\n& \\leq & \\sum_{j \\in J}\n\\left\\{\\P_{\\lambda}\\left(|\\widehat{T_j}(\\lambda) |>\n\\frac{t_j}{3}\\right) + \\P_{\\lambda} \\left(|T_j^\\spadesuit |\n> \\frac{t_j}{3}\\right) + \\P_{\\lambda} \\left(|T_j^\\clubsuit (\\lambda)| >\n\\frac{t_j}{3}\\right)\\right\\} \\nonumber\\end{aligned}$$ Due to Lemma \\[etude-stat\\] and using Markov Inequality, we obtain $$\\begin{aligned}\np_{\\lambda}& \\leq & \\sum_{j\\in J}\n\\P_{\\lambda}\\left(|\\widehat{T_j}(\\lambda) |> \\frac{t_j}{3} \\right) +\n\\sum_{j\\in J} \\left\\{\\frac{\\E_{\\lambda}| T_j^\\spadesuit |}{\n(t_j/3)}+\\frac{ \\E_{\\lambda}( T_j^\\clubsuit (\\lambda))^2}{\n(t_j/3)^2}\\right\\} \\nonumber \\\\\n&\\leq& \\sum_{j\\in J} \\P_{\\lambda}\\left(|\\widehat{T_j}(\\lambda) |>\n\\frac{t_j}{3}\\right)\\\\&& + K\\sum_{j\\in\nJ}\\left\\{\\left(t_j/3\\right)^{-1}\\frac{\\log(n_1)}{n_1}+\\left(t_j/3\\right)^{-2}\\left(\n\\frac{2^{j} \\log(n_1)}{ n_1n_2 }\\right)\\right\\}.", "\n \\label{THM1firstIneq}\\end{aligned}$$ Notice that $\\widehat{T_j}(\\lambda)$ is centered under $\\P_\\lambda$, then applying Lemma \\[LDUstat\\], where $t_j$ is $t_j= 3\\mu\n\\;2^j\\,n_2^{-1}\\sqrt{\\log(\\log(n_2))}$, the constant $\\mu$ is defined in (\\[seuil\\]) and since $\\mbox{{\\rm card}} (J) \\leq \\log\n(n_2)$, one obtains $$\\begin{aligned}\np_{\\lambda}& \\leq & K_g\\mbox{{\\rm card}} (J)\\; (\\log(n_2))^{-\\delta} + K \\mbox{{\\rm card}} (J)\n2^{-j_0}\\left(\\frac{ n_2\\log(n_1)}{ n_1\\sqrt{\\log\\log(n_2)} }\\right)\\\\\n& & +K \\mbox{{\\rm card}} (J)\n2^{-j_0}\\left( \\frac{\\log(n_1) n_2^2}{ n_2 n_1 \\sqrt{\\log\\log(n_2)}}\\right)\\\\\n& \\leq & K_g (\\log(n_2))^{1-\\delta} + K 2^{-j_0} \\left(\\frac{\\log(n_1)\\log(n_2)}{\\sqrt{\\log\\log(n_2)} }\\right),\\end{aligned}$$ where the last inequality holds since $\\delta$ satisfies $\\delta \\leq \\mu^2(2 K_g K_1)^{-1}$ (see Lemma \\[LDUstat\\]). ", "Since $\\mu$ is such that $\\mu >\n\\sqrt{2 K_g K_1}$, relation (\\[PremE\\]) is proved if one takes $\\delta=\\mu^2(2 K_g K_1)^{-1}$.\n\nProof of Relation (\\[SecondE\\]) (Second type error)\n---------------------------------------------------\n\nLet us fix $\\tau \\in {\\cal S}$ and $c\\in\\Gamma(Av_{nt_n^{-1}}(\\tau))$ and set $$p_c=\\P_{c} \\left(\\max_{j \\in J} \\inf_{\\lambda \\in\n\\Lambda} \\widetilde{T_j}(\\lambda) - t_j \\leq 0 \\right).$$ Using the expansion (\\[expansion\\]), we get, for any $j^\\star\\in\nJ$ $$\\begin{aligned}\np_{c}& \\leq & \\P_c \\left( \\inf_\\lambda \\left\\{2 T_{j^\\star}^\\diamond\n(\\lambda)+T_j(\\lambda)+ T_{j^\\star}^\\heartsuit +\nT_{j^\\star}^\\spadesuit +2T_{j^\\star}^\\clubsuit (\\lambda)\\right\\}\\leq\nt_{j^\\star}\n\\right)\\nonumber\\\\\n & \\leq & \\P_c \\left( \\inf_\\lambda \\left\\{2\nT_{j^\\star}^\\diamond (\\lambda)+T_{j^\\star}(\\lambda)\\right\\}\\leq 2t_{j^\\star}\n\\right)\\nonumber\\\\\n&&+ \\P_c \\left(T_{j^\\star}^\\heartsuit + T_{j^\\star}^\\spadesuit +2\n\\inf_\\lambda \\left\\{T_{j^\\star}^\\clubsuit (\\lambda)\\right\\}\\geq\nt_{j^\\star}\\right)\\nonumber\\\\\n& \\leq & \\P_c \\left( \\inf_\\lambda \\left\\{2 T_{j^\\star}^\\diamond\n(\\lambda)+T_{j^\\star}(\\lambda)\\right\\}\\leq 2t_{j^\\star} \\right)+ \\P_c\n\\left(T_{j^\\star}^\\heartsuit \\geq t_{j^\\star}/3\\right)\\nonumber\\\\\n&&\n + \\P_c \\left(T_{j^\\star}^\\spadesuit \\geq t_{j^\\star}/3\\right)\n+\\P_c \\left( \\inf_\\lambda \\left\\{T_{j^\\star}^\\clubsuit\n(\\lambda)\\right\\}\\geq t_{j^\\star}/6\\right)\n\\nonumber\\\\\n&=&p_{c1}(j^\\star)+p_{c2}(j^\\star)+p_{c3}(j^\\star)+p_{c4}(j^\\star).\\label{2proba}\\end{aligned}$$ Let us explain how $j^\\star$ is chosen. ", "From the wavelet expansion (\\[decomposition\\]) and Lemma \\[etude-stat\\], one has $$\\E_c\\widehat{T_{j^\\star}}(\\lambda)=T_{j^\\star}(\\lambda)=\\int(c-c_\\lambda)^2-B_{j^\\star}(\\lambda),$$ where $T_{j^\\star}$, $B_{j^\\star}$ are defined in (\\[TB\\]) and $t_j^\\star$ is the critical value given in (\\[seuil\\]). ", "Since $c$ is in $\\Gamma(Av_{nt_n^{-1}}(\\tau))$ and $c_\\lambda$ lies in $b_{s_\\Lambda, p_\\Lambda , \\infty}(M_\\Lambda)\n \\subset b_{s, p , \\infty} (M)$, the function $(c-c_\\lambda)$ is in $b_{s,p,\\infty}(M)$. We can choose $j^{\\star}$ such that $$\\begin{aligned}\n\\label{jstar}\n2^{j^\\star} & =& \\left(\\frac{{\\tilde\nK}}{3\\mu}\\frac{n_2}{\\sqrt{\\log\\log(n_2)}}\\right)^{1/(2s+1)},\\end{aligned}$$ which is possible due to our choice of $j_\\infty$ and because $s\\geq 1/2$; the constant ${\\tilde\nK}$ appears in (\\[reg\\]). ", "It implies that $B_{j^\\star}\\leq t_{j^\\star}$ since $B_{j^{\\star}} \\leq {\\tilde K} 2^{-2j^{\\star}s}$ (see Inequality (\\[reg\\])). ", "Next, since $c\\in \\Gamma(Av_{nt_n^{-1}}(\\tau))$, one has $\\int(c-c_{\\lambda^\\prime})^2\\geq A^2(v_{nt_n^{-1}}(\\tau))^2$ for all $\\lambda^\\prime \\in \\Lambda$. Focusing on rates $v_{nt_n^{-1}}(\\tau)$ combined with positive constant $A$ which satisfy $4t_{j^\\star}\\leq (A v_{nt_n^{-1}}(\\tau))^2$, one obtains $$\\begin{aligned}\n\\label{tT}\n\\frac{t_{j^\\star}}{\\E_c\\widehat{T_{j^\\star}}(\\lambda)}\\,=\\,\\frac{t_{j^\\star}}{T_{j^\\star}(\\lambda)}&\\leq\n&1/3.\\end{aligned}$$ Coming back to the evaluation of the probability terms (see relation (\\[2proba\\])), we first consider $p_{c1}(j^\\star)$. Consider an $\\eta$-net $\\Lambda_{\\eta}$ on the set $\\Lambda$ that is for any $\\lambda$ in $\\Lambda$, denote ${\\tilde \\lambda}$ the closest (in the Euclidean sense) element in $\\Lambda_{\\eta}$ to $\\lambda$ (closer than $\\eta$). ", "Due to assumption [**A0**]{}, let us prove that for any $j \\in J$, $T_{j}^\\diamond\n({\\tilde \\lambda})+T_{j}({\\tilde \\lambda})$ is close to $T_{j}^\\diamond\n(\\lambda)+T_{j}(\\lambda)$: $$\\begin{aligned}\n|T_{j}^\\diamond ({\\tilde \\lambda})-T_{j}^\\diamond ({\\lambda})| & = &\n\\left | \\sum_k \\left [ \\frac{1}{n_2} \\sum_{i_1 \\in {\\cal I}_2}\n\\left(\\phi_{j,k}(F(X_{i_1}), G(Y_{i_1}))-c_{j,k} \\right)\n\\left(c_{j,k}({\\tilde \\lambda})-c_{j,k}(\\lambda) \\right) \\right ]\\, \\right|\\nonumber \\\\\n&\\leq & \\left[ \\sum_k \\left(c_{j,k}({\\tilde\n\\lambda})-c_{j,k}(\\lambda)\\right)^2 \\right]^{1/2}\\nonumber \\\\\n&&\\hspace{1cm}\\times \\,\\left[ \\sum_k \\left(\\frac{1}{n_2}\n\\sum_{i_1 \\in {\\cal I}_2} \\left(\\phi_{j,k}(F(X_{i_1}), G(Y_{i_1}))-c_{j,k}\\right) \\right)^2\\right]^{1/2}\\nonumber \\\\\n&\\leq & Q \\eta^{\\nu} \\; 2^{2j}(2\\|\\phi\\|_{\\infty} + \\|c\\|_{\\infty}\n2^{-4j})^{1/2}. ", "\\nonumber\\end{aligned}$$ In the same way, one has, $$\\begin{aligned}\n|T_j(\\lambda) - T_j({\\tilde \\lambda})| & \\leq &\n \\kappa Q \\eta^{\\nu}\\;\\left(2\\mbox{{\\rm max}}(\\|c\\|, \\|c_{{\\tilde \\lambda}}\\|) + Q \\eta^{\\nu}\\right).\\end{aligned}$$ Choosing $\\eta=n^{-b}$ with $b \\nu >1$, then by (\\[tT\\]) and applying Lemma \\[Bernstein\\], we get $$\\begin{aligned}\np_{c1}(j^\\star)&\\leq& \\sum_{\\lambda\\in\\Lambda_{n^{-b}}} \\P_c\\left( 2\nT_{j^\\star}^\\diamond\n(\\lambda)+T_{j^\\star}(\\lambda)\\leq 2t_{j^\\star}\\right)\\nonumber\\\\\n&\\leq& \\sum_{\\lambda\\in\\Lambda_{n^{-b}}} \\P_c\\left( T_{j^\\star}^\\diamond\n(\\lambda)\\leq -T_{j^\\star}(\\lambda)/6\\right)\\nonumber\\\\\n&\\leq&\n\\sum_{\\lambda\\in\\Lambda_{n^{-b}}}\\exp\\left[-K\\,\\left(\\frac{n_2}{2^{2j^\\star}}T_{j^\\star}(\\lambda)\\wedge\n\\frac{n_2}{2^{j^\\star}}T_{j^\\star}(\\lambda)^{1/2} \\right)\\right]\\nonumber\\\\\n&\\leq&\n\\sum_{\\lambda\\in\\Lambda_{n^{-b}}}\\exp\\left[-K_2\\left(n_2 t_{j^\\star}\\wedge\n\\frac{n_2}{2^{j^\\star}}t_{j^\\star}^{1/2} \\right)\\right]\\nonumber\\\\\n&\\leq &(\\frac{D(\\Lambda)}{n^{-b}})^{d_{\\Lambda}}\\; \\exp\\left[\n-K_3\\,\\left(2^{j^{\\star}}(\\log\\log(n_2))^{1/2}\\right.\\right.\\nonumber\\\\\n&&\\hspace{4cm} \\left.\\left. ", "\\wedge \\left(\\frac{ n_2}\n{2^{j^{\\star}/2}}\\right)^{1/2}(\\log\\log(n_2))^{1/4}\\right)\n\\right]\\label{2ndeProbaT3}\n$$ where $D(\\Lambda)$ is the diameter of $\\Lambda$ and $K$, $K_2$ and $K_3$ are positive constants. ", "Both terms behind the minus sign in the exponential of the right hand side of the last inequality tend to infinity with a power of $n$ since $s > 1/2$. This implies that $p_{c1}(j^\\star)$ goes to zero as $n$ goes to infinity. ", "Now, it remains to verify that $p_{c2}(j^\\star),p_{c3}(j^\\star)$ and $p_{c4 }(j^\\star)$ are going to zero as $n_1\\wedge n_2$ goes to infinity. ", "Using again the bound (\\[tT\\]), Lemma \\[LDUstat\\] for some positive $\\delta$, Lemma \\[etude-stat\\] and the definition of the critical value (\\[seuil\\]), one gets\n\n$p_{c2}(j^\\star)+p_{c3}(j^\\star)+p_{c4}(j^\\star)$\n\n$$\\begin{aligned}\n&& \\leq \\P_c \\left(T_{j^\\star}^\\heartsuit \\geq\nt_{j^\\star}/3\\right)+\n + \\P_c \\left(T_{j^\\star}^\\spadesuit \\geq t_{j^\\star}/3\\right)+\n \\P_{c}\\left( T_{j^\\star}^\\clubsuit\n(\\lambda)\\geq t_{j^\\star}/6\\right)\\nonumber\\\\\n&& \\leq K_g (\\log (n_2))^{-\\delta}+ 9\n\\,\\frac{\\E_c|T_{j^\\star}^\\spadesuit (\\lambda)|^2 }{t_{j^\\star}^2}+\n\\,6\\,\\frac{\\E_c|T_{j^\\star}^\\clubsuit |\n}{t_{j^\\star}} \\nonumber \\\\\n&& \\leq K_g (\\log (n_2))^{-\\delta}+ 9 \\kappa\n\\frac{2^{j^\\star}\\log(n_1)}{n_1n_2\\,t_{j^\\star}^2}\n+\\,6 \\kappa \\frac{\\log(n_1)}{n_1\\,t_{j^\\star}}\\nonumber\\\\\n&&\\quad \\leq K_g (\\log (n_2))^{-\\delta}+ 6\\kappa \\;\n\\left(2^{-j^\\star}\\frac{n_2}{n_1}\\frac{\\log(n_1)}{\\sqrt{\\log\\log(n_2)}}\\right)\\label{2ndeProbaT2},\\end{aligned}$$\n\nwhich tends to zero with our choice of $j^\\star$ and where $\\kappa$ is the positive constant appearing in Lemma \\[etude-stat\\]. ", "Inequalities (\\[2ndeProbaT3\\]) and (\\[2ndeProbaT2\\]) entail that the right hand side of (\\[2proba\\]) is less than any $\\alpha \\in (0,1)$ as $n$ is large enough. ", "To finish the proof, observe that the choice of $v_{nt_n^{-1}}(\\tau)$ is driven by the fact that it corresponds to the smallest sequence such that $4 t_j^{\\star} \\leq (A\nv_{nt_n^{-1}}(\\tau))^2$, which leads to $$v_{nt_n^{-1}}(\\tau)\\geq \\left(2^{j^\\star} \\mu \\frac{\\sqrt{\\log(\\log n_2)}}\n{n_2 }\\right)^{1/2} \\geq \\left(\\frac{n_2}\n{\\sqrt{\\log\\log(n_2)}}\\right)^{2s/(4s+2)}.$$\n\nProof of Theorem \\[BInf\\]\n=========================\n\nWithout loss of generality, we suppose that the support of the scaling function $\\phi$ and its associated wavelet function $\\psi$ is $[0,1]$. Moreover recall that $\\int_0^1 \\psi^{\\epsilon} =0$. Let us give some $a>0$ which must be small enough.", "\n\n Discretisation of ${\\cal S}$\n-----------------------------\n\nFor any given $\\tau=(s,p,M) \\in {\\cal S}$, denote by $j(\\tau)$ the level $$2^{j(\\tau)}=(nt_n^{-1})^{2/(4s+2)}$$ and define $s_j$ the solution of the equation $j=j(s_j,p,M)$ for any resolution level $j \\in {\\tilde J}=\\{j_{s_{\\mbox{max}}},\\ldots, j_{s_{\\mbox{min}}}\\} \\subset \\{j_0,\\ldots,j_{\\infty}\\}$ with $$j_{s_{\\mbox{max}}}=\\lfloor j(s_{\\mbox{max}},p,M)\\rfloor \\mbox{ and } j_{s_{\\mbox{min}}}=\\lfloor\nj(s_{\\mbox{min}},p,M)\\rfloor.$$\n\nConsider now the set ${\\cal S}_n=\\{\\tau_j=(s_j,p,M), j\\in {\\tilde J}\\}$ which appears as a discretisation version of a subset of ${\\cal S}$ whose cardinality is of order $O(\\log (n))$.\n\nPrior and parametric family included in the alternatives\n--------------------------------------------------------\n\nFor any $s_j \\in {\\cal S}_n$, define a prior $\\pi_j$ which is concentrated on the class of the random functions $$c_j(u,v) = c_{\\lambda_0} (u,v) + \\sum_k \\sum_{\\epsilon =1}^3 \\delta_{k} u_{j}(n) \\psi^{\\epsilon}_{j,k}(u,v),$$ where $c_{\\lambda_0}$ is defined in assumption [**AInf**]{} and $$P(\\delta_{k} =1 )=P(\\delta_{k} =-1 )=1/2\\quad \\mbox{ and }\\quad\nu_{j}(n)= C_1 M (nt_n^{-1})^{-\\frac{2(s_j+1)}{4s_j+2}}$$ for $C_1$ such that $3 M^2 C_1^2=2a^2$. Let $j$ be any index in ${\\tilde J}$. Since $\\int \\psi=0$ and when $a$ is small enough (to guarantee that $c_j\\geq 0$), $c_j$ is a density. ", "Easy calculations imply that $$\\begin{aligned}\n\\| c_j-c_{\\lambda_0}\\|^2 = M^2 C_1^2 \\; (v_{nt_n^{-1}}(\\tau_j))^2\n> a^2 \\;(v_{nt_n^{-1}})^2\\label{L2} .\\end{aligned}$$ Moreover, if $a$ is small enough, we have $3\\; C_1^p <1$ and $$\\begin{aligned}\n2^{j(s_j+1-2/p)p} \\sum_k \\sum_{\\epsilon} | \\int\nc_j\\psi_{jk}^\\epsilon |^p&=&\n 2^{j(s_j+1-2/p)p} \\sum_k \\sum_{\\epsilon} | u_{j}(n) |^p\\\\\n & =& 3 C_1^p M^p \\leq M^p,\\end{aligned}$$ implying that $c_j\\in b_{s_j,p,\\infty}(M)$. Denote by ${\\cal A}_{j,n}\n(a)$ the set of densities $${\\cal A}_{j,n}\n(a)=\n\\{c \\in b_{s_j,p,\\infty}(M) : \\;\n\\displaystyle{\\inf_{\\lambda \\in \\Lambda}}\n \\|c - c_{\\lambda}\\|^2 > a^2 (v_{nt_n^{-1}}(\\tau_j))^2 \\}.$$ and consider the variation between both distributions $\\P_{\\lambda_0}$ and $\\P_\\Pi$ $$Var(\\P_{\\lambda_0}, \\P_\\Pi)= \\frac 1 2 \\int\n\\left|\\frac{d\\P_{\\Pi}}{d\\P_{\\lambda_0}}-1\\right| d\\P_{\\lambda_0},$$ where $$\\frac{d\\P_{\\Pi}}{d\\P_{\\lambda_0}} = \\frac{1}{N_n } \\sum_{j \\in {\\tilde J}}\n \\frac{d\\P_{j}}{d\\P_{\\lambda_0}}=\\frac{1}{N_n } \\sum_{j \\in {\\tilde J}}\n \\E_{\\pi_j}^{(n)} [\\frac{c_{j}}{c_{\\lambda_0}}],$$ and $N_n=\\mbox{ {\\rm card} }({\\tilde J})$. Assuming that the following assertion holds $$\\begin{aligned}\n\\lim_{n \\rightarrow \\infty} \\inf_{j \\in {\\tilde J}} \\pi_j (c \\in {\\cal\nA}_{j,n} (a))=1 ,\\label{alternative}\\end{aligned}$$ we deduce that the left hand side ($LHS$) of relation (\\[BI\\]) without the limit is bounded from below by $$\\begin{aligned}\nLHS&\\geq& \\P_{\\lambda_0}( D_n=1) + \\sup_{\\tau_j \\in {\\cal S}_n} \\sup_{c\n\\in {\\cal A}_{j,n} (a)} \\P_c ( { D}_n =0)\\nonumber\\\\\n & \\geq & 1 - Var(\\P_{\\lambda_0}, \\P_\\Pi)(1 + o_n(1)),\\label{Variation}\\end{aligned}$$ as $n$ large enough. ", "Since the supports of the functions $c_j$ and $c_{j'}$ are disjoint for $j \\neq j'$, one has $$\\begin{aligned}\n1 - Var(\\P_{\\lambda_0}, \\P_\\Pi) & \\geq & 1 - \\frac 1 2\n\\frac{1}{N_n^2} \\sum_{j \\in {\\tilde J}}\\E_{\\lambda_0}\n \\left[ \\left(\\int \\prod_{i=1}^n \\frac{c_j(U_i,V_i)}{c_{\\lambda_0}(U_i,V_i)} d\\pi_j(c_j)\\right)^2 -1 \\right] \\nonumber\\\\\n&\\geq& 1 -o_n(1)\n \\end{aligned}$$ provided that $$\\begin{aligned}\n \\lim_{n \\rightarrow \\infty} \\frac{1}{N_n^2} \\sum_{j \\in {\\tilde J}}\n \\E_{\\lambda_0} \\left[ \\left(\\int \\prod_{i=1}^n \\frac{c_j(U_i,V_i)}{c_{\\lambda_0}\n (U_i,V_i)} d\\pi_j(c_j)\\right)^2 \\right] = 0. ", "\\label{vrais}\\end{aligned}$$ Relation (\\[BI\\]) is thus proved if (\\[alternative\\]) and (\\[vrais\\]) are satisfied. ", "The remaining proofs are given in the sequel.", "\n\nProof of Relation (\\[alternative\\])\n-----------------------------------\n\nLet $\\Lambda^\\prime$ be a subsect of $\\Lambda$. We have $$\\begin{aligned}\n\\pi_j\\left(\\inf_{\\lambda \\in \\Lambda} \\|c_j-c_\\lambda\\|^2 \\leq a^2\n(v_{nt_n^{-1}}(\\tau_j))^2\\right)&\\leq& \\pi_j\\left(\\inf_{\\lambda \\in\n\\Lambda/\\Lambda^\\prime} \\|c_j-c_\\lambda\\|^2 \\leq a^2\n(v_{nt_n^{-1}}(\\tau_j))^2\\right)\\end{aligned}$$ $$\\begin{aligned}\n\\label{proba_inf}\n+\\quad \\pi_j\\left(\\inf_{\\lambda \\in \\Lambda^\\prime}\n\\|c_j-c_\\lambda\\|^2 \\leq a^2 (v_{nt_n^{-1}}(\\tau_j))^2\\right)\\end{aligned}$$ Consider the particular subset $\\Lambda'$ defined by $$\\Lambda'=\\{\\lambda \\in \\Lambda: \\; \\| c_{\\lambda_0} - c_\\lambda\\|^2 \\leq 6 C_1^2 M^{2}\n (v_{nt_n^{-1}}(\\tau_j))^2\\}.$$ Notice that $$\\lambda \\in \\Lambda/\\Lambda'\\Longrightarrow \\| c_\\lambda -c_j\\|^2\n\\geq a^2(v_{nt_n^{-1}}(\\tau_j))^2$$ due to the choice of $C_1$. It implies that the first term in the right hand side of (\\[proba\\_inf\\]) is null and then, it remains to prove that $$\\begin{aligned}\n\\lim_{n \\rightarrow \\infty} \\pi_j\\left(\\inf_{\\lambda \\in \\Lambda'}\n\\|c_j-c_\\lambda\\|^2 \\leq a^2 (v_{nt_n^{-1}}(\\tau_j))^2\\right) =0.", "\n\\label{lambda'}\\end{aligned}$$ Since $\\lambda$ is in $\\Lambda^\\prime$, we get $$\\begin{aligned}\n \\| c_\\lambda -c_j\\|_2^2 & =& \\|\nc_{\\lambda_0} - c_\\lambda\\|^2 + \\sum_k \\sum_{\\epsilon} u_{j}(n)^2 +\n2 \\sum_k\\sum_{\\epsilon} \\delta_{k} u_{j}(n)\nB_{j,k,\\lambda,\\lambda_0}\n\\\\\n& \\geq & 3 C_1^2 M^{2} \\; (v_{nt_n^{-1}}(\\tau_j))^2 + 2 \\sum_k\n\\delta_{k} u_{j}(n) \\sum_{\\epsilon} B_{j,k,\\lambda,\\lambda_0},\\end{aligned}$$ where $$B_{j,k,\\lambda,\\lambda_0}= \\int\n\\psi^{\\epsilon}_{j,k}(c_{\\lambda_0} - c_\\lambda).$$ Therefore assertion (\\[lambda’\\]) is equivalent to $$\\begin{aligned}\n\\lim_{n \\rightarrow \\infty} \\pi_j\\left(\\inf_{\\lambda \\in \\Lambda'}\n2\\sum_k \\delta_{k}u_{j}(n) B_{j,k,\\lambda,\\lambda_0}\\leq -a^2\n(v_{nt_n^{-1}}(\\tau_j))^2\\right)=0 \\nonumber.\\end{aligned}$$ or $$\\begin{aligned}\n\\lim_{n \\rightarrow \\infty} \\pi_j\\left(\\sup_{\\lambda \\in \\Lambda'}\n2\\sum_k (-\\delta_{k})u_{j}(n) B_{j,k,\\lambda,\\lambda_0}\\geq a^2\n(v_{nt_n^{-1}}(\\tau_j))^2\\right)=0 \\label{reseau}.\\end{aligned}$$\n\nWe can construct in the Euclidean metric an $\\eta$-net $\\Lambda'_{\\eta}$ on the subset $\\Lambda'$. For any $\\lambda$ in $\\Lambda'$, denote ${\\tilde \\lambda}$ the closest element in $\\Lambda'_{\\eta}$ to $\\lambda$ in the Euclidean sense. ", "Then for any $\\lambda \\in \\Lambda'$, we have by assumption [**A0**]{}: $$\\begin{aligned}\n|\\sum_k \\delta_{k}u_{j}(n)( B_{j,k,\\lambda,\\lambda_0}- B_{j,k,{\\tilde \\lambda},\\lambda_0})| & \\leq & u_{j}(n)\\sum_k |B_{j,k,\\lambda,\\lambda_0}- B_{j,k,{\\tilde \\lambda},\\lambda_0}| \\\\\n& \\leq & u_{j}(n)\\sum_k Q \\eta^{\\nu} 2^{-j} \\| \\psi^{\\epsilon}\\|_{\\infty}\\\\\n&\\leq & 2^{j(s+1)} 2^{2j} Q \\eta^{\\nu} 2^{-j} \\| \\psi^{\\epsilon}\\|_{\\infty}\\\\\n& \\leq & \\kappa 2^{j s} \\eta^{\\nu},\\end{aligned}$$ where $\\kappa$ is a positive constant depending on $Q$, $C_1$, $M$ and $\\| \\psi^{\\epsilon}\\|_{\\infty}$. Chosing $\\eta=n^{-b}$, with $b\\nu > \\frac{s_{{\\rm max}}}{ 2 s_{{\\rm max}} + 1}$, the proof of relation (\\[reseau\\]) is then reduced to the proof of $$\\begin{aligned}\n\\lim_{n \\rightarrow \\infty} {\\rm Card} (\\Lambda'_{n^{-b}}) \\pi_j\\left(\n2\\sum_k (-\\delta_{k})u_{j}(n) B_{j,k,{\\tilde \\lambda},\\lambda_0}\\geq a^2\n(v_{nt_n^{-1}}(\\tau_j))^2\\right)=&0& \\nonumber \\\\\n\\lim_{n \\rightarrow \\infty} (T n^{b})^{d_{\\Lambda}} \\pi_j\\left(\n2\\sum_k (-\\delta_{k})u_{j}(n) B_{j,k,{\\tilde \\lambda},\\lambda_0}\\geq a^2\n(v_{nt_n^{-1}}(\\tau_j))^2\\right)&=&0, \\label{Bern-reseau}\\end{aligned}$$ where $\\mbox{\\rm {Diam}}$ is the diameter of $\\Lambda$. Finally, relation (\\[lambda’\\]) is proved applying Bernstein inequality in the right hand side of relation (\\[Bern-reseau\\]). ", "Indeed Bernstein inequality is applied to $$\\displaystyle{\\pi_j\\left( 2\\sum_k (-\\delta_{k})u_{j}(n)\nB_{j,k,{\\tilde \\lambda},\\lambda_0}\\geq a^2\n(v_{nt_n^{-1}}(\\tau_j))^2\\right)},$$ with the i.i.d. ", "centered random variables $Z_k= -\\delta_{k} B_{j,k,{\\tilde\n\\lambda},\\lambda_0}$. In particular, Notice that $|Z_k| <\nK_1v_{nt_n^{-1}}(\\tau_j)$, $\\sum_k \\V (Z_k) \\leq K_2\n(v_{nt_n^{-1}}(\\tau_j))^2$, where $K_1$ and $K_2$ are positive constants. ", "Notice also that it leads to an exponential bound of order $\\exp (- 2 ^{j})$.\n\nProof of Relation (\\[vrais\\])\n-----------------------------\n\nSet $$l_{n,\\pi}= \\displaystyle{\\int \\prod_{i=1}^n\n\\frac{c_j(U_i,V_i)}{c_{\\lambda_0}(U_i,V_i)} d\\pi_j(c_j)}.$$ Due to the fact that the functions $\\psi^{\\epsilon}_{j,k}$ have disjoint support, it is possible to rewrite $c_j$ as follows $$c_j=c_{\\lambda_0} \\prod_{k} (1 + \\delta_{k} D_{j,k})$$ for $$D_{j,k}= u_{j}(n)\n\\sum_{\\epsilon}\\frac{\\psi^{\\epsilon}_{j,k}}{c_{\\lambda_0}}.$$ Then, $$\\begin{aligned}\nl_{n,\\pi}& =& \\prod_k \\int \\prod_{i=1}^n (1 + \\delta_{k}\nD_{j,k}(U_i,V_i)) d\\pi_j (\\delta_{k})\\\\& =& \\prod_k \\frac 1 2\n\\left\\{\\prod_{i=1}^n (1 + D_{j,k}(U_i,V_i))+ \\prod_{i=1}^n (1 -\nD_{j,k}(U_i,V_i))\\right\\},\\end{aligned}$$ and $$\\begin{aligned}\nl_{n,\\pi}^2 & =& \\prod_k \\frac 1 4 \\left\\{2 \\prod_{i=1}^n \\left[1 +\nD^2_{j,k}(U_i,V_i)\\right]\n + 2 \\prod_{i=1}^n \\left[1 - D^2_{j,k}(U_i,V_i)\\right]\\right. \\\\", "\n &&\\left.+ H\\left(D_{j,k}(U_i,V_i),\\left(D^{b_t}_{j,k}(U_t,V_t)\n \\right)_{t \\in \\{1,\\ldots,i-1,i+1,\\ldots,n\\}}\\right)\\right\\},\\end{aligned}$$ where $b_t$ is either $0$ or $2$. Due to the independence of the data and acting as in [@Pouet:2000], it can be shown that $$\\E_{\\lambda_0}\\left [H\\left(D_{j,k}(U_i,V_i),\\left(D^{b_t}_{j,k}(U_t,V_t)\n \\right)_{t \\in \\{1,\\ldots,i-1,i+1,\\ldots,n\\}}\\right)\\right]=0.$$ Therefore, $$\\begin{aligned}\n\\E_{\\lambda_0} [l_{n,\\pi}^2(U_i,V_i)] & \\leq & \\prod_k \\left\\{\n\\left(1 + \\E_{\\lambda_0}D_{j,k}^2(U_i,V_i)\\right)^n\n+\\left(1 - \\E_{\\lambda_0}D_{j,k}^2(U_i,V_i)\\right)^n \\right\\}\\\\\n& \\leq& \\prod_k \\cosh \\left( n\n\\E_{\\lambda_0}D_{j,k}^2(U_i,V_i)\\right).\\end{aligned}$$ Using the inequality $\\log (\\cosh(u)) \\leq K u^2$ where $K$ is a fixed constant and since $c_{\\lambda_0}$ is bounded from below by $m$, one obtains $$\\begin{aligned}\n\\frac{1}{N_n^2} \\sum_{j \\in {\\tilde J}} \\exp(\\log (\\E_{\\lambda_0}\nl_{n,\\pi})^2 ) &\\leq & \\frac{1}{N_n^2} \\sum_{j \\in {\\tilde J}} \\exp \\left\\{ K\nn^2\n \\sum_k \\left( \\E_{\\lambda_0}D_{j,k}^2(U_i,V_i)\\right)^2\\right\\} \\\\\n&\\leq & \\frac{1}{N_n^2} \\sum_{j \\in {\\tilde J}} \\exp \\left\\{ \\frac{3^2 K}{m^2}n^2 2^{2j} u_{j}(n)^4\\right\\}\\\\\n&\\leq& \\frac{\\log(n)^{\\kappa}}{\\log(n)(1+o_n(1)) },\\end{aligned}$$ where $\\kappa = K (3 C_1^2 M^2)^{2} m^{-2}=4Ka^4m^{-2}$. Choosing $a$ small enough and $\\kappa<1$, Relation (\\[vrais\\]) is then proved.", "\n\nAppendix A: Proof of Lemma \\[etude-stat\\]\n=========================================\n\nIn this part, $\\kappa$ denotes any positive constant which may depend on $\\phi$, $M$ and on $\\|c\\|,\\|c_\\lambda\\|$.\n\nNotations and Preliminaries\n---------------------------\n\nLet us define or recall some notations that will be used below. ", "For any $k \\in \\BBz^2$, set $$\\begin{aligned}\n\\xi_k(X_i,Y_i) & =&\n\\phi_{j,k}\\left(\\widehat{F}(X_i),\\widehat{G}(Y_i) \\right)-\n\\phi_{j,k}\\left(F(X_{i}),G(Y_{i})\\right) \\label{xi} \\\\\n\\omega_{j,k}^\\lambda(X_{i},Y_{i}) & =& \\phi_{j,k}(F(X_{i}),\nG(Y_{i}))-c_{j,k}(\\lambda) \\nonumber \\\\\n \\omega_{j,k}^\\infty(X_{i},Y_{i}) & =& \\phi_{j,k}(F(X_{i}),\nG(Y_{i}))-c_{j,k}, \\nonumber\\end{aligned}$$ where $i$ is in ${\\cal I}_2$. First, the localization property of the scaling function implies that only few $\\xi_k(X_i,Y_i)$ will be used since the others are zero. ", "Indeed, one has the following result\n\n\\[nombredei\\]For any $k \\in \\BBz^2$, let us denote $$\\begin{aligned}\nN_{j}&=& \\mbox{ {\\rm card } }\\left\\{i \\in {\\cal I}_2; \\xi_k(X_i,Y_i)\\not=\n0\\right\\}\\label{ni}.\\end{aligned}$$ Let $\\delta>0$. For any level $j$ such that $$2^j\\leq\n\\frac{2}{3\\sqrt{\\delta+1}}\\,\\left(\\frac{n_2}{\\log(n_2)}\\right)^{1/2},$$ one has $$\\begin{aligned}\n\\label{nx}\n\\mathbb{P}(N_{j}>2(2L+3)n_22^{-j}))&\\leq\n&K(n_1^{-\\delta}+n_2^{-\\delta}).\\end{aligned}$$\n\nWe refer to [@Genest/Masiello/Tribouley:2008] for the proof of this lemma since a similar result is established with an estimate $\\widehat{F}$ built on the whole sample: it guarantees in particular that $\\widehat{F}(X_{(i:n)})=i/n$, where $X_{(i:n)}$ denotes the $i-$th (among $n$) order statistic. ", "In our case, the situation is different since $\\widehat{F}(X_{(i:n_2)})$ is based on the observations lying in the subsample whose indices are in ${\\cal I}_1$ whereas it is calculated in an observation lying in the subsample whose indices are in ${\\cal I}_2$ ; nevertheless, applying the Dvoretsky–Kiefer–Wolfovitz Inequality, the following deviation inequality holds. ", "For any $\\epsilon>0$, $\\mathbb{P}_{\\widehat{F}}=\\mathbb{P}\\left(\\left|\\widehat{F}(X_{(i:n_2)})-\\frac{i}{n_2}\\right|\\geq\n2\\epsilon\\right)$ is bounded from above by\n\n$$\\begin{aligned}\n\\mathbb{P}_{\\widehat{F}}&\\leq&\n\\mathbb{P}\\left(\\left|\\widehat{F}(X_{(i:n_2)})-{F}(X_{(i:n_2)})\\right|\\geq\n\\epsilon\\right)+\n\\mathbb{P}\\left(\\left|{F}(X_{(i:n_2)})-\\widetilde{F}(X_{(i:n_2)})\\right|\\geq\n\\epsilon\\right)\\\\\n&\\leq& \\mathbb{P}\\left(\\|\\widehat{F}-{F}\\|_\\infty\\geq\n\\epsilon\\right)+ \\mathbb{P}\\left(\\|\\widetilde{F}-{F}\\|_\\infty\\geq\n\\epsilon\\right)\\\\\n&\\leq& K\\left(n_1^{-\\delta}+n_2^{-\\delta}\\right),\\end{aligned}$$ as soon as we take $\\epsilon=\\sqrt{\\delta\\log(n_1)/(2n_1)}\\vee\n\\sqrt{\\delta\\log(n_2)/(2n_2)}$. Here $\\widehat{F}$ represents the empirical margin computed with the subsample whose indices in ${\\cal I}_1$ and $\\widetilde{F}$, the empirical margin computed with the subsample whose indices in ${\\cal I}_2$.\n\n### Study of $\\widehat{T_j}(\\lambda)$\n\nRewrite $\\widehat{\\theta_{j,k}}(\\lambda)$ in $\\widehat{T_j}(\\lambda)=\\sum_k \\widehat{\\theta_{j,k}}(\\lambda)$ as follows $$\\begin{aligned}\n\\widehat{\\theta_{j,k}}(\\lambda)&=&\n\\frac{1}{n_2(n_2-1)}\\mathop{\\sum_{i_1, i_2 \\in {\\cal I}_2}}_{ i_1 \\neq i_2\n}\\omega_{j,k}^\\lambda(X_{i_1},Y_{i_1}) \\omega_{j,k}^\\lambda(X_{i_2},Y_{i_2}).\\end{aligned}$$ For all $i \\in {\\cal I}_2$, one has $\\E\n(\\omega_{j,k}^\\lambda(X_{i},Y_{i}))= c_{j,k} -c_{j,k}(\\lambda)$, which implies that $$\\begin{aligned}\n\\E ( \\widehat{T_j}(\\lambda)) = \\sum_k \\theta_{j,k}(\\lambda)=\nT_j(\\lambda).\\end{aligned}$$ Moreover for $p \\neq k$, one obtains\n\n$\\E\n(\\widehat{\\theta_{j,k}}(\\lambda)\\widehat{\\theta_{j,p}}(\\lambda))$ [$$\\begin{aligned}\n&= & \\frac{1}{(n_2(n_2-1))^2 }\\sum_{i_1\\not =i_2\\not=i_3\\not=i_4} \\E\n\\left[\\omega_{j,k}^\\lambda(X_{i_1},Y_{i_1})\\right] \\E\n\\left[\\omega_{j,p}^\\lambda(X_{i_3},Y_{i_3})\\right]\n\\E \\left[\\omega_{j,k}^\\lambda(X_{i_2},Y_{i_2})\\right]\\E \\left[\\omega_{j,p}^\\lambda(X_{i_4},Y_{i_4})\\right]\\\\\n& & +4\\frac{1}{(n_2(n_2-1))^2 }\\sum_{i_1\\not =i_2\\not=i_3} \\E\n\\left[\\omega_{j,k}^\\lambda(X_{i_1},Y_{i_1})\\right] \\E\n\\left[\\omega_{j,p}^\\lambda(X_{i_3},Y_{i_3})\\right]\n\\E \\left[\\omega_{j,k}^\\lambda(X_{i_2},Y_{i_2})\\omega_{j,p}^\\lambda(X_{i_2},Y_{i_2})\\right]\\\\\n& & +2\\frac{1}{(n_2(n_2-1))^2 }\\sum_{i_1\\not =i_2} \\E\n\\left[\\omega_{j,k}^\\lambda(X_{i_1},Y_{i_1})\\omega_{j,p}^\\lambda(X_{i_1},Y_{i_1})\\right]\n\\E\n\\left[\\omega_{j,k}^\\lambda(X_{i_2},Y_{i_2})\\omega_{j,p}^\\lambda(X_{i_2},Y_{i_2})\\right]\\\\\n&\\leq &\\theta_{j,k}(\\lambda)\\theta_{j,p}(\\lambda)+\n\\frac{4}{n_2}\\,\\left(c_{j,k} -c_{j,k}(\\lambda)\\right)\\left(c_{j,p}\n-c_{j,p}(\\lambda)\\right) \\left[\\int\\left(\\phi_{j,k}-\\int\n\\phi_{j,k}c_\\lambda\\right)\\left(\\phi_{j,p}-\\int\n\\phi_{j,p}c_\\lambda\\right)c\\right]\\\\&&\\,+\n\\frac{2}{n_2(n_2-1)}\\left[\\int\\left(\\phi_{j,k}-\\int\n\\phi_{j,k}c_\\lambda\\right)\\left(\\phi_{j,p}-\\int\n\\phi_{j,p}c_\\lambda\\right)c\\right]^2,\\end{aligned}$$]{} which implies that [$$\\begin{aligned}\n\\V (\\widehat{T_j}(\\lambda))&=\n&\\E\\left(\\sum_{k}\\widehat{\\theta_{j,k}}(\\lambda)\\right)^2-\\left(\\E\\sum_{k}\\widehat{\\theta_{j,k}}(\\lambda)\\right)^2\n\\\\\n&\\leq& \\frac{4}{n_2}\\,\\sum_{k,p}\\left(c_{j,k}\n-c_{j,k}(\\lambda)\\right)\\left(c_{j,p} -c_{j,p}(\\lambda)\\right)\n\\left[\\int\\left(\\phi_{j,k}-\\int\n\\phi_{j,k}c_\\lambda\\right)\\left(\\phi_{j,p}-\\int\n\\phi_{j,p}c_\\lambda\\right)c\\right]\\\\&&\\,+\n\\frac{2}{n_2(n_2-1)}\\sum_{k,p}\\left[\\int\\left(\\phi_{j,k}-\\int\n\\phi_{j,k}c_\\lambda\\right)\\left(\\phi_{j,p}-\\int\n\\phi_{j,p}c_\\lambda\\right)c\\right]^2.", "\n \\end{aligned}$$]{} Applying the Hölder Inequality and the consequence of the Parseval Equality, we get\n\n[ $\\displaystyle{\n \\sum_{k\n p}\\left[\\int\\left(\\phi_{j,k}-\\int\n\\phi_{j,k}c_\\lambda\\right)\\left(\\phi_{j,p}-\\int\n\\phi_{j,p}c_\\lambda\\right)c\\right]^2}$ $$\\begin{aligned}\n& \\leq & 2^2 \\left(\n\\sum_{k,p}\\left[\\int\\phi_{j,k}\\phi_{j,p}c\\right]^2+\n2\\sum_{k,p}\\left[\\int\\phi_{j,k}c_\\lambda\\int\\phi_{j,p}c\\right]^2\n+\\left(\\sum_{k}\\left[\\int\\phi_{j,k}c_\\lambda\\right]^2\\right)^2\n\\right)\n\\\\\n&\\leq & 2^2 \\left( \\left( \\sum_{k} \\int\\phi_{j,k}^2c \\right)^2 +2\n\\int c^2 \\int c_\\lambda^2 +\\left( \\int c_\\lambda^2\\right)^2\n \\right)\\leq \\kappa \\,2^{2j}.\\end{aligned}$$]{} We conclude that $$\\begin{aligned}\n\\V (\\widehat{T_j}(\\lambda))&\\leq & \\kappa \\left(\n\\frac{4}{n_2}\\left(\\sum_{k,p}\\theta_{j,k}(\\lambda)\\theta_{j,p}\n(\\lambda)\\right)^{1/2}\\;2^{j}+\\frac{2^{2j}}{n_2(n_2-1)}\\right)\\\\\n&\\leq & \\kappa \\left( \\frac{2^j}{n_2}T_j(\\lambda)+\n\\,\\frac{2^{2j}}{n_2^2}\n \\right),\\end{aligned}$$ which is the announced result for $\\widehat{T_j}(\\lambda)$.\n\n### Study of $T_j^\\spadesuit $ and $T_j^{\\clubsuit }(\\lambda)$ \n\nLet us denote $$\\begin{aligned}\n A_{i_1}&=&\n\\left[\\xi_{k}(X_{i_1},Y_{i_1})\\right],\\quad D_{i_1}=\n\\sum_{k,p}\\left(\\E\n\\left[\\xi_{k}(X_{i_1},Y_{i_1}) \\xi_{p}(X_{i_1},Y_{i_1})\\right]\\right)^2\\\\\nB_{i_1,i_2}&=&\\sum_{k}\\left[\\xi_{k}(X_{i_1},Y_{i_1})\n\\xi_{k}(X_{i_2},Y_{i_2}\\right],\\; C_{i_1,i_2}=\n\\xi_{k}(X_{i_1},Y_{i_1}) \\xi_{p}(X_{i_2},Y_{i_2}).\\end{aligned}$$ We need the following results which are stated in the lemma below\n\n\\[horreur\\] Assume that the scaling function is $q$-differentiable. ", "For any level $j\\leq j_\\infty$, there exists some positive constant $\\kappa$ depending on $\\phi$, its derivatives and on $\\|c\\|_\\infty$ (which might be $\\|c_\\lambda\\|_\\infty$ for some $\\lambda \\in \\Lambda$) such that for any distinct indices $i_1,i_2$, one obtains $$\\begin{aligned}\n\\label{A}\n\\E |A_{i_1}|&\\leq &\\kappa\\left(\\frac{\\log(n_1)}{n_1}\\right)^{1/2}\\end{aligned}$$ $$\\begin{aligned}\n\\label{BCD}\n\\E |B_{i_1,i_2}| & \\leq & \\kappa\n2^{2j}\\left(\\frac{\\log(n_1)}{n_1}\\right),\\quad \\E |C_{i_1,i_2}|\\leq\n\\kappa \\left(\\frac{\\log(n_1)}{n_1}\\right)\\\\\n|D_{i_1}|& \\leq &\n2^{6j}\\left(\\frac{\\log(n_1)}{n_1}\\right)^2.\\nonumber\\end{aligned}$$\n\nWe prove relation (\\[A\\]) in the next section, relations (\\[BCD\\]) are proven in [@Genest/Masiello/Tribouley:2008]. ", "We have $$\\begin{aligned}\n\\E T_{j}^\\spadesuit &=&\\frac{1}{n_2(n_2-1)}\\mathop{\\sum_{i_1, i_2\n\\in {\\cal I}_2 }}_{ i_1 \\neq i_2 }\\E[B_{i_1,i_2}].\\end{aligned}$$ Using Lemma \\[nombredei\\] and Lemma \\[horreur\\], it follows $$\\begin{aligned}\n\\E |T_{j}^\\spadesuit | &\\leq &\\frac{1}{n_2(n_2-1)}\\,(n_22^{-j})^2\\,\n2^{2j}\\left(\\frac{\\log(n_1)}{n_1}\\right)\\leq\\left(\\frac{\\log(n_1)}{n_1}\\right). ", "\\end{aligned}$$ Moreover, we get $$\\begin{aligned}\n T_{j}^\\clubsuit (\\lambda) &=&\\frac{1}{n_2(n_2-1)}\\mathop{\\sum_{i_1,\ni_2 \\in {\\cal I}_2}}_{ i_1 \\neq i_2 }\\sum_k\n\\left[\\xi_k(X_{i_1},Y_{i_1})\\omega_{j,k}^\\lambda(X_{i_2},Y_{i_2})\\right].\\end{aligned}$$ By Hölder Inequality and from lemmas \\[nombredei\\] and \\[horreur\\], one obtains $$\\begin{aligned}\n\\E| T_{j}^\\clubsuit (\\lambda)|&\\leq\n&\\frac{1}{n_2(n_2-1)}\\mathop{\\sum^{n_2}_{i_1, i_2 \\in {\\cal I}_2}}_{\ni_1 \\neq i_2 }\\left(\\sum_k\\,(\\E(A_i))^2\\;\\sum_k(\n\\E\\omega_{j,k}^\\lambda(X_{i_2},Y_{i_2}))^2\\right)^{1/2}.\\end{aligned}$$ Remembering that $\\E\n\\omega_{j,k}^\\lambda(X_{i_2},Y_{i_2})=(c_{j,k}-c_{j,k}(\\lambda))$ for any index $i_2$, we get $$\\begin{aligned}\n\\E| T_{j}^\\clubsuit (\\lambda)|&\\leq&\n\\frac{1}{n_2(n_2-1)}\\,(n_22^{-j})n_2\\,\n\\left[2^{2j}\\frac{\\log(n_1)}{n_1}\\,T_j(\\lambda)\\right]^{1/2}\\\\\n&\\leq&K\\,\\left(\\frac{\\log(n_1)}{n_1}\\,T_j(\\lambda)\\right)^{1/2}.\\end{aligned}$$ Let us study the moments of $T_{j}^\\clubsuit (\\lambda)$ under $\\P_\\lambda$. Since $\\E_\\lambda\\omega_{j,k}^\\lambda(X_i,Y_i)=0$ for any $k$ and $i$, we obviously have $\\E_\\lambda T_{j}^\\clubsuit\n(\\lambda)=0$ and $$\\E_\\lambda (T_{j}^\\clubsuit (\\lambda))^2=\\left(\\frac{1}{n_2(n_2-1)}\\right)^2\\sum_{i_1\\not=i_2\n}T_{i_1,i_2}+\\left(\\frac{1}{n_2(n_2-1)}\\right)^2\\sum_{i_1\\neq i_2\\neq\ni_3 }S_{i_1,i_2,i3},$$ where $$\\begin{aligned}\nT_{i_1,i_2}&=&\\sum_{k,p}\\left(\\E_\\lambda\\left[\\xi_k(X_{i_1},Y_{i_1})\\xi_p(X_{i_1},Y_{i_1})\\right]\n\\E_\\lambda \\left[\\omega_{j,k}^\\lambda(X_{i_2},Y_{i_2})\\omega_{j,p}^\\lambda(X_{i_2},Y_{i_2})\\right]\\right),\\\\\nS_{i_1,i_2,i_3}&=& \\sum_{k,p}\\left(\\E_\\lambda\n\\left[\\xi_{k}(X_{i_1},Y_{i_1})\n\\xi_{p}(X_{i_2},Y_{i_2})\\right]\\E_\\lambda\\left[\\omega_{j,k}^\\lambda(X_{i_3},Y_{i_3})\\omega_{j,p}^\\lambda(X_{i_3},Y_{i_3})\\right]\\right).\\end{aligned}$$ By Hölder Inequality, we have $$\\begin{aligned}\nT_{i_1,i_2}&=&\\sum_{k,p}\\E_\\lambda\n\\left[\\xi_k(X_{i_1},Y_{i_1})\\xi_p(X_{i_1},Y_{i_1})\\right] \\E_\\lambda\n\\left[\\omega_{j,k}^\\lambda(X_{i_2},Y_{i_2})\\omega_{j,p}^\\lambda(X_{i_2},Y_{i_2})\\right]\\\\&\\leq&D_{i_1}^{1/2}\n\\left(\\sum_{k,p}\\left( \\E_\\lambda\n\\left[\\omega_{j,k}^\\lambda(X_{i_2},Y_{i_2})\\omega_{j,p}^\\lambda(X_{i_2},Y_{i_2})\\right]\\right)^2\\right)^{1/2}\\end{aligned}$$ With Parseval Equality, we get $$\\begin{aligned}\n\\sum_{k,p}\\left(\\E_\\lambda\n\\left[\\omega_{j,k}^\\lambda(X_{i_2},Y_{i_2})\\omega_{j,p}^\\lambda(X_{i_2},Y_{i_2})\\right]\\right)^2&\\leq&\n\\sum_{k,p}\\left(\\int\\phi_{j,k}\\phi_{j,p}c_\\lambda\\right)^2\\\\\n&\\leq& K\\,\\sum_{k}\\int\\phi^2_{j,k}c_\\lambda^2\\leq K\\,2^{2j},\\end{aligned}$$ which combining with Lemma \\[horreur\\], implies that $$\\begin{aligned}\nT_{i_1,i_2}&\\leq&K\\,\\left(2^{6j}\\left(\\frac{\\log(n_1)}{n_1}\\right)^2\\right)^{1/2}\n\\left(2^{2j}\\right)^{1/2}\\leq\n2^{4j}\\left(\\frac{\\log(n_1)}{n_1}\\right).\\end{aligned}$$ In the same way, $$\\begin{aligned}\nS_{i_1,i_2,i_3}&\\leq &K\\,\\left(\\sum_{k,p}(\\E_\\lambda\nC_{i_1,i_2})^2\\right)^{1/2} \\left(\\sum_{k,p}\\left(\n\\E_\\lambda\\left[\\omega_{j,k}^\\lambda(X_{i_2},Y_{i_2})\\omega_{j,p}^\\lambda(X_{i_2},Y_{i_2})\\right]\\right)^2\\right)^{1/2}\n\\\\&\\leq&\n\\left(2^{2j}\\right)^{1/2}\\left(2^{4j}\\left(\\frac{\\log(n_1)}{n_1}\\right)^2\\right)^{1/2}\n\\leq 2^{3j}\\left(\\frac{\\log(n_1)}{n_1}\\right).\\end{aligned}$$ From Lemma \\[nombredei\\], one has $$\\begin{aligned}\n\\E_\\lambda( T_{j}^\\clubsuit (\\lambda))^2&\\leq&\nK\\;\\frac{1}{n_2^2(n_2-1)^2}(n_22^{-j})n_22^{4j}\\left(\\frac{\\log(n_1)}{n_1}\\right)\\\\\n&&+\nK\\;\\frac{1}{n_2^2(n_2-1)^2}(n_22^{-j})^2n_22^{3j}\\left(\\frac{\\log(n_1)}{n_1}\\right)\\\\\n&\\leq&K\\,2^{j}\\left(\\frac{\\log(n_1)}{n_2n_1}\\right).\\end{aligned}$$\n\n Proof of Lemma \\[horreur\\]\n---------------------------\n\nThe following expansion is crucial because it allows to reduce the study to univariate variables. ", "$$\\begin{aligned}\n\\label{biuni}\n&& \\xi_{k}(X_i,Y_i)= \\xi_{k_1}(X_i)\\xi_{k_2}(Y_i)\\\\&&\\hspace{2cm}+\n \\xi_{k_1}(X_i)\\phi_{jk_2}\\left(G(Y_i)\\right)\n+\n \\xi_{k_2}(Y_i)\\phi_{jk_1}\\left(F(X_i)\\right)\n,\\nonumber \\end{aligned}$$ where the univariate statistics $\\xi_{k_1}(X_i)$ and $\\xi_{k_2}(Y_i)$ are defined as follows $$\\begin{aligned}\n\\label{xiuni}\\xi_{k_1}(X_i)&=&\\phi_{j,k_1}\\left(\\frac{\\widehat{F}(X_i)}{n_1}\\right)-\\phi_{j,k_1}(F(X_i))\\\\\n\\xi_{k_2}(Y_i)&=&\\phi_{j,k_2}\\left(\\frac{\\widehat{G}(Y_i)}{n_1}\\right)-\\phi_{j,k_1}(G(Y_i)).\\end{aligned}$$ Assuming that $\\phi$ is continuously $q-$differentiable, we get $$\\begin{aligned}\n\\label{taylor}\\xi_{k_1}(X_i)&=&\\hat z_{k_1}(X_i)+\\hat\nw_{k_1}(X_i),\\end{aligned}$$ where $$\\begin{aligned}\n \\hat\nz_{k_1}(X_i)&=&\\sum_{\\ell=1}^{q-1}\\frac{2^{j\\ell}}{\\ell!}(\\widehat{F}(X_i)-F(X_i))^\\ell\\;\\phi^{(\\ell)}_{j,k_1}(F(X_i))\\end{aligned}$$ and $$\\begin{aligned}\n\\hat w_{k_1}(X_i)&=&2^{qj}\\int_{\\widehat{ F}(X_i)}^{F(X_i)}\n\\phi^{(q)}_{j,k_1}(t)\\;( F(X_i)-t)^{q-1}dt.\\end{aligned}$$ A direct application of the Dvoretsky, Kiefer and Wolfovitz Inequality leads to the following bound $$\\begin{aligned}\n\\label{DKW}\n\\P(\\|\\widehat{ F}-F\\|_\\infty\n>\\epsilon)\\leq K\\,\\exp(-2n_1\\epsilon^2)\\leq Kn_1^{-\\delta},\\end{aligned}$$ as soon as $\\epsilon= \\sqrt{0.5\\;\\delta\\log(n_1)/n_1}$. In the sequel, we take such an $\\epsilon$ with $\\delta$ large enough. ", "Since $j\\leq j_\\infty$ where $j_\\infty$ is defined in (\\[j0jinfty\\]), observe that $2^j\\epsilon\\leq 1$ and then we get $$\\begin{aligned}\n|\\hat z_{k_1}(X_i)|&\\leq &K\\,2^{j}\\epsilon\\max_{\\ell=1,\\ldots\nq-1}\\,|\\phi^{(\\ell)}_{j,k_1}(F(X_i))|(1+o_P(1))\\\\\n|\\hat w_{k_1}(X_i)|&\\leq & K\\,2^{(q+1/2)j}\\epsilon^{q}(1+o_P(1))\\end{aligned}$$ which leads to the following bound [$$\\begin{aligned}\n\\label{xibound1}|\\xi_{k_1}(X_i)|&\\leq &K\\left(\n\\;2^{(q+1/2)j}\\epsilon^q+2^j\\epsilon\\max_{\\ell=1,\\ldots q-1}\n\\,|\\phi^{(\\ell)}_{j,k_1}(F(X_i))|\\right)(1+o_P(1)).\\end{aligned}$$]{} The same kind of result obviously holds for $\\xi_{k_2}(Y_i)$. In the sequel, we need the following evaluations (which also hold for any derivatives of $\\phi$). ", "Using expansion (\\[biuni\\]), we get $$\\begin{aligned}\n \\xi_k(X_i,Y_i)&=&S_1+S_2,\\end{aligned}$$ where $$\\begin{aligned}\nS_1&=& \\xi_{k_1}(X_i)\\xi_{k_2}(Y_i), \\\\S_2&=&\n\\xi_{k_1}(X_i)\\phi_{j,k_2}\\left(G(Y_i)\\right)+\n\\xi_{k_2}(Y_i)\\phi_{j,k_1}\\left(F(X_i)\\right).", "\n \\end{aligned}$$ Using (\\[xibound1\\]), we get $$\\begin{aligned}\n\\E |S_1|&\\leq &K\\,\\left( 2^{(2q+1)j}\\epsilon^{2q}+2^{(q+1)j}\\epsilon^{q+1}+2^j\\epsilon^2\\right),\\\\\n\\E |S_2|&\\leq&K\\,\\left(2^{qj}\\epsilon^{q}+\\epsilon\\right).", "\n \\end{aligned}$$ If $2^j\\leq (n_1/\\log(n_1))^{1/2-1/2q}$, we obtain $\n\\E|\\xi_k(X_i,Y_i)|\\leq\\epsilon$ which ends the proof.", "\n\nAppendix B : Proof of Lemma \\[Bernstein\\]\n=========================================\n\nLet us denote $T_{j}^\\diamond(\\lambda)=n_2^{-1}\\sum_{i \\in {\\cal I}_2} Z_i$ where $$\\begin{aligned}\nZ_i&=&\\sum_k\\left( \\phi_{jk}(F(X_i),G(Y_i))-c_{jk}\\right)\\left(\nc_{jk}-c_{jk}(\\lambda)\\right),\\\\\n\\E_cZ_i&=&0,\\\\\n |Z_i|&\\leq&\\left(\\sum_k\\left(\n\\phi_{jk}(F(X_i),G(Y_i))-c_{jk}\\right)^2 \\sum_k \\left(\nc_{jk}-c_{jk}(\\lambda)\\right)^2\\right)^{1/2}\\\\\n&\\leq&K_1\\,\\left(\\left( 2^j\\right)^2 T_j(\\lambda)\\right)^{1/2}\\leq\nK\\,2^{j}\nT_j(\\lambda)^{1/2},\\end{aligned}$$ and $$\\begin{aligned}\n \\V_c(Z_i)&\\leq& \\sum_{k,p} \\E\\left(\n\\phi_{jk}(F(X_i),G(Y_i))-c_{jk}\\right)\\left(\\phi_{jp}(F(X_i),G(Y_i))-c_{jp}\\right)\\\\\n&&\\hspace{1 cm}\\times \\left|\\left(\nc_{jk}-c_{jk}(\\lambda)\\right)\\left(\nc_{jp}-c_{jp}(\\lambda)\\right)\\right|\\\\\n&\\leq& \\sum_{k,p}\\left( \\E\n\\phi_{jk}^2(F(X_i),G(Y_i)) \\E\\phi_{jp}^2(F(X_i),G(Y_i))\\right)^{1/2}\\\\\n&&\\hspace{1 cm}\\times \\left|\\left(\nc_{jk}-c_{jk}(\\lambda)\\right)\\left(\nc_{jp}-c_{jp}(\\lambda)\\right)\\right|\\\\\n&\\leq& \\|c\\|_\\infty\\,\\;\\sum_{k}\\left(\nc_{jk}-c_{jk}(\\lambda)\\right)^2=\\|c\\|_\\infty\\, T_j(\\lambda).\\end{aligned}$$ Applying Bernstein Inequality to the $Z_i'$ s leads to prove Lemma \\[Bernstein\\].", "\n\nAppendix C: Proof of Lemma \\[LDUstat\\]\n======================================\n\n$U$-Statistic\n-------------\n\nLet us first recall the result of [@Gine/Latala/Zinn:2000].", "\n\n[(Theorem 3.3 p. 21 [@Gine/Latala/Zinn:2000])]{}\\\nIt exists an universal positive constant $K_g < \\infty$ such that, if $\\Omega$ is a bounded canonical kernel of two variables for the i.i.d. ", "$Z_{i_1},Z_{i_2}$, $i_1, \\; i_2 \\in \\{1,\\ldots,{\\tilde n}\\}$, where ${\\tilde n} \\in \\BBn $, for any $x>0$, we have $$\\begin{aligned}\n\\P \\left( |\\sum_{i_1,i_2} \\Omega(Z_{i_1},Z_{i_2})| > x \\right) \\leq\nK_g \\exp \\left( - \\frac 1 K_g \\; \\min \\left\\{\\frac{x^2}{C^2},\n\\frac{x}{D},\\left(\\frac{x}{B}\\right)^{2/3},\n\\left(\\frac{x}{A}\\right)^{1/2} \\right\\}\\right), \\label{Gine}\\end{aligned}$$ where $$\\begin{aligned}\nA &= &\\|\\Omega (\\cdot,\\cdot)\\|_{\\infty},\\, B^2= \\tilde n \\left[ \\|\n\\E[ \\Omega^2(Z_1,\\cdot)] \\|_{\\infty}+ \\| \\E [\\Omega^2(\\cdot,Z_2)]\n\\|_{\\infty} \\right],\\\\\n C^2& =& \\tilde n^2 \\E [\n(\\Omega(Z_1,Z_2))^2] \\end{aligned}$$ and $$\\begin{aligned}\nD= \\tilde n \\sup_{\\Omega_1,\\Omega_2} \\{ \\E[\\Omega(Z_{1},Z_{2})\n\\Omega_1(Z_{1}) \\Omega_2(Z_{2})]:\n \\E [\\Omega_1^2(Z_{1})]\\leq 1; \\E [\\Omega_2^2(Z_2)] \\leq 1 \\}.\\end{aligned}$$\n\nWe apply this proposition for $Z_i=(F(X_i),G(Y_i))$, $\\tilde n=n_2$ and the kernel $$\\begin{aligned}\n\\Omega_{{\\tilde c}}\\left( Z_{i_1}, Z_{i_2}\\right) & =& \\sum_k \\left\\{\n\\phi_{j,k}\\left(Z_{i_1} \\right) - \\E_{{\\tilde c}}[ \\phi_{j,k}\\left(Z_{i_1}\n\\right)]\\right\\} \\times \\left\\{ \\phi_{j,k}\\left(Z_{i_2} \\right) -\n\\E_{{\\tilde c}}[ \\phi_{j,k}\\left(Z_{i_2} \\right)]\\right\\},\\end{aligned}$$ which is considered under the distribution $\\P_{{\\tilde c}}$ where ${\\tilde c}$ is either $c_\\lambda$ or $c$. The quantities $A$, $B$, $C$ and $D$ are evaluated in the following lemma which is proved in the next section.", "\n\n\\[ABCDBornes\\] There exists some positive constant $K_1$ larger than either $$\\left(12 L^2\\|\\phi\\|_{\\infty}^2\\right) \\vee \\left(\n2\\|{\\tilde c}\\|_\\infty\\right) \\vee \\left( 2 L^2\n\\|\\phi\\|_{\\infty}^2\\right)\\vee \\left(4\n\\|{\\tilde c}\\|_\\infty(\\|{\\tilde c}\\|_\\infty+3L^4\\|\\phi\\|_\\infty^2))\\right)$$ such that $$A \\leq K_1\n2^{2j},\\;B^2 \\leq K_1 \\,n_2 2^{2j},\\;C^2 \\leq K_1\\,n_2^2 2^{2j},\\;D \\leq\nK_1\\, n_2 ,$$ where ${\\tilde c}$ is either $c_\\lambda$ or $c$.\n\nAgain define ${\\tilde c}$ as $c_\\lambda$ or $c$, then applying both the result of [@Gine/Latala/Zinn:2000] and Lemma \\[ABCDBornes\\], for any level $j$ and any $x\\geq\n2^j((n_2-1)n_2)^{-1/2}\\sqrt{\\log(\\log(n_2))}$, it immediately follows that [$$\\begin{aligned}\n\\P_{{\\tilde c}}\\left(|\\frac{1}{n_2(n_2-1)}\\mathop{\\sum_{i_1, i_2 \\in {\\cal\nI}_2}}_{ i_1 \\neq i_2 } \\Omega_{{\\tilde c}}\\left( Z_{i_1}, Z_{i_2}\\right)|>\n\\mu x \\right)&\\leq& K_g\\exp\\left(-\\delta\\,\\log(\\log(n_2))\\right).\\end{aligned}$$]{} which ends the proof of Lemma \\[LDUstat\\].", "\n\nProof of Lemma \\[ABCDBornes\\]\n-----------------------------\n\nLet us denote $(U,V)=(F(X),G(Y))$ any pair of random variables whose marginal distribution are both uniform on $[0,1]$. Denote ${\\tilde c}$ the copula density which is $c_\\lambda$ or $c$; in the same spirit, the coefficients ${\\tilde c}_{j,k}$ stand for $c_{j,k}(\\lambda)$ or $c_{j,k}$. Recall that $$\\begin{aligned}\nc_{j,k}(\\lambda)& =& \\E_{\\lambda} \\left[ \\phi_{j,k}(F(X),G(Y))\n\\right] =\\int c_\\lambda (u,v )\\phi_{j,k}(u,v)du dv \\nonumber.\\\\\nc_{j,k}& =& \\E \\left[ \\phi_{j,k}(F(X),G(Y)) \\right] = \\int c (u,v\n)\\phi_{j,k}(u,v)du dv \\nonumber.", "\n$$ Notice that $$\\begin{aligned}\n\\sum_{k,p} \\E_{{\\tilde c}} [\n\\phi_{j,k}(U_{i_1},V_{i_1})\\phi_{j,p}(U_{i_1},V_{i_1})\n]& \\leq & 2^{2j}, \\\\\n\\sum_k (\\E_{{\\tilde c}}[ \\phi_{j,k}(U,V )])^2 & =& \\sum_k {\\tilde c}_{j,k}^2 \\leq \\|{\\tilde c}\\|^2\n\\leq M.\\end{aligned}$$ We get $$\\begin{aligned}\n A & =& \\| \\sum_{k}\\left(\\phi_{j,k}(u_1,v_1) - \\E_{{\\tilde c}}[\n\\phi_{j,k}(U,V )] \\right) \\left(\\phi_{j,k}(u_2,v_2) - \\E_{{\\tilde c}}[\n\\phi_{j,k}(U,V )] \\right) \\|_{\\infty} \\\\\n&\\leq & \\| \\sum_{k} \\phi_{j,k}(u_1,v_1)\\phi_{j,k}\n(u_2,v_2)\\|_{\\infty} + 2 \\|\n \\sum_{k} \\phi_{j,k}(u_1,v_1) \\E_{{\\tilde c}}[ \\phi_{j,k}(U,V )] \\|_{\\infty}\\\\&&+\n \\|\n \\sum_{k} \\left(\\E_{{\\tilde c}}[ \\phi_{j,k}(U,V )] \\right)^2\\|_\\infty\n \\\\\n& \\leq & L^2\\,2^{2j} \\|\\phi\\|_{\\infty}^2 +\n2L^2\\|\\phi\\|_{\\infty}\\|{\\tilde c}\\|_2 2^{j} + \\|{\\tilde c}\\|_2^2 \\leq K 2^{2j},\\end{aligned}$$ where $K\\geq 2 L^2 \\|\\phi\\|_{\\infty}^2 $ and $$\\begin{aligned}\nB^2& =& 2 n_2 \\left \\| \\sum_{k,p} \\E_{\\tilde c} \\left[\n\\left(\\phi_{j,k}(U_{i_1},V_{i_1}) - \\E_{\\tilde c}[ \\phi_{j,k}(U,V )] \\right)\n\\left(\\phi_{j,p}(U_{i_1},V_{i_1}) - \\E_{\\tilde c}[ \\phi_{j,p}(U,V) ] \\right)\n\\right]\\right. ", " \\\\\n& & \\times \\left.\\left(\\phi_{j,k}(u_2,v_2) - \\E_{\\tilde c}[ \\phi_{j,k}(U,V )]\n\\right) \\left(\\phi_{j,p}(u_2,v_2) - \\E_{\\tilde c}[ \\phi_{j,p}(U,V) ]\n\\right)\\right\\|_{\\infty} \\\\\n& \\leq & 2n_2 \\left \\| \\sum_{k,p} \\left|\\int \\phi_{j,k} \\phi_{j,p}\n{\\tilde c} -\\int \\phi_{j,k}\\,{\\tilde c}\\int \\phi_{j,p}\\,{\\tilde c} \\right|\n\\left(\\phi_{j,k}(u_2,v_2) - \\E_{\\tilde c}[\n\\phi_{j,k}(U,V)] \\right) \\right.\\\\\n& & \\times \\left. ", " \\left(\\phi_{j,p}(u_2,v_2) - \\E_{\\tilde c}[ \\phi_{j,p}(U,V)]\n\\right)\\right\\|_{\\infty} \\\\\n& \\leq & 2n_2 (2\\|{\\tilde c}\\|_\\infty) \\left[\\left \\| \\sum_{k,p}\n\\phi_{j,k}(u_2,v_2)\\phi_{j,p}(u_2,z_2)\\right\\|_{\\infty} + 2\\left \\|\n\\sum_{k,p} \\phi_{j,k}(u_2,v_2)\\E_{\\tilde c}[ \\phi_{j,k}(U,V\n)]\\right\\|_{\\infty}\\right.\\\\&&\\left. ", "+ \\left \\| \\sum_{k,p}\\E_{\\tilde c}[\n\\phi_{j,k}(U,V )]\\E_{\\tilde c}[ \\phi_{j,p}(U,V )]\\right\\|_{\\infty}\n\\right] \\\\\n& \\leq & (4 n_2\\|{\\tilde c}\\|_\\infty) \\left( 2^{2j}L^4 2\n\\|\\phi\\|_\\infty^2+2L^22^{j}\n \\|\\phi\\|_{\\infty}+2^{2j}\\|{\\tilde c}\\|_\\infty\\right)\n \\leq K\\,n_2 2^{2j}\\end{aligned}$$ where $K\\geq 4 \\|{\\tilde c}\\|_\\infty(\\|{\\tilde\nc}\\|_\\infty+3L^4\\|\\phi\\|_\\infty^2)$. Moreover, $$\\begin{aligned}\nC^2 & =&n_2^2 \\sum_{k,p} \\E_{\\tilde c} \\left[ \\left(\\phi_{j,k}(U_{i_1},V_{i_1}) -\n\\E_{\\tilde c}[ \\phi_{j,k}(U,V )] \\right) \\left(\\phi_{j,p}(U_{i_1},V_{i_1}) -\n\\E_{\\tilde c}[\n\\phi_{j,p}(U,V) ] \\right) \\right] \\\\\n & & \\times \\E_{\\tilde c} \\left[\n\\left(\\phi_{j,k}(U_{i_2},V_{i_2}) - \\E_{\\tilde c}[ \\phi_{j,k}(U,V )] \\right)\n\\left(\\phi_{j,p}(U_{i_2},V_{i_2}) - \\E_{\\tilde c}[ \\phi_{j,p}(U,V) ] \\right)\n\\right] \\\\&&\\\\ & =& n_2^2 \\sum_{k,p}\\left(\\E_{\\tilde c} \\left[\n\\phi_{j,k}(U_{i_1},V_{i_1})\\phi_{j,p}(U_{i_1},V_{i_1})\\right]-\\E_{\\tilde c}\n\\left[\\phi_{j,k}(U,V)\\right]\\E_{\\tilde c}\\left[\\phi_{j,p}(U,V)\\right]\\right)^2\\\\\n&=&n_2^2 \\sum_{k,p}\\left(\n\\int \\phi_{j,k} \\phi_{j,p}\\, {\\tilde c} - \\int \\phi_{j,k}\\,{\\tilde c} \\int \\phi_{j,p}\\,{\\tilde c} \\right)^2\\\\\n&\\leq &n_2^2 \\sum_{k,p}\\left(\n\\int \\phi_{j,k} \\phi_{j,p}\\, {\\tilde c} \\right)^2+n_2^2 \\left(\\sum_{k}\\left(\\int \\phi_{j,k}\\,{\\tilde c} \\right)^2\\right)^2\\\\\n&\\leq &n_2^2 \\sum_{k}\n\\int \\phi_{j,k}^2 \\, {\\tilde c}^2 +n_2^2 \\left(\\int {\\tilde c}^2\\right)^2\\\\\n& \\leq & \\|{\\tilde c}\\|_\\infty^2n_2^2 2^{2j} \\,+ n_2^2\\,\\|{\\tilde c}\\|_2^4\\leq K \\,n_2^2\n2^{2j},\\end{aligned}$$ where $K\\geq 2\\|{\\tilde c}\\|_\\infty^2$. Denote $u_{\\Omega_1,\\Omega_2}=\\E_{\\tilde c}[\\Omega_{\\tilde c}(Z_{1},Z_{2})\n\\Omega_{1,{\\tilde c}}(Z_{1})\n \\Omega_{2,{\\tilde c}}(Z_{2})]$ and for $i=1,\\;2$, put $$c_i(k)=\\int (\\phi_{j,k} -\\E{\\tilde c}\n\\phi_{j,k} (U,V)) \\Omega_{i,{\\tilde c}}\\, {\\tilde c}.$$ By Hölder Inequality, we get $$\\begin{aligned}\nu_{\\Omega_1,\\Omega_2} & = & \\sum_k \\left(\\int (\\phi_{j,k} -\\E_{\\tilde c}\n\\phi_{j,k} (U,V))\n\\Omega_{1,{\\tilde c}}\\, {\\tilde c}\\right)\\left(\\int (\\phi_{j,k} -\\E_{\\tilde c} \\phi_{j,k} (U,V)) \\Omega_{2,{\\tilde c}}\\,{\\tilde c}\\right) \\nonumber \\\\\n & \\leq & \\sqrt{\\sum_k (c_1(k))^2 \\sum_k (c_2(k))^2 } \\label{FirstD}.", "\n$$ Applying again the inequality of Hölder to $ \\sum_k (c_1(k))^2$ (the same occurs for $c_2(k)$), one gets $$\\begin{aligned}\n\\sum_k (c_1(k))^2 \\!\\!\\!\\!& ", "\\leq & \\!\\!\\!\\sum_k (\\int (\\phi_{j,k} -\\E_{\\tilde c}\n\\phi_{j,k} (U,V))\n\\Omega_{1,{\\tilde c}}\\, {\\tilde c} \\I_{[\\frac{k_1}{2^{j}}, \\frac{2N-1+k_1}{2^{j}}] \\times [\\frac{k_2}{2^{j}}, \\frac{2N-1+k_2}{2^{j}}]})^2 \\nonumber \\\\\n& \\leq & \\sum_k \\left( \\int (\\phi_{j,k} -\\E_{\\tilde c}\n\\phi_{j,k} (U,V))^2 {\\tilde c} \\right) \\times \\nonumber \\\\\n& & \\hspace{0.5cm}\\left(\\int (\\Omega_{1, {\\tilde c}})^2\\, \\I_{[\\frac{k_1}{2^{j}}, \\frac{2L-1+k_1}{2^{j}}] \\times [\\frac{k_2}{2^{j}}, \\frac{2L-1+k_2}{2^{j}}]} \\; {\\tilde c} \\right) \\nonumber \\\\\n& \\leq & \\|\\tilde c\\|_\\infty \\int (\\Omega_{1, {\\tilde c}})^2\\, {\\tilde c} \\sum_k \\I_{[\\frac{k_1}{2^{j}}, \\frac{2L-1+k_1}{2^{j}}] \\times [\\frac{k_2}{2^{j}}, \\frac{2L-1+k_2}{2^{j}}]}\\nonumber \\\\\n\\label{SecondD} \\!\\!\\!\\! & ", "\\leq & \\!\\!\\!\\! ", "12 \\|\\phi\\|_{\\infty}^2\nL^2,\\end{aligned}$$ since $\\E_{\\tilde c} (\\Omega_{1, {\\tilde c}}(U))^2) \\leq 1$. It follows that $D \\leq K\\,n_2$, where $K > 12 L^2\\|\\phi\\|_{\\infty}^2$.\n\n ------------- ------------- ----------------------------------- ----------------------------------- ----------------------------------\n \n Copula True copula $\\tau=0.25$ $\\tau=0.50$ $\\tau=0.75$\n under $H_0$ \n \n \n Gumbel Clayton 1.00 [*0.0000*]{} (0.72) 1.00 [*0.0000*]{} (0.99) 1.00 [*0.0000*]{} (1.00)\n Gumbel [**0.05 [*0.0105* ]{} (0.05)**]{} [**0.01 0.0049 (0.05)**]{} [**0.00 [*0.0000*]{} (0.05)**]{}\n Frank 0.31 [*0.0207* ]{} (0.15) 0.36 [*0.0214* ]{} (0.40 ) 0.24 [*0.0191*]{} (0.84)\n Normal 0.25 [*0.0195*]{} (0.10) 0.25 [*0.0194* ]{} (0.18) 0.09 [*0.0127*]{} (0.61)\n Student(4) 0.99 [*0.0035*]{} (0.14) 0.89 [*0.0139* ]{} (0.22) 0.62 [*0.0218*]{} (0.55)\n \n \n Clayton Clayton [**0.05 [*0.0101*]{} (0.05)**]{} [**0.51 [*0.0224*]{} (0.05)**]{} [**1.00 [*0.0000*]{} (0.05)**]{}\n Gumbel 0.99 [*0.0028*]{} (0.86) 1.00 [*0.0000*]{} (1.00) 1.00 [*0.0000*]{} (1.00)\n Frank 0.97 [*0.0079*]{} (0.56) 1.00 [*0.0000*]{} (0.96) 1.00 [*0.0000*]{} (1.00)\n Normal 0.77 [*0.0190*]{} (0.50) 0.85 [*0.0178*]{} (0.93) 1.00 [*0.0000*]{} (1.00)\n Student(4) 0.45 [*0.0223*]{} (0.56) 0.14 [*0.0157*]{} (0.95) 1.00 [*0.0000*]{} (1.00)\n \n \n Frank Clayton 0.99 [*0.0022*]{} (0.40) 1.00 [*0.0000*]{} (0.89) 1.00 [*0.000*]{} (0.97)\n Gumbel 0.19 [*0.0175*]{} (0.33) 0.23 [*0.0190*]{} (0.63) 0.22 [*0.0184*]{} (0.82)\n Frank [**0.05 [*0.0108* ]{} (0.05)**]{} [**0.01 [*0.0035* ]{} (0.05)**]{} [**0.00 [*0.000*]{} (0.05)**]{}\n Normal 0.14 [*0.0155*]{} (0.08) 0.39 [*0.0218*]{} (0.20) 0.72 [*0.0201*]{} (0.41)\n Student(4) 0.95 [*0.0096*]{} (0.18) 0.83 [*0.0167*]{} (0.08) 0.92 [*0.0121*]{} (0.06)\n \n \n Normal Clayton 0.97 [*0.0076*]{} (0.31) 1.00 [*0.0000*]{} (0.80) 1.00 [*0.0000*]{} (0.92)\n Gumbel 0.19 [*0.0176*]{} (0.24) 0.13 [*0.0151*]{} (0.38) 0.01 [*0.0040*]{} (0.38)\n Frank 0.23 [*0.0190*]{} (0.08) 0.35 [*0.0214*]{} (0.20) 0.60 [*0.0219*]{} (0.42)\n Normal [**0.05 [*0.0099*]{} (0.05)**]{} [**0.01 [*0.0045*]{} (0.05)**]{} [**0.00 [*0.0000*]{} (0.05)**]{}\n Student(4) 0.87 [*0.0149*]{} (0.10) 0.22 [*0.0185*]{} (0.08) 0.08 [*0.0120*]{} (0.06)\n \n \n Student(4) Clayton 0.71 [*0.0204*]{} (0.27) 1.00 [*0.000*]{} (0.77) 1.00 [*0.000*]{} (0.93)\n Gumbel 0.98 [*0.0056*]{} (0.19) 0.74 [*0.0331*]{} (0.34) 0.29 [*0.0202* ]{}(0.42)\n Frank 0.28 [*0.4485*]{} (0.09) 0.80 [*0.0182*]{} (0.27) 0.02 [*0.0061* ]{}(0.41)\n Normal 0.84 [*0.0166*]{} (0.05) 0.20 [*0.0178* ]{} (0.04) 0.03 [*0.0076* ]{}(0.04)\n Student(4) [**0.03 [*0.0074*]{} (0.05)**]{} [**0.01 [*0.0034*]{} (0.05)**]{} [**0.00 [*0.0000*]{} (0.05)**]{}\n \n ------------- ------------- ----------------------------------- ----------------------------------- ----------------------------------\n\n : $n_{MC}=500,n_B=20,n=2048,nn=2048$. Seed $1$. Empirical power for the test of $H_0: c=c_{\\lambda_0}$ at the given level $\\alpha=10\\%$ where $c_{\\lambda_0}$ is specified in the first column and the data are issue from a copula density specified in the second column. ", "The parameter of each copula density is chosen such that the Kendall’s tau is respectively $\\tau=0.25,0.50,0.75$. []{data-label=\"outside\"}\n\n Family parameter grid Cardinal $\\hat \\alpha$ Decision\n ---------- --------------------- ---------- --------------- ----------\n Gumbel $1.05:0.1:1.95$ 10 0.00 Yes\n Gaussian $0.0:0.1:0.9$ 10 0.04 Yes\n Clayton $0.5:0.1:1.4$ 10 0.42 Yes\n Frank $1.5:0.5:6.0$ 10 1.00 No\n Gumbel $1.0:0.05:1.95$ 20 0.00 Yes\n Gaussian $0.0:0.05:0.95$ 20 0.00 Yes\n Clayton $0.5:0.05:1.45$ 20 0.54 No\n Frank $1.25:0.25:6.0$ 20 1.00 No\n Gumbel $1.45$ 1 0.10 Yes\n Gaussien $0.48$ 1 0.12 Yes\n Clayton $0.92$ 1 0.86 No\n Frank $3.20$ 1 1.00 No\n Gumbel $1.36\\quad(1.07\\%)$ 1 0.02 Yes\n Gaussien $0.45\\quad(3.27\\%)$ 1 0.08 Yes\n Clayton $0.41\\;(13.15\\%)$ 1 0.62 No\n Frank $2.88\\quad(3.93\\%)$ 1 1.00 No\n\n : Empirical probability $\\hat \\alpha$ to reject the fit to a fixed parametrical family given in the first column and Decision at the prescribed level $\\alpha=5\\%$. Multivariate null hypotheses (first and second part); $H_0 \\; :\\: c=c_{\\hat\\lambda}$, where $\\hat \\lambda$ is obtained by inversion of the empirical Kendall’s tau (third part); $H_0 \\; :\\; c=c_{\\tilde\\lambda}$, where $\\tilde \\lambda$ is obtained by minimizing the ASE quantity which is given into brackets (fourth part).[]{data-label=\"test_frees\"}\n" ]

{ "pile_set_name": "ArXiv" }

[ 0, 0, 0, 0, 0, 0, 0, 0.0035714285714285713, 0.010351966873706004, 0, 0, 0.007936507936507936, 0, 0, 0.009433962264150943, 0, 0.03225806451612903, 0, 0, 0, 0, 0, 0.006389776357827476, 0.006349206349206349, 0.006329113924050633, 0, 0.005714285714285714, 0, 0, 0.018518518518518517, 0, 0.015, 0, 0, 0.00641025641025641, 0, 0, 0.00909090909090909, 0, 0, 0, 0, 0, 0, 0, 0, 0.012578616352201259, 0, 0.007782101167315175, 0.006369426751592357, 0, 0, 0.010869565217391304, 0, 0.012658227848101266, 0, 0.009345794392523364, 0, 0, 0.0064516129032258064, 0, 0, 0, 0, 0, 0, 0.012048192771084338, 0.009569377990430622, 0.005747126436781609, 0.025, 0.007042253521126761, 0.01, 0, 0.00819672131147541, 0, 0.00819672131147541, 0, 0.00819672131147541, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.005649717514124294, 0, 0.009725400457665904, 0, 0, 0.005012531328320802, 0, 0.00545950864422202, 0, 0.01152073732718894, 0.002173913043478261, 0, 0.010624169986719787, 0.014970059880239521, 0.010471204188481676, 0.0044444444444444444, 0.007010515773660491, 0.006476683937823834, 0, 0, 0.006024096385542169, 0.015151515151515152, 0, 0.00909090909090909, 0.005738880918220947, 0, 0, 0, 0.002325581395348837, 0, 0.005925925925925926, 0.002652519893899204, 0.004830917874396135, 0, 0.0033755274261603376, 0, 0, 0, 0, 0, 0.008, 0.010341261633919338, 0.015360983102918587, 0.008708272859216255, 0.0026595744680851063, 0.0025, 0.008064516129032258, 0.010822510822510822, 0.005319148936170213, 0, 0, 0, 0, 0, 0, 0.007518796992481203, 0, 0, 0, 0.0033783783783783786, 0.017167381974248927, 0, 0, 0, 0, 0.005076142131979695, 0.005154639175257732, 0, 0, 0.006329113924050633, 0.00641025641025641, 0, 0.016260162601626018, 0.010752688172043012, 0.0196078431372549, 0, 0.0053475935828877, 0, 0.005208333333333333, 0, 0, 0, 0, 0.025380710659898477, 0.007575757575757576, 0, 0, 0, 0, 0, 0.006329113924050633, 0.0037313432835820895, 0, 0.005154639175257732, 0, 0, 0, 0.004524886877828055, 0, 0.008, 0, 0, 0.0049261083743842365, 0, 0, 0.009140767824497258, 0, 0.008814102564102564, 0.003968253968253968, 0.011531365313653136, 0, 0, 0.008641975308641974, 0.0044526901669758815, 0.009313154831199068, 0.007792207792207792, 0.0033112582781456954, 0.00392156862745098, 0.015503875968992248, 0.0012360939431396785, 0.01070154577883472, 0.007036059806508356, 0.004739336492890996, 0, 0, 0.011204481792717087, 0, 0.002976190476190476, 0.004301075268817204, 0.009563658099222952, 0.011532125205930808, 0.008771929824561403, 0, 0.014925373134328358, 0.005723630417007359, 0.008968609865470852, 0.01020408163265306, 0.00819672131147541, 0.014721345951629864, 0.01783166904422254, 0.0030864197530864196, 0.00909090909090909, 0.0039011703511053317, 0.008130081300813009, 0.008499413833528723, 0.007532956685499058, 0.006648936170212766, 0.007792207792207792, 0.008149959250203748, 0.00436046511627907, 0.005547850208044383, 0.007722007722007722, 0.0045045045045045045, 0, 0.004166666666666667, 0.005917159763313609, 0.0051813471502590676, 0.00697350069735007, 0.012084592145015106, 0.009884678747940691, 0.00618921308576481, 0.009456264775413711, 0.009202453987730062, 0.007329363261566651, 0.012903225806451613, 0.003952569169960474, 0.0625, 0.0018171906232963837, 0.0063025210084033615 ]

[ "Europe Raiders (2018)\n\nOu Zhou gong lue is a movie starring Ji-hyun Jun, Tony Chiu-Wai Leung, and Kris Wu. ", "The rivalry between two bounty hunters intensifies as they both attempt to track down a destructive device known as the \"Hand of God.\"" ]

{ "pile_set_name": "Pile-CC" }

[ 0.037383177570093455, 0 ]

[ "1. ", "Field of the Invention\nThe present invention relates to a technology allowing a setting item of an image forming apparatus to be changed even in power-saving mode.", "\n2. ", "Background Technology\nIn order to suppress power consumption, an output apparatus, such as an image forming apparatus, on a network enters power-saving mode in which power supply to a portion that consumes large power, such as a printing unit, is cut off. ", "However, in power-saving mode, no reply can be made to a response from the network to the output apparatus about a setting state. ", "Therefore, in order to check the setting state of the output apparatus, it is required to transmit a packet for making a transition from power-saving mode to normal mode from a terminal device through the network to make a transition from power-saving mode to normal mode, and then check the setting state.", "\nHowever, returning to normal mode only for the purpose of checking the setting state poses some problems, such as requiring large power consumption and too long time for mode transition.", "\nTo solve these problems, in a method, a status register independently of the CPU of the output apparatus is provided and, at the time of making an inquiry about the setting state from the terminal device to the output apparatus in power-saving mode through the network, status information stored in the status register is sent back (Japanese Patent Laid-Open Publication No. ", "2002-99399).", "\nHowever, in this method, the amount of information with which the output apparatus can respond is extremely small from several bits to several tens of bits storable in the status register. ", "Moreover, although reading the setting information of the output apparatus (responding to an inquiry) is possible, writing (changing the setting information) is impossible.", "\nIn another method, an image processing apparatus is provided with a first ASIC (Application Specific Integrated Circuit) operating in conjunction with the CPU and a second ASIC that can respond, in conjunction with a network interface, to a signal from a terminal device for requesting a change of power-saving mode of the image processing apparatus, and a startup request signal is transmitted from the terminal device to the second ASIC to start up the first ASIC and the CPU (Japanese Patent Laid-Open Publication No. ", "2005-94679).", "\nHowever, this method is merely a way of starting up the CPU and, a transition to normal mode is still required in order to read and write the setting state of the image processing apparatus.", "\nTherefore, when the setting information, such as user account information, of the image forming apparatus is desired to be changed on the image forming apparatus, it is required to write the setting information of the image forming apparatus. ", "This operation does not require power supply to a printing unit, which consumes power most in the image forming apparatus, and can be performed in power-saving mode, which however cannot be achieved in the conventional technology." ]

{ "pile_set_name": "USPTO Backgrounds" }

[ 0, 0, 0, 0.00390625, 0, 0, 0, 0, 0, 0, 0, 0.0019157088122605363, 0, 0, 0, 0 ]

[ "Q:\n\nLeast square principles with Lagrange multiplier\n\nI have a function to minimize:\n$$f(a_1,a_2,a_3,a_4)=\\sum_{i=1}^n\\left(\\sum_{k=1}^3 a_k\\ p_i^k -a_4\\right)^2$$\nsubjected to this constraint:\n$$a_1^2+a_2^2+a_3^2=1$$\nand \n$$a_4\\geq0$$\nI am trying to cast it in least square principles and Lagrange multiplier and solve it, but I am not quite sure how to do it. ", "\nThe reason I want to do it this way is because I want to leverage existing library like alglib to solve it, it contains the least square principle toolbox, but it doesn't contain least square principles with Lagrange multiplier.", "\nAny idea how to cast this problem in a appropriate format that can be attacked by standard numerical libraries?", "\n\nA:\n\nI don't know if quadratic programming is OK with you because I don't know what alglib does. ", "It takes only to rewrite your function in terms of matrices as follows:\nDefine a vector $$P_i := \\begin{pmatrix}p^1_i\\\\p^2_i\\\\p_i^3 \\\\-1\\end{pmatrix}, a =\\begin{pmatrix}a_1\\\\a_2\\\\a_3 \\\\a_4\\end{pmatrix} $$ Then, your function to be minimized can be expressed as $$f(a) = a^T\\left( \\sum_k P_kP_k^T\\right)a$$ such that $$a^T\\begin{pmatrix} I_3 &0\\\\0&0_1\\end{pmatrix}a = 1$$ and $$\\begin{pmatrix}1&a^T \\end{pmatrix}\\begin{pmatrix} 0 &0 &0 &0 &1\\\\0 &0 &0 &0 &0\\\\0 &0 &0 &0 &0\\\\0 &0 &0 &0 &0\\\\1 &0 &0 &0 &0\\end{pmatrix}\\begin{pmatrix}1\\\\a \\end{pmatrix} \\geq 0$$\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0.008287292817679558, 0.004366812227074236, 0, 0, 0.0017953321364452424 ]

[ "//import {Text} from '../lib/text.mjs';\nimport * as THREE from 'three';\nimport { Text, Rotation, Position, ParentObject3D } from '../components/index.js';\n\nvar newsTicker = {\n url: 'assets/tweets.json',\n hashtag: '#helloWebXR',\n hashtagText: null,\n authorText: null,\n messageText: null,\n news: [],\n current: 0\n};\n\nvar screenMaterial;\n\nexport function setup(ctx, hall) {\n const newsTickerMesh = hall.getObjectByName('newsticker');\n\n ctx.world.createEntity();\n\n screenMaterial = hall.getObjectByName('screen').material;\n\n newsTicker.hashtagText = ctx.world.createEntity();\n newsTicker.hashtagText.addComponent(Text, {\n color: '#f6cdde', //0xdaa056,\n fontSize: 0.1,\n anchor: 'right',\n textAlign: 'right'\n }).addComponent(ParentObject3D, {value: newsTickerMesh});\n\n newsTicker.authorText = ctx.world.createEntity();\n newsTicker.authorText.addComponent(Text, {\n color: '#7f0c38', //0x67bccd,\n fontSize: 0.1,\n anchor: 'left',\n }).addComponent(ParentObject3D, {value: newsTickerMesh});\n\n newsTicker.messageText = ctx.world.createEntity();\n newsTicker.messageText.addComponent(Text, {\n color: 0x000000,\n fontSize: 0.13,\n maxWidth: 2.3,\n lineHeight: 1,\n textAlign: 'left',\n baseline: 'top',\n anchor: 'left'\n }).addComponent(ParentObject3D, {value: newsTickerMesh});\n\n ['hashtag', 'author', 'message'].forEach( i => {\n\n newsTicker[`${i}Text`].addComponent(Position, hall.getObjectByName(i).position);\n newsTicker[`${i}Text`].addComponent(Rotation, {x: 0, y: Math.", "PI, z: 0});\n });\n newsTicker.hashtagText.getMutableComponent(Text).text = newsTicker.hashtag;\n\n fetch(newsTicker.url).then(res => res.json()).then(res => {\n newsTicker.news = res;\n nextNews();\n });\n}\n\nfunction nextNews() {\n const n = newsTicker;\n n.authorText.getMutableComponent(Text).text = n.news[n.current].author;\n n.messageText.getMutableComponent(Text).text = n.news[n.current].message;\n n.current = (n.current + 1) % n.news.length;\n setTimeout(nextNews, 3000);\n}\n\nexport function execute(ctx, delta, time) {\n var v = 0.98 + Math.sin(time * 40) * 0.02;\n screenMaterial.color.setRGB(v, v, v);\n}\n" ]

{ "pile_set_name": "Github" }

[ 0.0006574621959237344, 0.003236245954692557 ]

[ "Q:\n\nNeed Help on entity framework\n\nI have 3 tables(Roles,Actions and RoleActionLinks). ", "Roles table has few columns(RoleID,RoleName,Desc). ", "Actions table has few colums(ActionID,ActionName,Desc). ", "In RoleActionLink is created for store the association between Roles and Actions and this table has the columns such as RoleID,ActionID\nWhen I created the data model(edmx). ", "it shows only Role and Action as entity. ", "i did not find RoleActionLink table. ", "but even there is no direct relation between Roles and Actions table, both tables are automatically related using RoleActionLink table.", "\nWhen i create the new Action, a action record should be populated in Action table(this is works fine). ", "At the same time, i need to populate record in RoleActionLinks table. ", "But i dont have the entity to populate.", "\nPlease tell me how to accomplish my needs.", "\n\nA:\n\nThis should work:\nnewAction.", "Roles.", "Add(role1);\nnewAction.", "Roles.", "Add(role2);\n\nLook at navigation properties in your model. ", "There should be EntityCollection called Roles (name may differ).", "\nEntity framework automatically handles n-n tables and creates collections on both sides.", "\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0, 0.0392156862745098, 0.03571428571428571, 0.005780346820809248, 0, 0, 0.007407407407407408, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.015625, 0, 0 ]

[ "Q:\n\nAppending to PATH in a Windows Docker container\n\nI need to append to the PATH within a Windows Docker container, and I've tried many permutations.", "\nENV PATH=%PATH%;C:\\\\Foo\\\\bin\nENV PATH=$PATH;C:\\\\Foo\\\\bin\nENV PATH=\"%PATH%;C:\\Foo\\bin\"\nENV PATH=\"$PATH;C:\\Foo\\bin\"\nRUN \"set PATH=%PATH%;C:\\Foo\\bin\"\n\nNone of these work: they don't evaluate the preexisting PATH variable.", "\nWhat is the right syntax to append to the PATH? ", "Can I even append to the PATH inside Docker? (", "I can on similar Linux containers)\n\nA:\n\nUnfortunately ENV won't work, because windows environment variable work a little differently than linux. ", "more info\nAs of now the only way to do this is through RUN\nBut you don't need to create a separate file to do this. ", "This can be done by the following much simpler one line command:\nRUN setx path \"%path%;C:\\Foo\\bin\"\n\nA:\n\nYou can set environment variables permanently in the container using a powershell script.", "\nCreate a powershell script in yout docker context (e.g. setpath.ps1 ) containing this:\n[Environment]::SetEnvironmentVariable(\"Path\", $env:Path + \";C:\\Foo\\bin\", [EnvironmentVariableTarget]::Machine)\n\nAdd this script to your container dockerfile and RUN the script. ", "Add something like this to your dockerfile:\nADD ./setpath.ps1 c:/MyScripts/setpath.ps1\nRUN powershell -Command c:\\MyScripts\\setpath.ps1\n\nA:\n\n[Environment]::SetEnvironmentVariable is a good way, but will not work in nanoserver.", "\nThe best choice is:\nRUN setx path '%path%;C:\\Foo\\bin'\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0.02666666666666667, 0.0228310502283105, 0.02040816326530612, 0.043478260869565216, 0.006896551724137931, 0.008620689655172414, 0, 0.0037735849056603774, 0.004424778761061947, 0.017857142857142856 ]

[ "Does maternal serum screening for Down syndrome induce anxiety in younger mothers?", "\nTo assess anxiety levels in mothers with low-risk pregnancies before and after offering routine serum screening. ", "A prospective study was carried out at the Kandang Kerbau Women's and Children's Hospital in Singapore from February 2000 to August 2000. ", "We used standard statistical analysis and Spielberger's state-trait anxiety inventory (STAI) which consists of 40 items to assess anxiety. ", "Anxiety levels were assessed at several stages: before serum screening counselling, after counselling but before serum screening, before the routine 20-week obstetrical screening ultrasound scan, and after ultrasound scan results were acknowledged four to six weeks later. ", "As the STAI questionnaire has only been validated for an English-speaking population, only English-speaking women were recruited for the study. ", "The subjects included 111 women between 15 to 20 weeks gestation that were randomly selected (without any risk factors) for serum screening counselling. ", "Anxiety levels did not decline significantly after counselling by a trained nurse-counsellor. ", "They were highest prior to counselling and were significantly higher compared to all other times in which anxiety was assessed. ", "Anxiety levels were lowest after the serum screening and routine 20-week screening ultrasound scan results were acknowledged. ", "They were also significantly lower compared to all other times in which anxiety was assessed. ", "Anxiety before serum screening was not abnormally high and routine serum screening offered by trained nurse counsellors did not significantly increase maternal anxiety in mothers with low risk pregnancies." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0.007246376811594203, 0.007194244604316547, 0, 0, 0, 0, 0, 0, 0, 0 ]

[ "---\nabstract: 'Finding methods for making generalizable predictions is a fundamental problem of machine learning. ", "By looking into similarities between the prediction problem for unknown data and the lossless compression we have found an approach that gives a solution. ", "In this paper we propose a compression principle that states that an optimal predictive model is the one that minimizes a total compressed message length of all data and model definition while guarantees decodability. ", "Following the compression principle we use Bayesian approach to build probabilistic models of data and network definitions. ", "A method to approximate Bayesian integrals using a sequence of variational approximations is implemented as an optimizer for hyper-parameters: Bayesian Stochastic Gradient Descent (BSGD). ", "Training with BSGD is completely defined by setting only three parameters: number of epochs, the size of the dataset and the size of the minibatch, which define a learning rate and a number of iterations. ", "We show that dropout can be used for a continuous dimensionality reduction that allows to find optimal network dimensions as required by the compression principle.'", "\nauthor:\n- |\n Michael Tetelman\\\n BayzAI, Volkswagen Group of America, Innovation Center California\\\n `michael.tetelman@gmail.com`\nbibliography:\n- 'bibtex.bib'\ntitle: On Compression Principle and Bayesian Optimization for Neural Networks \n---\n\nIntroduction\n============\n\nOne of the most intriguing properties of neural networks models is their ability to learn from available data samples and make predictions about new data, see [@lecun2015deep].", "\n\nThe predictions for new data cannot be absolutely accurate and the prediction errors give a very important characteristic of the neural network model.", "\n\nMaking a prediction about some vector value $x$ means to compute a probability $P(x)$. Unless specifically stated here and throughout the paper $x$ includes all data sample components which for predictive models typically are pairs $x=(input, labels) \\equiv (X,Y)$. We will consider models that predict both input and label components, so the probabilistic model is a chain of submodels for input and label parts $P(x) \\equiv P(X,Y) = P(Y|X)P(X)$.\n\nThe definition of a neural network model provides a specific recipe - an algorithm for computing the probability of $x$ that depends on an exact definition of the network architecture $arch$ and a set of numerical parameters commonly called weights $w$ [@goodfellow2016deep]\n\n$$\\label{model:1}\nP(x|w,arch) \\equiv P(x|w).$$\n\nTo simplify notations we will omit $arch$ argument when it is possible without effecting the meaning.", "\n\nPrediction error or loss for a known ground truth value $x$ is defined as\n\n$$\\label{loss}\nl(x,w) = -\\log{P(x|w)}.$$\n\nWhile it is possible to minimize the loss for a known $x$ by selecting an appropriate network definition and network weights the prediction error for a new data sample which is unknown at the time of prediction will be most of the time higher than for optimized loss.", "\n\nThe novelty gap between the prediction loss for unknown and known data points is inevitable and reflects the information gap between unknown and available data. ", "The goal of the predictive method is to find an algorithm that is expected to minimize that gap.", "\n\nWe would like to note that the prediction problem is very much similar to the problem of a lossless compression of data. ", "The lossless compression problem typically set as follows: there is a sender that creates an encoded message about data points and sends it to a receiver who decodes the message and finds what data are. ", "Like in the prediction problem a sender’s goal is to minimize a message length that is given by negative log of probability of the model in use and send enough information so the receiver will be able to decode the message [@mackay2003information].", "\n\nThe decodability of the compressed message is a very important constraint. ", "To decode the message the receiver must know the definition of the model used by the sender. ", "If the sender will use the weights $w_0$ that minimizes the message length it is necessary also to send an additional message describing $w_0$ by using a probabilistic model of weights $P(w)$ that must be known by a receiver. ", "Then the total message length will be\n\n$$\\label{totalloss}\nl_{tot}(x,w_0) = -\\log{P(x|w_0)} - \\log{P(w_0)}.$$\n\nThe eq. ", "above shows that the optimal weights correspond to a minimum of the total message length for both data and weight descriptions and not the minimum of the message length for the data sample alone.", "\n\nThe encoded message for weights must be send to a receiver because it is required for decodability. ", "Essentially the extra message length is equivalent (but not equal) to a novelty gap for predicting new data. ", "We consider that equivalence in to be more than a coincidence and rather as a fundamental property.", "\n\nWe propose here *the compression principle* that states that an optimal predictive model is the one that minimizes a total compressed message length of all data and model definition while guarantees decodability.", "\n\nThe message length of the model works as a regularizer that allows to avoid overfitting: if model has a big number of weights we can find the weight values that will reduce a compressed data length to be very small, however, the compressed model size will be larger and so the total length.", "\n\nAnother consequence of *the compression principle* is that the optimal model size will increase with the size of available data. ", "In particular this is an important outcome of the neural architecture search (NAS) based on *the compression principle* - this is not in a scope of the current paper and will be considered elsewhere.", "\n\nThe contributions of the paper are follows:\n\n1. ", " A Bayesian stochastic gradient descent optimizer for hyper-parameters (BSGD) is derived by using *the compression principle* applied to a neural network optimization;\n\n2. ", " A method for computing Bayesian integrals is developed based on renormalization group ideas by computing differentials of the hyper-parameters via sequential variational approximations;\n\n3. ", " We show that dropout effectively works as a dimensionality reduction method, that allows to continuously control an effective number of network parameters.", "\n\nThe paper is organized as follows: in section \\[compression\\] we discuss the compression approach, in section \\[dropout\\] dropout as a dimensionality reduction tool, in section \\[integrals\\] the method for approximate computing of Bayesian integrals; in section \\[BSGD\\] we derive an optimizer - Bayesian SGD, examples of using BSGD are discussed in section \\[experiments\\].", "\n\nThe Compression Principle {#compression}\n=========================\n\nWe will consider data samples as randomly and independently drawn from an unknown source distribution.", "\n\nThe probability of a dataset $\\{x_n;n=1..N\\}$ derived from a defined model $P(x|w)$ is given by the following Bayesian integral over weights multiplied by corresponding prior of hyper-parameters and network architecture $$\\label{compression:integral}\n P(\\{x_n\\},\\mathbf{H},arch) = P(\\mathbf{H}|arch) P(arch) \\int\\displaylimits_{w}\n {P(w|\\mathbf{H},arch)\\prod_{n=1}^{N}{P(x_n|w)} \\mathrm{d}w},$$ where $P(w|\\mathbf{H},arch)$ is a prior distribution of weights that depends on hyper-parameters $\\mathbf{H}$ and network architecture $arch$, $P(\\mathbf{H}|arch)$ is a prior distribution of hyper-parameters for a given architecture and $P(arch)$ is a prior of the network architecture.", "\n\nTo make a prediction about data point $x_0$ we need to compute a ratio of two Bayesian integrals\n\n$$\\label{compression:prediction}\nP(x_0|\\{x_n\\},\\mathbf{H},arch) = P(\\{x_0, x_n\\},\\mathbf{H},arch)/P(\\{x_n\\},\\mathbf{H},arch)$$\n\nTo achieve a better compression we need to maximize the probability of dataset w.r.t. ", "hyper-parameters $\\mathbf{H}$ and architecture. ", "To simplify notations further we will include architecture of the network into hyper-parameter variables $\\mathbf{H}$.\n\n*The compression principle* requires decodability, which means that the compressed message should provide a complete description of all hyper-parameters including network architecture that allows to reconstruct both weight prior $P(w|\\mathbf{H})$ and model $P(x|w)$. The decodability results in an expanded message length by the additional part that describes all hyper-parameters\n\n$$\\label{compression:length}\nl_{total} = l_{data} + l_{\\mathbf{H}}; \\;\nl_{\\mathbf{H}} = -\\log{P(\\mathbf{H}|arch)} -\\log{P(arch)}.$$\n\nReducing the number of parameters helps to avoid overfitting. ", "However there are other methods known to improve the predictive performance like dropout without explicitly reducing dimensions of parameters [@dropout2014].", "\n\nWe show here that dropout effectively works as a dimensionality reduction method.", "\n\nDropout as a dimensionality reduction {#dropout}\n=====================================\n\nDropout layer is implemented by randomly replacing feature values with zeros with some rate $r, 0\\leqslant r < 1$. Due to dropout the information content of a feature vector is reduced. ", "To account for that let’s consider a representation of features as a binary string with one bit per feature. ", "Each bit has value 1 if the feature is positive and zero otherwise. ", "The information content in the bit-string is compatible with the definition of ReLU activations, which are essentially binary gate units that transfer input as is if it is positive and output zero otherwise.", "\n\nDropout noise is reducing channel capacity for each bit in the bit-string of features which in the presence of noise is given by a mutual information $I(X;Y)$ between one bit of input $X$ and one bit of output $Y$ of the dropout layer:\n\n$$I(X;Y) = \\sum_{X,Y}{p(X)p(Y|X)\\log_2{\\left(\\frac{p(Y|X)}{p(Y)}\\right)}},$$\n\nwhere $p(Y|X=0)=\\delta_{Y,0}$ and $p(Y|X=1)=r\\delta_{Y,0}+(1-r)\\delta_{Y,1}$. Considering most informative input distribution $p(X=0)=p(X=1)=1/2$ dropout with rate $r$ reduces one bit of information to a fraction of it\n\n$$\\label{dropout:dimred}\nI(X;Y) = 1 - 0.5\\left\\lbrace{r\\log_2{\\frac{1}{r}} + (1+r)\\log_2{(1+r)}}\\right\\rbrace.$$\n\nSee details for general case in the Appendix \\[appendix:dropout\\]. ", "For example, the dropout rate $r=0.5$ in eq. ", "results in a factor 0.3113, which reduces the effective number of features to almost $1/3$ of the original number. ", "Then all effective dimensions of weight matrices should be adjusted by multiplying by square of the corresponding reduction factors, which for $r=0.5$ reduces the effective number of weights by the factor $0.0969 \\approx 0.1$.\n\nComputing Bayesian integrals {#integrals}\n============================\n\nFor approximate computing of the Bayesian integrals in the eq. ", "we will use an idea of renormalization group combined with the variational inference [@mackay2003information].", "\n\nThe integral over weights is taken from a product of the prior of the weight distribution and a product of a large number of model probabilities for each data sample that can be represented as an exponential of the negative loss for a dataset of $N$ samples\n\n$$\\label{int:int}\nI(\\{x_n\\},\\mathbf{H}) = \\int{\\mathrm{d}w P(w|\\mathbf{H})\\prod_{n=1}^N P(x_n|w)} = \\int{\\mathrm{d}w P(w|\\mathbf{H}) e^{-L(w)}},\nL = -\\sum_{n=1}^N{l_n(w)}.$$\n\nThe renormalization group approach here is based on using the same parametrization for the posterior as for the prior, so to approximate the posterior distribution at the initial hyper-parameter point $\\mathbf{H_0}$ we use a prior distribution at a different hyper-parameter point $\\mathbf{H_1}$\n\n$$P(w|\\mathbf{H_0}) e^{-L(w)} \\sim P(w|\\mathbf{H_1}).$$\n\nTo achieve better accuracy we will do it incrementally in a sequence of small steps as follows: we split a small part of loss $L(w)=\\epsilon L + (1-\\epsilon L)$ with a small $\\epsilon \\ll 1$ and factorize the integral in eq. ", "as follows\n\n$$\\label{int:factor}\nI(\\{x_n\\},\\mathbf{H_0}) = \\mathbf{\\langle AB \\rangle_{H_1}} \\approx \\mathbf{\\langle A \\rangle_{H_1} \\langle B \\rangle_{H_1}}$$\n\nwhere $\\mathbf{A}$ and $\\mathbf{B}$ are\n\n$$\\label{int:ab}\n\\mathbf{A} = \\frac{P(w|\\mathbf{H_0})e^{-\\epsilon L}}{\nP(w|\\mathbf{H_1})}, \\; \\; \\mathbf{B} = e^{-(1-\\epsilon)L}.$$\n\nand the average of some $\\mathbf{C}$ is defined as $\\mathbf{\\langle C \\rangle_{H_1}} = \\int{\\mathrm{d}w P(w|\\mathbf{H_1}) \\mathbf{C}(w)}$ where $\\mathbf{C}$ could be $\\mathbf{A, B}$ or $\\mathbf{AB}$.\n\nThe error of replacing the average of a product with a product of averages can be estimated by using the Cauchy-Schwarz inequality [@cauchySchwarz] in the following form\n\n$$\\label{cauchy-schwarz}\n\\left(\\mathbf{\\langle AB \\rangle_{H_1} - \\langle A \\rangle_{H_1} \\langle B \\rangle_{H_1}}\\right)^2 \\leqslant \n\\mathbf{Var\\left(A\\right) Var\\left(B\\right)}.$$\n\nThe error in eq. ", "is minimized by selecting hyper-parameters $\\mathbf{H_1}$ at a minimum of the variance of $\\mathbf{A}$ which gives the following update rule:\n\n$$\\label{int:update}\n\\mathbf{H_1} = \\mathbf{H_0} - \n\\epsilon \\frac{1}{\\langle\\left(\\frac{\\partial\\log{P(w|\\mathbf{H_0})}}{\\partial\\mathbf{H}}\\right)^2\\rangle_{\\mathbf{H_0}}} \n\\frac{\\partial \\langle L \\rangle_{\\mathbf{H_0}}}{\\partial\\mathbf{H}}$$\n\nwhere\n\n$$\\langle L \\rangle_{\\mathbf{H_0}} = \\int{\\mathrm{d}w P(w|\\mathbf{H_0}) L(w)},$$\n\n$$\\langle \\left(\\frac{\\partial\\log{P(w|\\mathbf{H_0})}}{\\partial\\mathbf{H}}\\right)^2 \\rangle_{\\mathbf{H_0}} = \n\\int{\\mathrm{d}w P(w|\\mathbf{H_0})\\left(\\frac{\\partial\\log{P(w|\\mathbf{H_0})}}{\\partial\\mathbf{H}}\\right)^2}.$$\n\nThe eq. ", "describes a gradient descent in the hyper-parameter space. ", "The variance of $\\mathbf{A}$ at a minimal point $\\mathbf{H_1}$ has an order of value $O(\\epsilon^2)$ and goes to zero with $\\epsilon \\rightarrow 0$ and so the error of the approximation of the integral for each step goes to zero as $\\epsilon$. Taking into account the independence of $N$ random samples we can find that for the one step integral approximation the error is $O(\\epsilon \\sqrt{N})$. See details in the Appendix \\[appendix:integral\\].", "\n\nNow, after the first step the integral in eq. ", "is approximately equal to\n\n$$I(\\{x_n\\},\\mathbf{H_0}) \\approx \\\\\n\\int{\\mathrm{d}w P(w|\\mathbf{H_0}) e^{-\\epsilon L}}\n\\int{\\mathrm{d}w P(w|\\mathbf{H_1}) e^{-(1-\\epsilon) L}}.$$\n\nThe whole integral is computed by repeating these steps exactly $T$ times using the update rule with $\\mathbf{H_0, H_1}$ replaced by $\\mathbf{H_t, H_{t+1}}$ for all $t=0..(T-1)$, where $T=1/\\epsilon$ is a number of epochs. ", "Then the integral is equal to\n\n$$\\label{int:total}\nI(\\{x_n\\},\\mathbf{H_0}) = \\prod_{t=0}^{T-1}\n\\left( \\int{\\mathrm{d}w P(w|\\mathbf{H_t}) e^{-\\epsilon L(w)}} \\right).$$\n\nThe posterior distribution is given by $P(w|\\mathbf{H}_{T-1})$ where $\\mathbf{H}_{T-1}$ is a final value of hyper-parameters from recurrent updates in eq..\n\nThe total error of approximating the integral after $T$ epochs accumulates to an exponential factor $\\exp(O(\\sqrt{N}))$ which for large $N$ gives a small correction relative to the value of the integral $I \\propto \\exp(O({N}))$ in the eq.\\[int:total\\].", "\n\nBayesian SGD {#BSGD}\n============\n\nThe developed method for an approximate computation of the Bayesian integrals gives us a general approach for the hyper-parameter optimization. ", "Unlike the approaches based on heuristics and searching the hyper-parameter space [@yu2020hyperparameter] our method by derivation completely determines the optimal parameter values like learning rate and number of iterations for the solution.", "\n\nIn the previous section we considered a method of computing the Bayesian integrals with a sequence of steps where each step accounts for a contribution of a fraction of the total loss that corresponds to a full batch gradient descent. ", "Here we are considering another multi-step method where we are using for each step a fraction of a loss for a single randomly selected data sample or a minibatch $l_n(w)$. For the minibatch $l_n$ is a normalized loss per sample. ", "The update rule will become then a stochastic gradient descent in the hyper-parameter space\n\n$$\\label{bsgd-update}\n\\mathbf{H_{t+1}} = \\mathbf{H_t} - \n\\epsilon \\frac{b}{\\langle\\left(\\frac{\\partial\\log{P(w|\\mathbf{H_t})}}{\\partial\\mathbf{H}}\\right)^2\\rangle_{\\mathbf{H_t}}} \n\\frac{\\partial \\langle l_n \\rangle_{\\mathbf{H_t}}}{\\partial\\mathbf{H}},$$\n\nwhere $b$ is a minibatch size.", "\n\nTotal number of the iterations needed to complete computing the integral $T$ is equal to a number of epochs $N_e$ times a number of minibatches in the dataset $N_b$: $T=N_e N_b$, while parameter $\\epsilon=1/N_e$.\n\nThe eq. ", "defines a general hyper-parameter optimization method based on Bayesian SGD. ", "For the following we will consider an important practical case of a fixed architecture with Gaussian prior distribution of $d$ independent weights where hyper-parameters are means and variances $(\\mu_i, \\sigma_i), i=1..d$\n\n$$P(w|\\mathbf{H}) = \\prod_{i=1}^d{\\frac{e^{-\\frac{(w_i-\\mu_i)^2}{2\\sigma_i^2}}}\n{\\sqrt{2\\pi\\sigma_i^2}}}, \\quad \\mathbf{H} = \\{(\\mu_i,\\sigma_i); i=1..d\\}.$$\n\nWhile in the eq. ", "the factor with a square of the the gradient contains a direct product and actually is a matrix by hyper-parameters we will simplify it by using only diagonal elements of the matrix. ", "Then it is easy to find that for each $i=1..d$\n\n$$\\langle\\left(\\frac{\\partial\\log{P(w|\\mathbf{\\mu,\\sigma)}}}{\\partial\\mu_i}\\right)^2\\rangle = \\frac{1}{\\sigma_i^2}, \\quad\n\\langle\\left(\\frac{\\partial\\log{P(w|\\mathbf{\\mu,\\sigma)}}}{\\partial\\sigma_i}\\right)^2\\rangle = \\frac{2}{\\sigma_i^2}$$\n\nand update rules will look as follows\n\n$$\\label{bsgd-mu-sigma}\n\\mu_{i,t+1} = \\mu_{i,t} - \\epsilon b\\; \\sigma_{i,t}^2\n\\frac{\\partial\\langle l_n\\rangle_t}{\\partial \\mu_{i,t}}, \\quad\n\\sigma_{i,t+1} = \\sigma_{i,t} - \\epsilon b\\; \\frac{\\sigma_{i,t}^2}{2}\n\\frac{\\partial\\langle l_n\\rangle_t}{\\partial \\sigma_{i,t}}.$$\n\nThe gradients of the Gaussian average of the minibatch loss could be transformed to averages of the gradients w.r.t weights\n\n$$\\frac{\\partial\\langle l_n\\rangle_t}{\\partial \\mu_{i,t}} = \n\\langle\\frac{\\partial l_n}{\\partial w_{i,t}}\\rangle_t, \\quad\n\\frac{\\partial\\langle l_n\\rangle_t}{\\partial \\sigma_{i,t}} = \n\\sigma_{i,t}\\langle\\frac{\\partial^2 l_n}{\\partial w_{i,t}^2}\\rangle_t.$$\n\nLet’s define a scaled inverse variance $s_i = 1/(\\sigma_i^2 b)$, where $b$ is the minibatch size with total number of samples $N=bN_b$. Then finally, the equations for Bayesian SGD have the following form\n\n$$\\label{bsgd:final}\n\\mu_{i,t+1} = \\mu_{i,t} - \\frac{\\epsilon}{s_{i,t}}\n\\langle\\frac{\\partial l_n}{\\partial w_{i,t}}\\rangle_t, \\quad\ns_{i,t+1} = s_{i,t} + \\epsilon\\;\n\\langle\\frac{\\partial^2 l_n}{\\partial w_{i,t}^2}\\rangle_t,$$\n\nhere the learning rate is $\\epsilon=1/N_e$ with the total number of iterations $T=N_e N_b$.\n\nIn practical computing the second gradient by weights is replaced by a square of the fist gradient by weights $\\partial^2 l/\\partial w^2 \\rightarrow (\\partial l/\\partial w)^2$. Analysis of that equivalence is given in the Appendix \\[appendix:bsgd\\].", "\n\nBefore a first iteration all inverse variances $s$ are initialized with ones and all means $\\mu$ are initialized with a standard weight initialization. ", "Then training is performed for exactly $T$ iterations. ", "Hyper-parameters $(\\mu,\\sigma)$ are updated at each step as in Algorithm \\[algo:bsgd\\].", "\n\nselect minibatch index: $n$ get weight sample from normal distribution: $w \\gets \\mathcal{N}(\\mu, 1/{\\sqrt{sb}})$ compute the minibatch loss per sample: $l_n(w)$ compute gradient of the loss: $\\mathbf{grad_n} \\gets {\\partial l_n}/{\\partial w}$ compute update for $\\mu$: $\\mu \\gets \\mu - \\frac{1}{N_e} \\mathbf{grad_n}/s$ compute update for $s$: $s \\gets s +\\frac{1}{N_e}(\\mathbf{grad_n})^2$\n\nThe sampling of weights from the weight prior is needed to approximate the averaging of losses by the weight prior. ", "We should note that the weight sampling that was proposed earlier was based on heuristics: the *dropconnect* in [@dropconnect] and the *Gaussian dropout* in [@pmlr-v28-wang13a]. ", "In [@kingma2015variational] the variational dropout was proposed as generalization of the Gaussian dropout.", "\n\nThe Bayesian SGD has only three free parameters: a number of epochs $N_e$, a number of minibatches in a training set $N_b$ and the size of the minibatch that together define a total number of iterations and a learning rate.", "\n\nExperiments\n===========\n\nThe experiments were implemented in Pytorch for MNIST classification problem [@mnist]. ", "Data were used as is without augmentation, preprocessing or pre-training. ", "The convolutional network was trained on a standard training set of 60K images with 28x28 pixels per an image with three different optimizers: BSGD, ADAM [@kingma2014adam], SGD [@ruder2016overview] with the same training conditions: minibatch size 60, training length 10 epochs, with total number of iterations 10K. Training was done on a single GPU machine, with a total training time for any optimizer about 10 min. ", "Testing and validation were computed with a standard 10K-sample testing set. ", "No validation or testing results were used in training.", "\n\nThe learning rate for ADAM was set to $0.0001$ while learning rate for SGD was set to $0.1$. Learning rate for BSGD is set automatically to the inverse number of epoch which is $0.1$. Dropout was used for all non-linear layers with the dropout rate $0.01$.\n\nThe standard cross-entropy loss was used for training with any optimizer.", "\n\nThe network consists of 26 layers total: 1 input convolutional layer, 9 dual convolutional residual blocks, adaptive pooling, 3 fully connected dual residual blocks and a final linear layer. ", "The detailed network definition can be found in Appendix \\[appendix:net\\].", "\n\nThere were no significant differences found in a training progress for the train and validation losses per sample between all 3 optimizers as shown in Figure. ", "\\[fig:t-v-loss\\].", "\n\n![", "Progress of training and validation losses for BSGD, ADAM, SGD given in log scale vs number of epochs. ", "Every 10 points on horizontal axis corresponds to one epoch with 10 epochs total.[]{data-label=\"fig:t-v-loss\"}](optimizers-compare.png)\n\nHowever, the prediction accuracy on the testing set was notably different, with the maximum accuracy over multiple runs 0.9975 obtained by BSGD. ", "Average accuracy over multiple runs for each optimizer is given in the Table \\[table:accuracy\\].", "\n\n BSGD ADAM SGD\n ---------- ---------- ---------\n 0.996323 0.994642 0.99298\n\n : Average accuracy for each tested optimizer.", "\n\n\\[table:accuracy\\]\n\nThe performance for ADAM and SGD optimizers depends on specific values of a number of parameters and first of all on the learning rate. ", "The used learning rate $10^{-4}$ for ADAM and $0.1$ for SGD was selected by trial and error. ", "There is a possibility that a better performance could be achieved with a different selection of the learning rate and other optimizer parameters. ", "However, there is no method known other than exhaustive search for finding the optimal parameters in this case. ", "In practical computing that results in a significant time spent on tuning optimizer parameters via multiple trials.", "\n\nBayesian SGD does not have that problem. ", "In fact it has the highest performance when all parameters in the Algorithm \\[algo:bsgd\\] have values that derived from approximating the Bayesian integrals and lower performance for different values.", "\n\n![", "Average number of errors vs dropout rate for the same network[]{data-label=\"fig:error-dropout\"}](error-dropout.png)\n\nInteresting to see an effect of dropout on the accuracy. ", "Figure \\[fig:error-dropout\\] shows that an average number of test errors does not change significantly for a large interval of dropout rates from 0 to 0.09. ", "The errors were averaged over multiple training runs of the same network definition for the same dropout rate.", "\n\nOn the other side using the estimate of the effective number of feature dimensions \\[dropout:dimred\\] we can find that the effective number of parameters of the network is reduced to 0.6 of the original size. ", "That allows to pose a question “What is an actual number of parameters that parametrize a solution?”", "\n\nConclusion\n==========\n\nWe developed a method for an approximate computing of the posterior given by the Bayesian integral over weights via iterative updates of the hyper-parameters using variational approximations and reparameterization of the prior to represent the posterior.", "\n\nWe derived the Bayesian SGD optimizer from a fundamental Bayesian approach. ", "The BSGD algorithm is automatically defines an optimal learning rate per hyper-parameter and a number of iterations.", "\n\nWe tested BSGD, ADAM, and SGD on a standard MNIST classification problem by training with equal conditions for each optimizer. ", "After averaging over multiple runs the BSGD shows the best performance.", "\n\nWe studied the effect of dropout on the test performance of the networks trained with BSGD optimizer and found no significant difference in the performance for an interval of dropout rates from 0 to 0.09.", "\n\nBroader Impact {#broader-impact .unnumbered}\n==============\n\nWe think that the proposed *compression principle* will attract an interest of the research community.", "\n\nWe think that the developed method for estimating the Bayesian integrals will be found useful for researchers in the field of Deep Learning.", "\n\nWe think that the proposed Bayesian SGD optimizer will allow researchers and practitioners to significantly speed up training neural networks and finding solutions by eliminating a need for multiple trials for searching optimal training parameters.", "\n\nDropout {#appendix:dropout}\n=======\n\nTo model the degradation of the information in features due to dropout we consider the information content of a sign of a feature with one-bit representation of the feature that corresponds to its sign only with values $X=0,1$. We consider the dropout layer representation that is completely equivalent to a standard definition. ", "In this model the dropout layer is a gate unit that transfers the feature. ", "The gate is controlled by a random bit - the dropout noise. ", "When the feature is positive the feature input to the dropout layer is one $X=1$, otherwise the input bit is zero $X=0$. When the random dropout bit is one the gate transfers the feature as is and when the random dropout bit is zero the feature is transferred as zero.", "\n\nFor input probability distribution $P(X)$ the output distribution of the dropout layer $P(Y|X)$ will be\n\n$$\\label{a:dropout:distr}\nP(Y|X=0) = \\delta_{Y,0},\\; P(Y|X=1) = r\\delta_{Y,0}+(1-r)\\delta_{Y,1}.$$\n\nHere r is the dropout rate from the interval $0\\leqslant r <1$.\n\nDue to dropout the information content in the features is reduced. ", "The information measure of the output of the dropout layer is given by the mutual information $I(X;Y)$\n\n$$I(X;Y) = \\sum_{X,Y}{p(X)p(Y|X)\\log_2{\\left(\\frac{p(Y|X)}{p(Y)}\\right)}}.$$\n\nUsing the distributions in the eq. ", "we can find\n\n$$\\label{a:dropout:mutual}\n\\begin{split}\nI(X;Y|r) = & -\\left\\{P(X=0)+rP(X=1)\\right\\}\\log_2\\left\\{P(X=0)+rP(X=1)\\right\\}\\\\\n& - P(X=1)\\left\\{r\\log_2{\\frac{1}{r}}+(1-r)\\log_2P(X=1)\\right\\}.", "\n\\end{split}$$\n\nWhen $r=1$ the information is completely lost $I=0$. When $r=0$ the mutual information in the eq. ", "is equal to the information in $X$\n\n$$I(X;Y|r=0) = -P(X=0)\\log_2{P(x=0)}-P(X=1)\\log_2{P(X=1)}.$$\n\nThe ratio $R=I(X;Y|r)/I(X;Y|r=0)$ gives an effective reduction factor for the amount of information in a feature due to dropout noise. ", "The reduction factor $R=1$ when $r=0$ and $R=0$ when $r=1$. Essentially this factor is reducing the effective dimensionality of the features due to dropout. ", "Then using the dropout with different rates we can continuously control the effective number of feature dimensions.", "\n\nThe integral factorization error {#appendix:integral}\n================================\n\nThe factorization of the integral in the eq. ", "results in the error that is defined by the right side of the Cauchy-Schwarz inequality in the eq.. To compute the error we have to expand the $\\mathbf{A}$ and $\\mathbf{B}$ to the power series by $\\epsilon$ up to the second order $O(\\epsilon^2)$ and compute corresponding expansion of the product of variances $\\mathbf{Var(A)}, \\mathbf{Var(B)}$ up to the second-order of $\\epsilon$. Remember that $\\mathbf{\\langle A \\rangle_{H_1}} = \\int{\\mathrm{dw}P(w|\\mathbf{H_1}) A(w)}$. So the mean of $\\mathbf{A}$ by $P(w|\\mathbf{H_1})$ is\n\n$$\\mathbf{Mean(A)_{H_1}} = \\mathbf{Mean}\\left(\\frac{P(w|\\mathbf{H_0})}\n{P(w|\\mathbf{H_1})}e^{-\\epsilon L}\\right)_{\\mathbf{H_1}} = \\mathbf{Mean}\\left( e^{-\\epsilon L} \\right)_{\\mathbf{H_0}}.$$\n\nThen the variance is\n\n$$\\begin{split}\n\\mathbf{Var(A)} = & \\langle\\left(\\frac{P(w|\\mathbf{H_0})}\n{P(w|\\mathbf{H_1})}e^{-\\epsilon L}\\right)^2\\rangle_{\\mathbf{H_1}} - \n\\left(\\mathbf{Mean(A)_{H_1}}\\right)^2 = \\\\\n& \\langle\\left(\\frac{P(w|\\mathbf{H_0})}\n{P(w|\\mathbf{H_1})}e^{-2\\epsilon L}\\right)\\rangle_{\\mathbf{H_0}} -\n\\langle e^{-\\epsilon L}\\rangle_{\\mathbf{H_0}}^2.", "\n\\end{split}$$\n\nThe new hyper-parameter point $\\mathbf{H_1=H_0 + \\Delta H}$, where $\\Delta H$ has an order of $O(\\epsilon)$. The expansion of the variance of $\\mathbf{A}$ by $\\epsilon$ up to the second order gives\n\n$$\\mathbf{Var(A)_{H_1}} = \\epsilon^2 \\mathbf{Var}(L)_{\\mathbf{H_0}} + \n2\\epsilon \\langle L \\mathbf{\\Delta H}\\frac{\\partial\\log{P(w|\\mathbf{H_0})}}{\\partial\\mathbf{H}}\\rangle_{\\mathbf{H_0}} + \n\\langle \\left( \\mathbf{\\Delta H}\\frac{\\partial\\log{P(w|\\mathbf{H_0})}}{\\partial\\mathbf{H}}\\right)^2\\rangle_{\\mathbf{H_0}}.$$\n\nThe variance of $\\mathbf{B}$ at $\\mathbf{H_1}$ is\n\n$$\\mathbf{Var(B)_{H_1}} = \n\\langle e^{-2(1-\\epsilon)L}\\rangle_{\\mathbf{H_1}} - \n\\langle e^{-(1-\\epsilon)L}\\rangle_{\\mathbf{H_1}}^2.$$\n\nThe leading term in the variance of $\\mathbf{B}$ is $O(1)$ by $\\epsilon$ and does not depend on $\\mathbf{\\Delta H}$. Then we can take into account that $e^{-L}$ is a bounded random variable and so its variance is bounded by a constant $O(1)$, see details in [@popoviciu1935equations], [@doi:10.1080/00029890.2000.12005203].", "\n\nNow we can reduce the integral factorization error by minimizing $\\mathbf{Var(A)_{H_1}}$ by $\\mathbf{\\Delta H}$, which immediately gives the update rule in the eq.. Using the $\\mathbf{\\Delta H}$ from the eq. ", "we get the minimal value of the $\\mathbf{Var(A)_{H_1}}$\n\n$$\\label{a:int:var:A}\n\\min_{\\mathbf{H_1}}\\{\\mathbf{Var(A)_{H_1}}\\} = \\epsilon^2 \\mathbf{Var}(L)_{\\mathbf{H_0}} - \\epsilon^2 \n\\frac{\\partial \\langle L \\rangle_{\\mathbf{H_0}}}{\\partial\\mathbf{H}}\n\\left\\langle \\left(\\frac{\\partial\\log{P(w|\\mathbf{H_0})}}{\\partial\\mathbf{H}} \\right)^2\\right\\rangle_{\\mathbf{H_0}}^{-1}\n\\frac{\\partial \\langle L \\rangle_{\\mathbf{H_0}}}{\\partial\\mathbf{H}}.$$\n\nBecause total loss $L$ is proportional to a number of samples $N$ the second term in the eq. ", "seems to be quadratic by $N$. However, this is not the case we would like to consider. ", "We would like to select hyper-parameters of the prior to make the prior $P(w|\\mathbf{H})$ having a sharp peak by $w$ to be important in the integral relative to losses and that requires the $\\log{P(w|\\mathbf{H})}$ to be at least of order $O(N)$ near maximum. ", "Then the second term in the eq. ", "is $O(1)$ by $N$. The first term is the variance of a large sum of $N$ individual sample losses. ", "Considering that the samples are independent and random the variance of $\\mathbf{A}$ is proportional to $N$.\n\nFinally, the factorization error is a square root of the product of variances of $\\mathbf{A}$ and $\\mathbf{B}$ which gives the integral factorization error estimate for one $\\epsilon$-step to be $O(\\epsilon \\sqrt{N})$. After $T$ steps the error accumulates to a factor $\\exp(O(\\sqrt{N}))$ because $T\\epsilon=1$.\n\nBSGD: Approximating second gradient of loss {#appendix:bsgd}\n===========================================\n\nWhile computing of the first gradient of a sample loss w.r.t. ", "weights is a standard operation in typical neural network frameworks which is implemented via backpropagation algorithm the computing of the second gradient of a sample loss function usually is not available.", "\n\nHowever, availability of the second gradient is not a main problem. ", "Because the loss of a sample is defined as a negative log of a model probability $l(x,w)=-\\log{P(x|w)}$ the second gradient of a loss can be represented as a sum of two parts\n\n$$\\label{a:bsgd:second_grad}\n\\frac{\\partial^2 l}{\\partial w^2} = \\left(\\frac{\\partial l}{\\partial w}\\right)^2 - \\frac{1}{P(x|w)}\\frac{\\partial^2 P(x|w)}{\\partial w^2}.$$\n\nThe second term on the right side of the equation could have large negative values from sharp minima of the model probability that may destabilize a convergence of BSGD in the eq.. To stabilize the iterations of the variances we need to modify that second term. ", "After $t$ iterations the variance will be\n\n$$\\frac{1}{\\sigma_{i,t}^2} = \\frac{1}{\\sigma_{i,0}^2} + \\sum_t{\\frac{1}{T}\\frac{\\partial^2{l_t(w)}}{\\partial{w_{i,t}^2}}}$$\n\nWe can notice that the total contribution from the term with the second gradient of the probability in eq. ", "for all updates of the variance in one epoch in the eq. ", "accumulates to the following value\n\n$$\\int{\\mathrm{dw}P(w|\\mathbf{H})\\;\\frac{1}{N}\\sum_{n=1}^{N}{\\frac{1}{P(x_n|w)} \\frac{\\partial^2 P(x_n|w)}{\\partial w^2}}}.$$\n\nThe sum over samples is an approximation of the average with an unknown true distribution $P(x)$\n\n$$\\label{a:bsgd:int}\n\\frac{1}{N}\\sum_{n=1}^{N}{\\frac{1}{P(x_n|w)} \\frac{\\partial^2 P(x_n|w)}{\\partial w^2}} \\approx \\int{\\mathrm{dx}\\frac{P(x)}{P(x|w)} \\frac{\\partial^2 P(x|w)}{\\partial w^2}}.$$\n\nThe sampling error of replacing the sum with the integral estimate in the eq. ", "is $O(1/\\sqrt{N})\\ll 1$. For typically used large $N$ the sampling error is small.", "\n\nNow we can see that if the model distribution $P(x|w)$ is close or rather proportional up to constant in the area of known samples to the true distribution $P(x)$, which we could expect if the model converges, then the integral over $x$ is diminishing as well as an overall contribution of the term with second gradient of the model probability to the cumulative sum.", "\n\nThe solution to avoid the destabilizing spikes from the second term on the right side in the eq. ", "could be a smoothing of the term compatible with its overall disappearance from the complete epoch sum, however just omitting the second term is found to be a good solution, so in the definition of the BSGD we will approximate the second gradient of a loss with a square of the first gradient of the loss.", "\n\nNetwork definition {#appendix:net}\n==================\n\nThe neural network used for experiments is a convolutional network with a feed-forward architecture.", "\n\nThe network diagram is in the Figure \\[a:fig:net\\].", "\n\n![", "Neural network diagram.[]{data-label=\"a:fig:net\"}](neurips-net.png)\n\nThe network consists of four large blocks: input block, feature extraction block, prediction block and output block.", "\n\nThe input block consists of a single convolutional layer with a kernel size 5x5 that converts one channel input to 100 channels followed by ReLU activations.", "\n\nFeature extraction block has 9 residual blocks with 2 convolutional layers per each block with a kernel size 3x3 and ReLU activations.", "\n\nFeature extraction block is followed by an adaptive pooling layer.", "\n\nThe following prediction block has 3 residual blocks each with 2 fully connected linear layers and ReLU activations.", "\n\nThe output block is a linear layer that converts 100 features to 10.", "\n\nThe cross-entropy loss is implemented via log-softmax function.", "\n" ]

{ "pile_set_name": "ArXiv" }

[ 0, 0, 0, 0, 0.010638297872340425, 0.004878048780487805, 0, 0.01098901098901099, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.008403361344537815, 0, 0, 0, 0, 0, 0, 0, 0.005025125628140704, 0, 0.005813953488372093, 0, 0, 0.0026595744680851063, 0, 0, 0.0031847133757961785, 0, 0.0014347202295552368, 0.006369426751592357, 0, 0, 0, 0, 0, 0.001392757660167131, 0, 0, 0, 0, 0.0009852216748768472, 0.006607929515418502, 0.0014084507042253522, 0, 0, 0, 0, 0.005190311418685121, 0.011049723756906077, 0, 0, 0, 0.0026455026455026454, 0, 0, 0.002512562814070352, 0, 0.009653605905735378, 0, 0, 0, 0, 0.011235955056179775, 0, 0.008888888888888889, 0.008771929824561403, 0, 0.011961722488038277, 0, 0, 0.009009009009009009, 0, 0.013513513513513514, 0, 0, 0, 0.02912621359223301, 0.0035460992907801418, 0.010416666666666666, 0.013513513513513514, 0.012658227848101266, 0.021505376344086023, 0, 0, 0, 0.023255813953488372, 0, 0, 0, 0, 0, 0, 0, 0, 0.01282051282051282, 0.008620689655172414, 0.023255813953488372, 0.014084507042253521, 0.0048543689320388345, 0, 0.007042253521126761, 0.004, 0, 0, 0, 0, 0.0029498525073746312, 0.009216589861751152, 0, 0, 0, 0, 0, 0, 0.0027624309392265192, 0.0047984644913627635, 0, 0.003703703703703704, 0, 0, 0, 0, 0.001692047377326565, 0, 0, 0.0016420361247947454, 0, 0, 0, 0, 0, 0, 0.003278688524590164, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]

[ "/**\n * @license\n * Copyright The Closure Library Authors.", "\n * SPDX-License-Identifier: Apache-2.0\n */\n\n/**\n * @fileoverview A Map that associates multiple values with a single key.", "\n */\n\ngoog.provide('goog.labs.structs.", "Multimap');\n\ngoog.require('goog.array');\n\n\n\n/**\n * Creates a new multimap.", "\n * @final\n * @template K, V\n */\ngoog.labs.structs.", "Multimap = class {\n constructor() {\n this.clear();\n }\n\n /**\n * Clears the multimap.", "\n */\n clear() {\n this.count_ = 0;\n this.map_ = new Map();\n }\n\n /**\n * Clones this multimap.", "\n * @return {!", "goog.labs.structs.", "Multimap<K, V>} A multimap that contains all\n * the mapping this multimap has.", "\n */\n clone() {\n var map = new goog.labs.structs.", "Multimap();\n map.addAllFromMultimap(this);\n return map;\n }\n\n /**\n * Adds the given (key, value) pair to the map. ", "The (key, value) pair\n * is guaranteed to be added.", "\n * @param {K} key The key to add.", "\n * @param {V} value The value to add.", "\n */\n add(key, value) {\n var values = this.map_.get(key);\n if (!", "values) {\n this.map_.set(key, (values = []));\n }\n\n values.push(value);\n this.count_++;\n }\n\n /**\n * Stores a collection of values to the given key. ", "Does not replace\n * existing (key, value) pairs.", "\n * @param {K} key The key to add.", "\n * @param {!", "Array<V>} values The values to add.", "\n */\n addAllValues(key, values) {\n goog.array.forEach(values, function(v) {\n 'use strict';\n this.add(key, v);\n }, this);\n }\n\n /**\n * Adds the contents of the given map/multimap to this multimap.", "\n * @param {!", "goog.labs.structs.", "Multimap<K, V>} map The\n * map to add.", "\n */\n addAllFromMultimap(map) {\n goog.array.forEach(map.getEntries(), function(entry) {\n 'use strict';\n this.add(entry[0], entry[1]);\n }, this);\n }\n\n /**\n * Replaces all the values for the given key with the given values.", "\n * @param {K} key The key whose values are to be replaced.", "\n * @param {!", "Array<V>} values The new values. ", "If empty, this is\n * equivalent to `removeAll(key)`.", "\n */\n replaceValues(key, values) {\n this.removeAll(key);\n this.addAllValues(key, values);\n }\n\n /**\n * Gets the values correspond to the given key.", "\n * @param {K} key The key to retrieve.", "\n * @return {!", "Array<V>} An array of values corresponding to the given\n * key. ", "May be empty. ", "Note that the ordering of values are not\n * guaranteed to be consistent.", "\n */\n get(key) {\n var values = this.map_.get(key);\n return values ? ", "goog.array.clone(values) : [];\n }\n\n /**\n * Removes a single occurrence of (key, value) pair.", "\n * @param {K} key The key to remove.", "\n * @param {V} value The value to remove.", "\n * @return {boolean} Whether any matching (key, value) pair is removed.", "\n */\n remove(key, value) {\n var values = this.map_.get(key);\n if (!", "values) {\n return false;\n }\n\n var removed = goog.array.removeIf(values, function(v) {\n 'use strict';\n return Object.is(value, v);\n });\n\n if (removed) {\n this.count_--;\n if (values.length == 0) {\n this.map_.delete(key);\n }\n }\n return removed;\n }\n\n /**\n * Removes all values corresponding to the given key.", "\n * @param {K} key The key whose values are to be removed.", "\n * @return {boolean} Whether any value is removed.", "\n */\n removeAll(key) {\n // We have to first retrieve the values from the backing map because\n // we need to keep track of count (and correctly calculates the\n // return value). ", "values may be undefined.", "\n var values = this.map_.get(key);\n if (this.map_.delete(key)) {\n this.count_ -= values.length;\n return true;\n }\n\n return false;\n }\n\n /**\n * @return {boolean} Whether the multimap is empty.", "\n */\n isEmpty() {\n return !", "this.count_;\n }\n\n /**\n * @return {number} The count of (key, value) pairs in the map.", "\n */\n getCount() {\n return this.count_;\n }\n\n /**\n * @param {K} key The key to check.", "\n * @param {V} value The value to check.", "\n * @return {boolean} Whether the (key, value) pair exists in the multimap.", "\n */\n containsEntry(key, value) {\n var values = this.map_.get(key);\n if (!", "values) {\n return false;\n }\n\n var index = goog.array.findIndex(values, function(v) {\n 'use strict';\n return Object.is(v, value);\n });\n return index >= 0;\n }\n\n /**\n * @param {K} key The key to check.", "\n * @return {boolean} Whether the multimap contains at least one (key,\n * value) pair with the given key.", "\n */\n containsKey(key) {\n return this.getKeys().includes(key);\n }\n\n /**\n * @param {V} value The value to check.", "\n * @return {boolean} Whether the multimap contains at least one (key,\n * value) pair with the given value.", "\n */\n containsValue(value) {\n return this.getValues().includes(value);\n }\n\n /**\n * @return {!", "Array<K>} An array of unique keys.", "\n */\n getKeys() {\n return [...this.map_.keys()];\n }\n\n /**\n * @return {!", "Array<V>} An array of values. ", "There may be duplicates.", "\n */\n getValues() {\n return goog.array.flatten([...this.map_.values()]);\n }\n\n /**\n * @return {!", "Array<!Array<K|V>>} An array of entries. ", "Each entry is of the\n * form [key, value].", "\n */\n getEntries() {\n var keys = this.getKeys();\n var entries = [];\n for (var i = 0; i < keys.length; i++) {\n var key = keys[i];\n var values = this.get(key);\n for (var j = 0; j < values.length; j++) {\n entries.push([key, values[j]]);\n }\n }\n return entries;\n }\n};\n\n\n/**\n * The backing map.", "\n * @private {!", "Map<K, !", "Array<V>>}\n */\ngoog.labs.structs.Multimap.prototype.map_;\n\n\n/**\n * @private {number}\n */\ngoog.labs.structs.Multimap.prototype.count_ = 0;\n" ]

{ "pile_set_name": "Github" }

[ 0.017543859649122806, 0.01639344262295082, 0, 0, 0.0392156862745098, 0, 0, 0.0625, 0, 0, 0, 0, 0, 0.027777777777777776, 0.025, 0, 0, 0, 0.027777777777777776, 0.06666666666666667, 0, 0.004608294930875576, 0.06666666666666667, 0, 0, 0, 0.01639344262295082, 0.06666666666666667, 0, 0, 0, 0.024390243902439025, 0.0625, 0, 0, 0, 0, 0, 0.02564102564102564, 0.023255813953488372, 0.013513513513513514, 0, 0.0027548209366391185, 0.016666666666666666, 0.018867924528301886, 0, 0, 0.004651162790697674, 0, 0.011235955056179775, 0.010638297872340425, 0.023809523809523808, 0.012987012987012988, 0, 0.008733624454148471, 0.008849557522123894, 0.008264462809917356, 0.008695652173913044, 0.009708737864077669, 0, 0.012345679012345678, 0, 0, 0.009523809523809525, 0.024390243902439025, 0, 0, 0.06666666666666667, 0, 0.007246376811594203 ]

[ "Introduction\n============\n\nCombinatorial chemistry and high-throughput screening are advancing drug discovery. ", "However, many lead compounds have been categorized as Biopharmaceutics Classification System classes II or IV, with poor solubility.[@b1-ijn-10-2521] The insoluble nature causes significant challenges in drug delivery. ", "Nano-drug systems are well-established approaches to meeting the formulation challenges of insoluble drugs.[@b2-ijn-10-2521] These promising carriers include liposomes,[@b3-ijn-10-2521],[@b4-ijn-10-2521] micro-/nanoemulsions,[@b5-ijn-10-2521] micelles,[@b6-ijn-10-2521],[@b7-ijn-10-2521] solid lipid nanoparticles,[@b8-ijn-10-2521] polymeric nanoparticles,[@b9-ijn-10-2521] and nanocrystals.[@b10-ijn-10-2521] Of these, liposomes and lipid emulsions have been successfully developed into injectable products such as Doxil^®^ and Diprivan^®^. However, other nanocarriers have seen a low rate of commercial success due to physical instability, poor reproducibility, and potential toxicity. ", "For those drugs that are insoluble both in water and in oil (or lipids), low drug load and significant drug leakage are the main obstacles to developing injectable nanocarriers. ", "In this regard, nanocrystals are advantageous because they can greatly enhance drug solubility. ", "It is noteworthy that a nanocrystal product of paliperidone for injection has been approved by the European Medicines Agency[@b11-ijn-10-2521] and five oral products of drug nanocrystals have been approved by the US Food and Drug Administration.[@b10-ijn-10-2521]\n\nNanocrystals (or nanosuspensions) are a sub-micron colloidal dispersion system of pure drug particles in water.[@b12-ijn-10-2521]--[@b14-ijn-10-2521] Nanocrystal systems are often stabilized by surfactants or polymers.[@b12-ijn-10-2521]--[@b14-ijn-10-2521] The advantages of nanocrystals include: (1) they offer a high drug load (because the drug is suspended in a solid state); (2) there is no significant toxicity concern because a limited quantity of excipient (stabilizer) is used; (3) they are suitable for a diverse range of administration routes including oral, parenteral, pulmonary, and ocular pathways;[@b15-ijn-10-2521]--[@b17-ijn-10-2521] and (4) they show potential in passive targeting and provide a depot effect.[@b18-ijn-10-2521]--[@b20-ijn-10-2521] Drug nanocrystals are usually prepared using two types of methods, namely, anti-solvent precipitation of dissolved drugs (bottom-up)[@b21-ijn-10-2521],[@b22-ijn-10-2521] and comminution of bulk drugs with high-pressure homogenization or media milling (top-down).[@b23-ijn-10-2521],[@b24-ijn-10-2521] Although bottom-up technology is a straight-forward procedure to obtain drug nanocrystals, it has several limitations, such as the use of harmful organic solvents[@b25-ijn-10-2521],[@b26-ijn-10-2521] and there being great difficulty in controlling particle size.[@b27-ijn-10-2521] Top-down technology is more applicable to industrial production due to the facile and energy-saving processes involved.[@b28-ijn-10-2521]\n\nSNX-2112 is a novel Hsp90 inhibitor currently in clinical evaluation for cancer treatment (delivered by its prodrug SNX-5422; [www.clinicaltrials.gov](http://www.clinicaltrials.gov)). ", "The anticancer action of the compound is through abrogation of signaling via human epidermal growth factor receptor 2 (HER2), protein kinase B (Akt), and extracellular-signal-regulated kinase (ERK).[@b29-ijn-10-2521],[@b30-ijn-10-2521] SNX-2112 is hardly bioavailable after oral administration due to rather limited solubility and possible P-gp efflux.[@b31-ijn-10-2521] The prodrug SNX-5422 has been designed with an improved solubility profile, providing acceptable systemic exposure of SNX-2112. ", "An effective alternative to circumvent the low bioavailability of SNX-2112 is to develop its parenteral formulations. ", "Due to poor solubility both in water (8 μg/mL) and oil (eg, 5 μg/mL in soybean oil), nanocrystals are an excellent candidate for the intravenous delivery of the drug.", "\n\nPhysiologically based pharmacokinetic (PBPK) modeling has gained increasing attention in drug research and development.[@b32-ijn-10-2521],[@b33-ijn-10-2521] A PBPK model quantitatively describes drug absorption, distribution, metabolism, and elimination in the body, facilitating a deep understanding of the effects of these intricate processes on drug exposure and of how these processes interact with each other.[@b34-ijn-10-2521],[@b35-ijn-10-2521] An important feature of an PBPK model is that it allows data extrapolation from a preclinical animal species to humans, thereby providing excellent guidance on clinical trials.[@b36-ijn-10-2521] Although PBPK modeling has become a powerful tool in identifying the factors determining drug pharmacokinetics, its ability to evaluate the effects of the formulation factor on drug disposition has not been fully established.[@b37-ijn-10-2521]\n\nIn the study reported here, we aimed to explore the possibility and feasibility of nanocrystals for the intravenous delivery of SNX-2112, and to determine the in vivo fate of the drug nanocrystals. ", "Drug nanocrystals were prepared using the wet-media milling technique and characterized by particle size, differential scanning calorimetry (DSC), in vitro release, etc. ", "PBPK modeling was performed to evaluate the drug disposition in rats following administration of the nanocrystals using drug cosolvent as a reference. ", "A rat PBPK model incorporating the parameters of particulate uptake (into the liver and spleen) and of the in vivo drug release was used to describe and explain the experimental data.", "\n\nMaterials and methods\n=====================\n\nMaterials\n---------\n\nSNX-2112 and AT-533 (used as an internal standard) were synthesized in our laboratory as described previously.[@b38-ijn-10-2521] Poloxamer 188 (P188) was obtained from BASF SE (Ludwigshafen, Germany). ", "Polysorbate 80 (Tween^®^ 80) was obtained from Aladdin Reagent Co Ltd (Shanghai, People's Republic of China). ", "All other chemicals were of analytical grade and used as received.", "\n\nPreparation of SNX-2112 nanocrystals\n------------------------------------\n\nDrug nanocrystals were prepared using a top-down wet-milling technique.[@b39-ijn-10-2521] In brief, the stabilizer (0.2, 2.0, or 20.0 mg) was dissolved in 2 mL water, into which 20 mg of drug powder was dispersed. ", "The resulting suspensions were subjected to homogenization in a homemade medium mill containing 2 g milling pearls (zirconium oxide, diameter of 1.2 and 0.8 mm at a ratio of 1:1) stirred at 1,100 rpm. ", "After 24 hours' homogenization, the nanocrystal suspensions were separated from the system by filtration against a nylon sieve.", "\n\nCharacterization of SNX-2112 nanocrystals\n-----------------------------------------\n\nThe particle size and zeta potential of the drug nanocrystals were determined using a Zetasizer Nano ZS (Malvern Instruments, Malvern, UK) at 25°C based on the dynamic light scattering principle. ", "The sample (0.1 mL) was diluted with deionized water to 1 mL and measured after equilibrium to the given temperature in the cell. ", "To calculate particle size, the data were recorded by and analyzed with the built-in software.", "\n\nThe morphology of the drug nanocrystals was observed by transmission electron microscopy (TEM). ", "In brief, an aliquot of nanocrystals was dropped on a carbon-coated copper grid and then anchored, drying at ambient atmosphere, to the supporter. ", "The supporter with fixed nanoparticles was immersed into 1% phosphotungstic acid solution for 100 seconds. ", "The negatively stained particles were allowed to dry at ambient atmosphere and photographed at an acceleration voltage of 100 kV with TEM (Tecnai 10, Philips, Amsterdam, the Netherlands).", "\n\nDSC\n---\n\nThe samples of lyophilized nanocrystals, stabilizer (P188), pure drug, and physical mixtures of about 5 mg were weighed into a non-hermetically sealed aluminum pan, and proceeded to differential calorimetric scanning on a DSC 204A/G Phoenix^®^ instrument (Netzsch GmbH & Co KG, Selb, Germany). ", "The samples were heated from 25°C to 350°C at a stepping rate of 10°C/minute. ", "The instrument was calibrated using indium. ", "All the DSC measurements were carried out in a nitrogen atmosphere at a flow rate of 100 mL/minute.", "\n\nIn vitro drug release study\n---------------------------\n\nThe bulk-equilibrium reverse dialysis method was used for the release study of SNX-2112 nanocrystals.[@b40-ijn-10-2521] The release medium was a phosphate buffer (pH 7.4) with 0.2% Tween 80. ", "After equilibrium with the bulk phase under magnetic agitation, 1 mL of nanocrystal (2 mg/mL) suspension was introduced and reversely dialyzed at 37°C against the dialysis bags. ", "At predetermined intervals, the dialysis bags were withdrawn. ", "SNX-2112 concentrations in the dialysate were determined and the percentage of drug release was calculated as mean ± standard deviation (n=3).", "\n\nPharmacokinetic study of SNX-2112\n---------------------------------\n\nPharmacokinetic study was performed with Sprague-Dawley rats (male; 193.0±8.9 g). ", "The experimental procedures were approved by the Animal Ethics Committee of Jinan University (Guangzhou, China). ", "The rats were randomly divided into two groups, the cosolvent (control) and nanocrystal groups. ", "The control group of rats received SNX-2112 cosolvent (45% propylene glycerol) by intravenous injection (tail vein) at a dose of 2 mg/kg body weight, whereas the nanocrystal group received SNX-2112 nanocrystals intravenously at a same dose. ", "At predetermined time points (10, 20, and 30 minutes; 1 hour; and 2, 4, 6, 8, 12, and 18 hours) after drug administration, the rats (n=5) were rendered unconscious for blood and tissue sampling by injecting phenobarbital. ", "The blood was collected by cardiac puncture. ", "After washout of blood with ice-cold saline, the heart, liver, spleen, lung, kidney, and intestine were rapidly removed, weighed, and stored at −80°C for further analysis.", "\n\nBlood samples were processed as described previously.[@b41-ijn-10-2521] The concentrations of SNX-2112 in the samples were analyzed by the ultra-performance liquid chromatography quadrupole time-of flight mass spectrometry (UPLC-QTOF/MS) system. ", "The heart, liver, spleen, lung, kidney, and intestine tissues were homogenized 1:2 (w/v) in saline solution. ", "A 0.2 mL aliquot of tissue homogenate was mixed with 1 mL acetonitrile containing 2 μM internal standard (AT-533). ", "The mixture was vortexed and centrifuged at 13,000× *g* for 15 minutes (4°C). ", "The supernatant was collected and dried using a Concentrator plus (Eppendorf AG, Hamburg, Germany). ", "The residue was reconstituted in 250 μL of water/methanol (50:50, v/v) and centrifuged at 13,000× *g* for 15 minutes (4°C). ", "A 3 μL aliquot of the supernatant was injected into the UPLC-QTOF/MS system.", "\n\nQuantification of SNX-2112 using UPLC-QTOF/MS\n---------------------------------------------\n\nQuantification of SNX-2112 was performed using a Waters UPLC-QTOF/MS system consisting of an ACQUITY UPLC system and Xevo™ G2 QTof (Waters Corporation, Milford, MA, USA) as detailed in our previous publications.[@b31-ijn-10-2521],[@b42-ijn-10-2521] The analytical methods were rigorously validated with respect to linearity (4.88--5,000.00 nM), precision (\\>90%), and recovery (\\>80%).", "\n\nNonparticulate PBPK model\n-------------------------\n\nA nonparticulate PBPK model was constructed to describe disposition of nonparticulate SNX-2112 in rats. ", "The whole-body PBPK model consisted of heart (ht), liver (li), spleen (sp), lung (lu), kidney (kd), and intestine (in) compartments. ", "All other tissues were lumped into a remainder (rm) compartment ([Figure 1A](#f1-ijn-10-2521){ref-type=\"fig\"}). ", "Elimination of SNX-2112 primarily occurred in the liver,[@b41-ijn-10-2521] which was represented by the apparent hepatic clearance *CL~h~*. ", "Drug concentrations in each tissue were determined by the tissue/plasma partition coefficient (*K~p,ti~*), plasma flow (*Q~ti~*), and organ volume (*V~ti~*). ", "The *K~p,ti~* values of SNX-2112 for the intestine, kidney, spleen, and heart were obtained experimentally from the tissue/plasma ratios toward the end of the study (close to 18 hours). [", "Table 1](#t1-ijn-10-2521){ref-type=\"table\"} summarizes the plasma flow and organ volume values that were fixed during model fitting. ", "Plasma cardiac output (*Q~co~*) for rats was set at 2.69 L/h.[@b43-ijn-10-2521] The mass balance equations ([Equations 1](#fd1-ijn-10-2521){ref-type=\"disp-formula\"}--[8](#fd8-ijn-10-2521){ref-type=\"disp-formula\"}) for the model follow. \"*", "C*\" denotes the concentration of SNX-2112 in each compartment. \"*", "V*\" denotes the volume in each compartment. ", "The subscripts represent the compartment names. ", "$$V_{pl}\\frac{dC_{pl}}{dt} = Q_{co}\\left( {\\frac{C_{lu}}{Kp_{lu}} - C_{pl}} \\right)$$ $$V_{ht}\\frac{dC_{ht}}{dt} = Q_{ht}\\left( {C_{pl} - \\frac{C_{ht}}{Kp_{ht}}} \\right)$$ $$V_{sp}\\frac{dC_{sp}}{dt} = Q_{sp}\\left( {C_{pl} - \\frac{C_{sp}}{Kp_{sp}}} \\right)$$ $$V_{in}\\frac{dC_{in}}{dt} = Q_{in}\\left( {C_{pl} - \\frac{C_{in}}{Kp_{in}}} \\right)$$ $$V_{rm}\\frac{dC_{rm}}{dt} = Q_{rm}\\left( {C_{pl} - \\frac{C_{rm}}{Kp_{rm}}} \\right)$$ $$V_{kd}\\frac{dC_{kd}}{dt} = Q_{kd}\\left( {C_{pl} - \\frac{C_{kd}}{Kp_{kd}}} \\right)$$ $$V_{li}\\frac{dC_{li}}{dt} = Q_{ha}C_{pl} + Q_{in}\\frac{C_{in}}{Kp_{in}} + Q_{sp}\\frac{C_{sp}}{Kp_{sp}} - Q_{li}\\frac{C_{li}}{Kp_{li}} - CL_{H}C_{li}$$ $$V_{lu}\\frac{dC_{lu}}{dt} = Q_{1i}\\frac{C_{li}}{Kp_{li}} + Q_{ht}\\frac{C_{ht}}{Kp_{ht}} + Q_{rm}\\frac{C_{rm}}{Kp_{rm}} + Q_{kd}\\frac{C_{kd}}{Kp_{kd}} - Q_{co}\\frac{C_{lu}}{Kp_{lu}}$$\n\nNanoparticulate PBPK model\n--------------------------\n\nFollowing the administration of drug nanocrystals, the drug in the plasma (and tissues when uptake of nanoparticles occurred) had two forms, nonparticulate (D) and nanoparticulate (N) ([Figure 1B](#f1-ijn-10-2521){ref-type=\"fig\"}). ", "It was necessary to distinguish the two forms in pharmacokinetic modeling due to their differences in disposition processes, such as tissue partitioning. ", "The nanoparticulate PBPK model was constructed based on the nonparticulate PBPK model ([Figure 1A](#f1-ijn-10-2521){ref-type=\"fig\"}). ", "The nanoparticulate model additionally incorporated a species of the nanoparticulate drug in the plasma, liver, and spleen compartments. ", "The nonparticulate drug was released from the nanoparticulate form following the first-order rate constant (*K~rel~*). ", "Uptake of nanoparticles into the liver and spleen is represented by the apparent clearance, *Up~li~* and *Up~sp~*, respectively. ", "The mass balance equations ([Equations 9](#fd9-ijn-10-2521){ref-type=\"disp-formula\"}--[19](#fd19-ijn-10-2521){ref-type=\"disp-formula\"}) for the model follow. \"*", "C*\" and \"*C^Nano^*\" denote the concentrations of the nonparticulate and particulate SNX-2112 in each compartment, respectively. \"*", "V*\" denotes the volume in each compartment. ", "The subscripts represent the compartment names. ", "$$V_{pl}\\frac{dC_{pl}}{dt} = Q_{co}\\left( {\\frac{C_{lu}}{Kp_{lu}} - C_{pl}} \\right) + K_{rel}V_{pl}C_{pl}^{Nano}$$ $$V_{pl}\\frac{dC_{pl}^{Nano}}{dt} = - K_{rel}V_{pl}C_{pl}^{Nano} - Up_{li}C_{pl}^{Nano} - Up_{sp}C_{pl}^{Nano}$$ $$V_{ht}\\frac{dC_{ht}}{dt} = Q_{ht}\\left( {C_{pl} - \\frac{C_{ht}}{Kp_{ht}}} \\right)$$ $$V_{sp}\\frac{dC_{sp}}{dt} = Q_{sp}\\left( {C_{pl} - \\frac{C_{sp}}{Kp_{sp}}} \\right) + K_{rel}V_{sp}C_{sp}^{Nano}$$ $$V_{sp}\\frac{dC_{sp}^{Nano}}{dt} = Up_{sp}C_{pl}^{Nano} - K_{rel}V_{sp}C_{sp}^{Nano}$$ $$V_{in}\\frac{dC_{in}}{dt} = Q_{in}\\left( {C_{pl} - \\frac{C_{in}}{Kp_{in}}} \\right)$$ $$V_{rm}\\frac{dC_{rm}}{dt} = Q_{rm}\\left( {C_{pl} - \\frac{C_{rm}}{Kp_{rm}}} \\right)$$ $$V_{kd}\\frac{dC_{kd}}{dt} = Q_{kd}\\left( {C_{pl} - \\frac{C_{kd}}{Kp_{kd}}} \\right)$$ $$V_{li}\\frac{dC_{li}}{dt} = Q_{ha}C_{pl} + Q_{in}\\frac{C_{in}}{Kp_{in}} + Q_{sp}\\frac{C_{sp}}{Kp_{sp}} - Q_{li}\\frac{C_{li}}{Kp_{li}} - CL_{H}C_{li} + K_{rel}V_{li}C_{li}^{Nano}$$ $$V_{li}\\frac{dC_{li}^{Nano}}{dt} = Up_{li}C_{pl}^{Nano} - K_{rel}V_{li}C_{li}^{Nano}$$ $$V_{lu}\\frac{dC_{lu}}{dt} = Q_{li}\\frac{C_{li}}{Kp_{li}} + Q_{ht}\\frac{C_{ht}}{Kp_{ht}} + Q_{rm}\\frac{C_{rm}}{Kp_{rm}} + Q_{kd}\\frac{C_{kd}}{Kp_{kd}} - Q_{co}\\frac{C_{lu}}{Kp_{lu}}$$\n\nData analysis\n-------------\n\nModeling and simulations were performed with MATLAB^®^ 8.0 software (The MathWorks, Inc, Natick, MA, USA). ", "All parameters were estimated using a weighing scheme of 1/y[@b2-ijn-10-2521]. ", "Compartmental and non-compartmental analyses of plasma/tissue concentrations versus time were performed using Pharsight^®^ Phoenix^®^ WinNonlin^®^ 6.3 software (Certara, LP, Princeton, NJ, USA). ", "Data are expressed as mean ± standard deviation. ", "The two-tailed Student's *t*-test was used to compare the parameter difference between the cosolvent (control) and nanocrystal groups. ", "The prior level of significance was set at 5% or *P*\\<0.05.", "\n\nResults\n=======\n\nStabilizer selection for SNX-2112 nanocrystals\n----------------------------------------------\n\nThe stabilizer played an important role in determining the properties of the drug nanocrystals.[@b24-ijn-10-2521] We prepared SNX-2112 nanocrystals using different types of stabilizers including nonionic surfactants (Tween 80, P188, and P407), ionic surfactants (sodium dodecyl sulfate and sodium taurocholate hydrate), and polymers (hydroxypropyl methylcellulose and polyvinylpyrrolidone) ([Figure 2](#f2-ijn-10-2521){ref-type=\"fig\"}). ", "Also, the effects of the amount of stabilizer on the nanocrystals were evaluated ([Figure 2](#f2-ijn-10-2521){ref-type=\"fig\"}). ", "Clearly, the drug nanocrystals (with a particle \\<1 μm) can be formed by using any of tested stabilizers at a higher amount of 20 mg ([Figure 2A](#f2-ijn-10-2521){ref-type=\"fig\"}). ", "Although all other stabilizers showed potential in generating SNX-2112 nanocrystals, P188 was chosen as the optimal stabilizer for further evaluation because (1) the resulting nanocrystals showed acceptable particle size with narrow size distribution ([Figure 2](#f2-ijn-10-2521){ref-type=\"fig\"}), and (2) P188 is a drug excipient for intravenous administration without safety concerns.[@b44-ijn-10-2521]\n\nPreparation and characterization of optimized SNX-2112 nanocrystals\n-------------------------------------------------------------------\n\nThe optimized SNX-2112 nanocrystals were prepared using P188 (20 mg) as the stabilizer. ", "The obtained nanocrystals showed a particle size of 203 nm with a small polydispersity index value of 0.153 ([Figure 3A](#f3-ijn-10-2521){ref-type=\"fig\"}). ", "The zeta potential was determined to be −11.6 mV. This electric charge would help to stabilize the nanosuspensions, though the steric effects of P188 may play a major role in stabilization of the colloidal systems.[@b45-ijn-10-2521] In addition, SNX-2112 nanocrystals presented acicular architectures in their morphology as observed by TEM ([Figure 3B](#f3-ijn-10-2521){ref-type=\"fig\"}), indicating that they were indeed nanocrystals rather than an amorphous substance. ", "Preliminary stability studies also revealed that the prepared nanocrystals with P188 were stable for at least 1 month ([Figure 4](#f4-ijn-10-2521){ref-type=\"fig\"}).", "\n\nDSC thermograms of the pure drug (SNX-2112), P188, physical mixture, and drug nanocrystals are shown in [Figure 3C](#f3-ijn-10-2521){ref-type=\"fig\"}. ", "The endothermic peaks at 57°C and 264°C corresponded to the melting points of P188 and SNX-2112, respectively ([Figure 3C](#f3-ijn-10-2521){ref-type=\"fig\"}). ", "The DSC profiles of the nanocrystals and physical mixture were superposition of those of individual SNX-2112 and P188. ", "This lends strong support to the notion that the nanoscaled drug was crystalline. ", "It has been reported in the literature that drug nanocrystals are microcrystalline rather than amorphous.[@b46-ijn-10-2521],[@b47-ijn-10-2521] The in vitro drug release profile of the nanocrystals was comparable to that of the control (drug cosolvent) ([Figure 3D](#f3-ijn-10-2521){ref-type=\"fig\"}). ", "The f~2~ value (similarity factor) between the two preparations was calculated to be 85. ", "The release data indicate that release of the drug from the nanocrystals was rather rapid.", "\n\nPharmacokinetic and biodistribution profiles\n--------------------------------------------\n\nThe plasma concentration--time profiles displayed two distinct phases following intravenous administration of either the cosolvent (control) or nanocrystals ([Figure 5A](#f5-ijn-10-2521){ref-type=\"fig\"}). ", "These two phases appear to be the distribution (10 minutes--2 hours) and elimination (4--18 hours) phases. ", "It was also noted that there were no significant differences in plasma concentrations (at the time points after 10 minutes) between the nanocrystal and cosolvent groups (*P*\\>0.05; [Figure 5A](#f5-ijn-10-2521){ref-type=\"fig\"}). ", "The data were analyzed using the compartmental approach and found to be well described by the two-compartment model ([Figure 5A](#f5-ijn-10-2521){ref-type=\"fig\"} and [Table 2](#t2-ijn-10-2521){ref-type=\"table\"}). ", "The derived parameters of the area under the concentration--time curve (AUC), elimination half-life (t~1/2β~), and systemic clearance (CL) were comparable between the nanocrystal and cosolvent groups (*P*\\>0.05; [Table 2](#t2-ijn-10-2521){ref-type=\"table\"}). ", "This was consistent with the high similarity in the pharmacokinetic profile.", "\n\nThe AUCs in the heart, liver, spleen, lung, kidney, and intestine were estimated by non-compartmental analysis ([Table 3](#t3-ijn-10-2521){ref-type=\"table\"}). ", "The organs in which drug exposure was increased by greater than 20% were considered to exhibit a nanoparticle uptake effect. ", "An apparently higher AUC~(0--18\\ h)~ (≥27%) was found for the liver and spleen in the nanocrystal group ([Table 3](#t3-ijn-10-2521){ref-type=\"table\"}). ", "This resulted largely from the significantly higher concentrations at the initial time points (within 1 hour) ([Figure 5B](#f5-ijn-10-2521){ref-type=\"fig\"} and C). ", "Therefore, uptake of nanoparticles into the liver and spleen had to be considered in pharmacokinetic modeling. ", "By contrast, the AUC~(0--18\\ h)~ for the heart, lung, kidney, and intestine remained relatively unchanged or slightly changed ([Table 3](#t3-ijn-10-2521){ref-type=\"table\"}). ", "It was also noted that organ exposure was reduced in the intestine.", "\n\nNonparticulate PBPK modeling\n----------------------------\n\nThe nonparticulate PBPK model ([Figure 1A](#f1-ijn-10-2521){ref-type=\"fig\"}) was used to describe SNX-2112 disposition in rats following cosolvent administration. ", "The obtained pharmacokinetic parameters associated with the nonparticulate drug will be fixed in later nanoparticulate PBPK modeling. ", "The partitioning coefficient of SNX-2112 for those tissues to which the inlet drug concentrations equaled the plasma concentrations (ie, intestine, spleen, kidney, and heart) was determined experimentally as the ratio of AUC~(0--18\\ h)~ of tissue normalized to that of plasma AUC~(0--18\\ h)~.[@b48-ijn-10-2521] Tissue/Plasma AUC ratios of 12.2, 6.78, 5.92, and 2.24 were reached at 18 hours for the intestine, spleen, kidney, and heart, respectively ([Figure 6](#f6-ijn-10-2521){ref-type=\"fig\"} and [Table 4](#t4-ijn-10-2521){ref-type=\"table\"}). ", "The partition coefficients for other tissues (liver, lung, and remaining tissues) and the hepatic clearance were estimated by simultaneous fitting of the model to the experimental data ([Table 4](#t4-ijn-10-2521){ref-type=\"table\"}). ", "The observed values were largely close to the predicted ones from the model, indicating that the data were well described by the PBPK model ([Figure 7](#f7-ijn-10-2521){ref-type=\"fig\"}).", "\n\nNanoparticulate PBPK modeling\n-----------------------------\n\nThe nanoparticulate PBPK model ([Figure 1B](#f1-ijn-10-2521){ref-type=\"fig\"}) was specifically constructed to distinguish the nanoparticulate from the nonparticulate drug as nanocrystals were administered to the rats. ", "Parameters associated the nonparticulate form were taken as the corresponding estimates from nonparticulate PBPK modeling and were fixed during model fitting. ", "The nanoparticle-specific parameters (ie, *Up~sp~*, *Up~li~*, and *K~rel~*) were obtained by simultaneous fitting of the model to the experimental data ([Table 4](#t4-ijn-10-2521){ref-type=\"table\"}). ", "The observed values versus the predicted ones from the model were plotted ([Figure 8](#f8-ijn-10-2521){ref-type=\"fig\"}). ", "It is clear the pharmacokinetic data were adequately modeled. ", "In addition, the drug (in two forms) concentrations in the plasma, liver, and spleen were simulated ([Figure 9](#f9-ijn-10-2521){ref-type=\"fig\"}). ", "The results reveal that the nanocrystals rapidly released drug molecules in vivo, accounting for their cosolvent-like pharmacokinetic behaviors and biodistribution.", "\n\nDiscussion\n==========\n\nIn this study, we, for the first time as far as we are aware, developed and characterized a nanocrystal system for systemic delivery of SNX-2112, a novel Hsp90 inhibitor currently in clinical evaluation for cancer treatment. ", "Further, we elucidated the in vivo fate of SNX-2112 nanocrystals through pharmacokinetic determination and PBPK modeling. ", "Our results reveal that the nanocrystals provided a rapid drug release in vivo. ", "Also, significant uptake of nanocrystals into the spleen and liver occurred, accounting for the initial drug accumulation (within 1 hour) in these two tissues. ", "The present study and another indicate that PBPK modeling is a powerful tool to describe body disposition of nanoformulations, and to facilitate a better understanding of their in vivo fate.[@b37-ijn-10-2521] In addition, our nanocrystals solve the solubility problem of SNX-2112 and overcome the difficulties in drug administration when it comes to clinical trials with the drug itself (the cosolvent is prohibited due to a large amount of organic solvent). ", "Hence, the potential for nanocrystals to achieve the systemic delivery of SNX-2112 is worthy of further investigation in clinical trials.", "\n\nDue to the insoluble nature of SNX-2112, a nanocrystal formulation is an excellent choice for its delivery. ", "Attempts to develop other types of nanoformulations such as solid lipid nanoparticles and micro-/nanoemulsions for the compound were unsuccessful due to its low solubility in the oil phase. ", "It is well known that the stabilizer plays an important role in determining the properties of drug nanocrystals. ", "The authors of the present paper prepared SNX-2112 nanocrystals using different types of stabilizers including nonionic surfactants (Tween 80, P188, and P407), ionic surfactants (sodium dodecyl sulfate and sodium taurocholate hydrate), and polymers (hydroxypropyl methylcellulose and polyvinylpyrrolidone). ", "It was found that P188-stabilized nanocrystals showed acceptable particle size with narrow size distribution. ", "Preliminary stability studies also revealed that the prepared nanocrystals with P188 were stable for at least 1 month. ", "Hence, P188-stabilized nanocrystals are a promising system to effectively deliver SNX-2112.", "\n\nThe SNX2112 nanocrystals showed a negative zeta potential (−11.6 mV) and extensive distribution in the liver and spleen. ", "Hence, in addition to contributing to the physical stability of the formulation, the negative nature of the zeta potential may be a key determinant in the rate as well as the efficiency of monocyte/macrophage uptake (particularly in the liver and spleen). ", "This is well supported by many other studies in the literature.[@b49-ijn-10-2521],[@b50-ijn-10-2521] For instance, the oridonin nanosuspensions, with a zeta potential value of −21.8 mV, developed by Gao et al demonstrated a high uptake by the liver and spleen.[@b49-ijn-10-2521] Rabinow et al evaluated the biodistribution of an intravenous itraconazole nanosuspension (with a zeta potential of −30.6 mV) and found that high accumulation of the drug occurred in the liver and spleen.[@b50-ijn-10-2521]\n\nIn vitro release study showed that drug release from the nanocrystals was rapid and was comparable to that from cosolvent ([Figure 3D](#f3-ijn-10-2521){ref-type=\"fig\"}). ", "Parallel release profiles for both the free drug (cosolvent) and the nanocrystals indicate that the nanocrystals rapidly dissolved in the medium upon dilution. ", "However, drug release from the cosolvent was not as immediate as expected. ", "The time for complete drug release was unusually long (more than 4 hours). ", "This was because membrane resistance (the dialysis bag) had a retarding effect on the process of drug release. ", "As a mimic of in vivo release, the dialysis technique is frequently used to study the release kinetics of drug nanocrystals.[@b51-ijn-10-2521],[@b52-ijn-10-2521] Although the release data obtained by the dialysis method did not fully represent the real release behavior of drug nanocrystals in the blood stream, the method was useful for the characterization and comparative analyses of drug release from different formulations.", "\n\nIn nonparticulate PBPK modeling, the tissue/plasma AUC ratios versus time were evaluated and steady-state in vivo conditions were verified ([Figure 6](#f6-ijn-10-2521){ref-type=\"fig\"}). ", "The K~p~ values were determined as the tissue/plasma AUC ratios under steady-state conditions, as described previously.[@b46-ijn-10-2521] The obtained values were used within the PBPK model without further optimization.[@b46-ijn-10-2521] It was noteworthy that the authors were unable to determine K~p~ values using such a method for the liver and lung because plasma did not directly flow to these two organs (ie, inlet drug concentrations to the two organs did not equal the plasma concentrations) ([Figure 1A](#f1-ijn-10-2521){ref-type=\"fig\"}). ", "However, the authors plotted the tissue/plasma AUC ratios versus time and also observed steady-state conditions for the liver and lung (data not shown). ", "The AUC ratios toward 18 hours were 8.17 (for the liver) and 4.32 (for the lung), which markedly differed from the optimized K~p~ values ([Table 4](#t4-ijn-10-2521){ref-type=\"table\"}). ", "Therefore, our PBPK modeling was carefully and appropriately performed.", "\n\nOur results reveal that the nanocrystals were significantly taken up into the liver and spleen. ", "This is consistent with the notion that nanoparticles possess the ability of targeting the macrophage system and can accumulate in the liver and spleen.[@b53-ijn-10-2521] The pharmacokinetic curves of the nanocrystals and cosolvent show a high similarity ([Figure 5A](#f5-ijn-10-2521){ref-type=\"fig\"}). ", "The fact that the nanocrystals did not markedly alter the pharmacokinetic profile of SNX-2112 was due to immediate drug release from the nanocrystals upon intravenous administration ([Figure 9](#f9-ijn-10-2521){ref-type=\"fig\"}). ", "It is noteworthy that such cosolvent-like behaviors of drug nanocrystals in the body have been observed in human subjects.[@b54-ijn-10-2521] Since the nanocrystals rapidly dissolved into the free drug, minimal systemic risk would be associated with particulate injection, highlighting the great potential of the nanocrystal formulation in practical uses.", "\n\nConclusion\n==========\n\nBy using the wet-media milling technique, we developed and characterized a nanocrystal system, allowing the systemic delivery of the insoluble drug SNX-2112. ", "The obtained SNX-2112 nanocrystals were 203 nm in size with a zeta potential of −11.6 mv. ", "In addition, the nanocrystals showed a comparable release profile to that of the control (drug cosolvent). ", "Moreover, the nanoparticulate PBPK model incorporating the parameters of particulate uptake (into the liver and spleen) and of in vivo drug release was well fitted to the experimental data. ", "The results reveal that the nanocrystals rapidly released drug molecules in vivo, accounting for their cosolvent-like pharmacokinetic behaviors. ", "Due to particulate uptake, drug accumulation in the liver and spleen were significant at the initial time points (within 1 hour). ", "This study is the first to report that the nanocrystal system is a good choice for the systemic delivery of SNX-2112. ", "Also, our study contributes to an improved understanding of the in vivo fate of nanocrystals.", "\n\nThis work was supported by the National Natural Science Foundation of China (No 81373496), the Program for Pearl River New Stars of Science and Technology in Guangzhou (No 2014059), and the Doctoral Fund of Ministry of Education of China (20134401120014).", "\n\n**Disclosure**\n\nThe authors declare no conflicts of interest in this work.", "\n\n{#f1-ijn-10-2521}\n\n{#f2-ijn-10-2521}\n\n{#f3-ijn-10-2521}\n\n{#f4-ijn-10-2521}\n\n{#f5-ijn-10-2521}\n\n{#f6-ijn-10-2521}\n\n{#f7-ijn-10-2521}\n\n{#f8-ijn-10-2521}\n\n{#f9-ijn-10-2521}\n\n###### \n\nList of plasma flows and organ volumes for rats (250 g) taken from the literature[@b42-ijn-10-2521]\n\n Tissue Plasma flow (L/h) Organ volume (L)\n ----------- ------------------- ------------------\n Liver 0.4920 0.009150\n Kidney 0.3790 0.001830\n Spleen 0.0269 0.000500\n Intestine 0.3850 0.006750\n Lung 2.6900 0.001250\n Heart 0.1320 0.000825\n Remaining 1.2800 0.221000\n\n###### \n\nList of pharmacokinetic parameters derived by fitting the conventional two-compartment model to the plasma data\n\n Parameter Unit Cosolvent %CV Nanocrystals %CV\n ------------- ---------- ----------- ------- -------------- -------\n k~10~ 1/h 0.403 10.50 0.239 16.00\n k~12~ 1/h 0.850 18.80 0.352 45.80\n k~21~ 1/h 0.325 14.30 0.276 39.60\n t~1/2α~ h 0.465 15.00 0.886 35.20\n t~1/2β~ h 7.890 9.97 8.230 21.90\n C~0~ μM 1.930 9.79 1.110 13.20\n V L 0.558 9.80 0.965 13.20\n CL L/h 0.225 3.93 0.230 7.68\n V~2~ L 1.460 9.90 1.230 23.40\n CL~2~ L/h 0.474 12.30 0.339 36.70\n AUC~0--t~ μM\\*h 4.020 3.92 3.870 7.67\n AUC~0--inf~ μM\\*h 4.790 3.90 4.670 7.68\n AUMC μM\\*h^2^ 43.000 12.20 44.500 24.00\n MRT h 8.980 9.31 9.520 17.90\n V~ss~ L 2.020 7.47 2.190 13.50\n\n**Abbreviation:** %CV, coefficient of variation.", "\n\n###### \n\nNon-compartmental estimates of SNX-2112 parameters in tissues and plasma of cosolvent- and nanocrystal-treated rats\n\n Cosolvent Nanocrystal Nanocrystal/cosolvent\n ----------------------------------- ----------- ------------- -----------------------\n Plasma AUC~(0--18\\ h)~ (μM\\*h) 4.200 4.16 0.991\n Systemic clearance, CL~s~ (L/h) 0.256 0.259 1.010\n Heart AUC~(0--18\\ h)~ (μM\\*h) 9.070 10.500 1.160\n Liver AUC AUC~(0--18\\ h)~ (μM\\*h) 33.900 56.300 1.660\n Spleen AUC~(0--18\\ h)~ (μM\\*h) 29.000 36.800 1.270\n Lung AUC~(0--18\\ h)~ (μM\\*h) 17.900 19.700 1.100\n Kidney AUC~(0--18\\ h)~ (μM\\*h) 24.700 25.800 1.040\n Intestine AUC~(0--18\\ h)~ (μM\\*h) 52.400 44.100 0.842\n\n**Abbreviations:** AUC, area under the concentration--time curve; CL~s~, systemic clearance.", "\n\n###### \n\nEstimated pharmacokinetic parameters of SNX-2112 in rats using the physiologically based pharmacokinetic (PBPK) models following administration of SNX-2112 cosolvent or nanocrystals\n\n Parameter Unit Estimate %CV\n ----------- ------ --------------------------------------------------- -----------------------------------------------\n K~p,kd~ 5.9200 --[b](#tfn4-ijn-10-2521){ref-type=\"table-fn\"}\n K~p,sp~ 6.7800 --[b](#tfn4-ijn-10-2521){ref-type=\"table-fn\"}\n K~p,in~ 12.2000 --[b](#tfn4-ijn-10-2521){ref-type=\"table-fn\"}\n K~p,ht~ 2.2400 --[b](#tfn4-ijn-10-2521){ref-type=\"table-fn\"}\n K~p,li~ 4.0300[a](#tfn3-ijn-10-2521){ref-type=\"table-fn\"} 8.1\n K~p,lu~ 0.0213[a](#tfn3-ijn-10-2521){ref-type=\"table-fn\"} 9.5\n K~p,rm~ 5.1500[a](#tfn3-ijn-10-2521){ref-type=\"table-fn\"} 4.3\n CL~H~ L/h 0.0246[a](#tfn3-ijn-10-2521){ref-type=\"table-fn\"} 7.6\n Up~li~ L/h 0.0906 36.0\n Up~sp~ L/h 0.1820 26.0\n K~rel~ 1/h 0.0257 24.0\n\n**Notes:**\n\nValues estimated by nonparticulate PBPK modeling;\n\nfixed value.", "\n\n**Abbreviation:** %CV, coefficient of variation.", "\n\n[^1]: \\*These authors contributed equally to this work\n" ]

{ "pile_set_name": "PubMed Central" }

[ 0, 0.0136986301369863, 0.01308139534883721, 0, 0, 0.010852713178294573, 0.006012024048096192, 0, 0, 0.005494505494505495, 0, 0, 0, 0.007434944237918215, 0, 0, 0.003436426116838488, 0, 0, 0.007067137809187279, 0, 0, 0.01020408163265306, 0, 0, 0.0053475935828877, 0.006557377049180328, 0, 0, 0.010101010101010102, 0.004, 0, 0, 0, 0.006535947712418301, 0.008849557522123894, 0, 0, 0.0045045045045045045, 0, 0, 0.008064516129032258, 0, 0, 0, 0.01, 0.008064516129032258, 0.013157894736842105, 0.016666666666666666, 0, 0.007518796992481203, 0, 0.014285714285714285, 0, 0, 0, 0.012605042016806723, 0, 0, 0, 0.0017543859649122807, 0, 0, 0, 0, 0.007751937984496124, 0, 0, 0, 0, 0.002197802197802198, 0.012658227848101266, 0, 0, 0, 0, 0.0054446460980036296, 0, 0, 0.001584786053882726, 0, 0.006382978723404255, 0, 0, 0, 0.008403361344537815, 0, 0.006666666666666667, 0, 0, 0, 0, 0, 0, 0.003861003861003861, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.0018315018315018315, 0, 0, 0, 0, 0.005, 0, 0, 0, 0, 0, 0, 0, 0, 0.002178649237472767, 0, 0, 0, 0, 0.006514657980456026, 0, 0, 0, 0, 0.00390625, 0.008915304606240713, 0, 0, 0, 0, 0.004672897196261682, 0, 0.0072992700729927005, 0, 0.010810810810810811, 0, 0, 0.0033003300330033004, 0, 0.002824858757062147, 0, 0, 0, 0, 0, 0, 0, 0, 0.011673151750972763, 0, 0, 0.010526315789473684, 0.006666666666666667, 0, 0, 0, 0, 0.007633587786259542, 0.019230769230769232, 0.014457831325301205, 0.014598540145985401, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.007692307692307693, 0, 0, 0, 0, 0, 0.022727272727272728, 0.00684931506849315, 0.007407407407407408, 0, 0, 0, 0, 0, 0.022727272727272728, 0, 0.006097560975609756, 0, 0, 0.002938295788442703, 0.007329842931937173, 0.006056527590847914, 0, 0 ]

[ "[Distal tubal obstructions: microsurgery or in vitro fertilization].", "\nLaser CO2 microsurgical salpingostomies were performed among 70 infertile patients from 1979 to 1982. ", "9 patients lost to follow up were excluded. ", "All patients underwent a preoperative hysterosalpingography and a laparoscopy in order to realise a tubal evaluation. ", "4 groups of tubal damage were defined with a subsequent pregnancy rate of: 66.66% in group I, 37.5% in group II, 5.55% in group III, 0% in group IV. ", "4 groups of adhesions were defined with a subsequent pregnancy rate of: 42.85% in group 0, 33.33% in group mild, 25% in group moderate, 5.85% in group severe. ", "From this results the authors propose a tubal and adhesions score in order to select patients for microsurgery or in vitro fertilization." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0, 0, 0.013422818791946308, 0, 0 ]

[ "How Does Occlusion Training Work in Increasing muscle?", "\n\nBefore I answer this question in detail, I will try and explain exactly what occlusion training is, because not a lot of people would have heard of this before.", "\n\nOcclusion training, a.k.a. ", "blood flow restriction training (makes it easier to understand) in fitness circles, is a technique where while training with heavy weights the blood flow to the target muscles is restricted. ", "It is the limbs where its done usually. ", "Seems crazy right?", "\n\nHow can restriction in blood flow impact muscle growth at all?", "\n\nWell, it does help.", "\n\nSo in this article, I am going to tell you how occlusion training works, my personal experience with this training form of training, and who is this training most useful for.", "\n\nThe working of occlusion training\n\nWhen you go to the gym and your trainer asks you to take deep breaths in between each set, the primary reason behind it is to provide additional oxygen to your body through improved blood flow, resulting in a surge of energy. ", "Restricting the blood flow is just like going opposite to this approach, isn’t it? ", "It’d suffocate your body for energy, right?", "\n\nNow here lays the secret: You don’t have to completely restrict the blood flow,[1] but restrict it only in veins of those muscles which are the focus of your exercise.", "\n\nOcclusion training increases your muscles in three steps:\n\nIf you are a regular visitor to the gym, you would have heard the phrase ‘I am all pumped’ at least a dozen times. ", "Even I use this phrase a lot while showing off my muscles after the workout, even though the ‘pump’ usually dies out by morning. ", "Since the blood is flowing in and not out, your muscle cells would eventually reach a point where they are so packed with fluid that they would have just two options, either to burst out or grow.[2] Guess which one is not ugly? ", "Once the muscle cells are packed with fluid, there is a drop in oxygen levels. ", "The body starts deputing more muscle fibers to assist them in the job.[3] The involvement of more muscle fibers during the same old workout results in their growth. ", "Lower oxygen levels increase the amount of accumulated lactic acid (a.k.a the burn) in the muscles. ", "This may sound bad, but studies show that lactic acid accumulation is an indicator of protein breakdown and growth.[4] I hope you remember, that amino acids formed after protein breakdown are the building blocks of our muscles.", "\n\nBeginning to grow bigger with bands\n\nWhen you start with occlusion training, you use a restriction band or a wrap to apply pressure on the veins of the limb under focus and restrict the blood flow. ", "Pay a little attention here: I said restricting blood flow in the veins, which actually means that the muscles will still receive fresh blood. ", "It will just not flow back to the heart.[5]\n\nMy experience with occlusion training\n\nWhen I started going to the gym and saw some of the big guys training with weights while wearing bands on their biceps, I assumed\n\nthey meant for support. ", "I was working with light weights and it seemed awkward that I should use bands for support. ", "I never even bothered to clarify on the purpose they served. ", "Mainly because I was afraid. ", "I could clearly see that people using those bands were lifting almost thrice the weight I was.", "\n\nSix months in, I moved on to heavier weights and decided to use one of those bands. ", "For support.", "\n\nJust after a couple of exercises, I found my biceps twice as pumped as they usually get. ", "In addition, I was unable to complete my sets. ", "I couldn’t even complete half of my scheduled reps. ", "Plus, my muscles were awfully sore. ", "That is when I decided to ask about their real purpose. ", "That was the last day when I tried my hands on occlusion training. ", "I was never in for heavy muscle building, and that was exactly what occlusion is meant for.", "\n\nSo next time you say ‘I am all pumped’, try your exercises with one of those bands on and you’ll realize what getting pumped means.", "\n\nSome words of caution\n\nWorkout is all about being cautious and you have to do everything right in order to extract the maximum advantage. ", "Here are a few things that you should keep in mind.", "\n\nOcclusion is not for beginners. ", "This method of training should start when you already have strong muscles. ", "This actually means that you should not try it until you have been going to the gym for at least 6 months. ", "It will make your muscles sore and by the end of the exercise even the lighter weights would feel very heavy. ", "Occlusion is not at all beneficial for people who are in for just fitness or medium muscle building as this training methodology is going to result in rapid muscle build up provided your technique and diet are right. ", "You have to make sure that the band is used properly. ", "Wrapping it very tight is going to restrict flow in your arteries which should not happen. ", "Studies say that on a scale of 10 if you say that the grip of the band is 6-7, you should be fine.[6] Any tight would mean your arteries would get collapsed, and any loose would mean no benefit at all. ", "You should restrict your breaks between every set to less than 30 seconds.", "\n\nIf you are really serious about muscle building and have been going to the gym for some time, try out this technique and share below what change you observed.", "\n\nRelated Articles\n\nLast Updated: May 26, 2014\n\nNext Scheduled Update: Jul 26, 2014" ]

{ "pile_set_name": "OpenWebText2" }

[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.0049504950495049506, 0, 0, 0 ]

[ "Q:\n\nhow to make a h2 background image 'dynamic' in height as well\n\nI have a 1px wide by 35px high image, that I set as background for my h2 elements. ", "The accompanying CSS is as follows:\nh2 {\n background:transparent url(./images/h1_bg.gif) repeat-x left top;\n font-size:18px;\n}\n\nAs a result, the background is set as the image no matter what the length of the heading is, as long as the heading fits in one line. ", "If the heading spills over to next line, then\nimage is not set as background for the wider area, as it repeats only in x direction. ", "I think I need to create another image, example 1px wide by 60px high image and set it as background of the two-line heading. ", "But I need some way to check if the heading is a two-line heading or not, so I can change the image background. ", "How do I do that? ", "Plus, is there some other way to accomplish this task?", "\n\nA:\n\nYou usually deal with this by just creating a very tall image, like 100 pixels or more. ", "Hold the bottom row colour for the remainder of the height.", "\n\nA:\n\nYou can do this by using height.", "\n if($j(\"h1\").height() > 30)\n {\n $j(\"h1\").css(\"background\",\"transparent url(./images/h1_bg_0.gif) repeat-x left top\")\n }\n else\n {\n $j(\"h1\").css(\"background\",\"transparent url(./images/h1_bg_1.gif) repeat-x left top\")\n }\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0, 0.0037313432835820895, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]

[ "5 Unique Valentine's Day Gifts Ideas for Her\n\nPosted by bigsmall.in on Dec 15, 2016\n\nWhen your girl is already making guesses for the gift you’re going to give her this Valentine’s Day, you find yourself in a dilemma of the century! ", "Ain’t you? ", "You know gift selection has never been your forte. ", "You’ll curse yourself for not being good at the job. ", "But there is one thing you know for sure- risking it is not an option this time.", "\n\nYour girl has always done a lot for you and this time you’d want to make up to her, and express your love for her though your gift. ", "That’s the plan, right? ", "We know it’s always easier said than done and you are battling inside, wondering what will be a perfect gift for your girlfriend/wife this Valentine’s day.", "\n\nWe, at bigsmall.in are resolute to help you make your girl happiest she even can. ", "And that’s why we have brought for your love most creative, unique and beautiful presents to make her day.", "\n\nTake away all the darkness and illuminate her world with the lights of love! ", "Fill her life with the colours of the world just like this lamp and let your lady love know how much you care for her happiness.", "\n\nDoes your wife/girlfriend always complain about how hard it is to get the nail polish right; especially when she has to take care of the tiny bottle? ", "It’s time for you to gift her nail polish holder and bring an end to all her problems!", "\n\nDoes she work for long hours? ", "Or does she have a habit of wearing high heels? ", "If your girl is a workaholic who hardly gets time to relax, then gift her spa socks and give her feet some comfort while she takes a break from a long day!", "\n\nLet the mermaid keep your girl safe for you! ", "This cozy mermaid blanket gives comfort wherever she may be. ", "Be is watching a movie, reading a book or taking a nap, the cute tail would make her fall in love with its awesomeness. ", "So, gift her this blanket and make her the mermaid princess!" ]

{ "pile_set_name": "Pile-CC" }

[ 0.004291845493562232, 0, 0, 0, 0, 0, 0, 0.0064516129032258064, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]

[ "#!", "/usr/bin/env python\n# -*- coding: utf-8 -*-\n\nimport math\nimport torch\nimport torch.nn as nn\nimport torch.utils.model_zoo as model_zoo\nimport torch.autograd as autograd\nfrom torch.autograd.variable import Variable\nfrom threading import Lock\nfrom torch.distributions import Categorical\n\nglobal_lock = Lock()\n\n\nmodel_urls = {\n 'resnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',\n 'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',\n 'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',\n 'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',\n 'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',\n}\n\n\ndef conv3x3(in_planes, out_planes, stride=1):\n \"\"\"3x3 convolution with padding\"\"\"\n return nn.", "Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,\n padding=1, bias=False)\n\n\nclass BasicBlock(nn.", "Module):\n expansion = 1\n\n def __init__(self, inplanes, planes, stride=1, downsample=None):\n super(BasicBlock, self).__init__()\n self.conv1 = conv3x3(inplanes, planes, stride)\n self.bn1 = nn.", "BatchNorm2d(planes)\n self.relu = nn.", "ReLU(inplace=True)\n self.conv2 = conv3x3(planes, planes)\n self.bn2 = nn.", "BatchNorm2d(planes)\n self.downsample = downsample\n self.stride = stride\n\n def forward(self, x):\n residual = x\n\n out = self.conv1(x)\n out = self.bn1(out)\n out = self.relu(out)\n\n out = self.conv2(out)\n out = self.bn2(out)\n\n if self.downsample is not None:\n residual = self.downsample(x)\n\n out += residual\n out = self.relu(out)\n\n return out\n\n\nclass Bottleneck(nn.", "Module):\n expansion = 4\n\n def __init__(self, inplanes, planes, stride=1, downsample=None):\n super(Bottleneck, self).__init__()\n self.conv1 = nn.", "Conv2d(inplanes, planes, kernel_size=1, bias=False)\n self.bn1 = nn.", "BatchNorm2d(planes)\n self.conv2 = nn.", "Conv2d(planes, planes, kernel_size=3, stride=stride,\n padding=1, bias=False)\n self.bn2 = nn.", "BatchNorm2d(planes)\n self.conv3 = nn.", "Conv2d(planes, planes * 4, kernel_size=1, bias=False)\n self.bn3 = nn.", "BatchNorm2d(planes * 4)\n self.relu = nn.", "ReLU(inplace=True)\n self.downsample = downsample\n self.stride = stride\n\n def forward(self, x):\n residual = x\n\n out = self.conv1(x)\n out = self.bn1(out)\n out = self.relu(out)\n\n out = self.conv2(out)\n out = self.bn2(out)\n out = self.relu(out)\n\n out = self.conv3(out)\n out = self.bn3(out)\n\n if self.downsample is not None:\n residual = self.downsample(x)\n\n out += residual\n out = self.relu(out)\n\n return out\n\n\n# ==============================\n# Original Model without Gating\n# ==============================\n\nclass ResNet(nn.", "Module):\n def __init__(self, block, layers, num_classes=1000):\n self.inplanes = 64\n super(ResNet, self).__init__()\n self.conv1 = nn.", "Conv2d(3, 64, kernel_size=7, stride=2, padding=3,\n bias=False)\n self.bn1 = nn.", "BatchNorm2d(64)\n self.relu = nn.", "ReLU(inplace=True)\n self.maxpool = nn.", "MaxPool2d(kernel_size=3, stride=2, padding=1)\n self.layer1 = self._make_layer(block, 64, layers[0])\n self.layer2 = self._make_layer(block, 128, layers[1], stride=2)\n self.layer3 = self._make_layer(block, 256, layers[2], stride=2)\n self.layer4 = self._make_layer(block, 512, layers[3], stride=2)\n self.avgpool = nn.", "AvgPool2d(7)\n self.fc = nn.", "Linear(512 * block.expansion, num_classes)\n\n for m in self.modules():\n if isinstance(m, nn.", "Conv2d):\n n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels\n m.weight.data.normal_(0, math.sqrt(2. / ", "n))\n elif isinstance(m, nn.", "BatchNorm2d):\n m.weight.data.fill_(1)\n m.bias.data.zero_()\n\n def _make_layer(self, block, planes, blocks, stride=1):\n downsample = None\n if stride !", "= 1 or self.inplanes !", "= planes * block.expansion:\n downsample = nn.", "Sequential(\n nn.", "Conv2d(self.inplanes, planes * block.expansion,\n kernel_size=1, stride=stride, bias=False),\n nn.", "BatchNorm2d(planes * block.expansion),\n )\n\n layers = []\n layers.append(block(self.inplanes, planes, stride, downsample))\n self.inplanes = planes * block.expansion\n for i in range(1, blocks):\n layers.append(block(self.inplanes, planes))\n\n return nn.", "Sequential(*layers)\n\n def forward(self, x):\n x = self.conv1(x)\n x = self.bn1(x)\n x = self.relu(x)\n x = self.maxpool(x)\n\n x = self.layer1(x)\n x = self.layer2(x)\n x = self.layer3(x)\n x = self.layer4(x)\n\n x = self.avgpool(x)\n x = x.view(x.size(0), -1)\n x = self.fc(x)\n return x\n\n\ndef resnet18(pretrained=False, **kwargs):\n \"\"\"Constructs a ResNet-18 model.", "\n Args:\n pretrained (bool): If True, returns a model pre-trained on ImageNet\n \"\"\"\n model = ResNet(BasicBlock, [2, 2, 2, 2], **kwargs)\n if pretrained:\n model.load_state_dict(model_zoo.load_url(model_urls['resnet18']))\n return model\n\n\ndef resnet34(pretrained=False, **kwargs):\n \"\"\"Constructs a ResNet-34 model.", "\n Args:\n pretrained (bool): If True, returns a model pre-trained on ImageNet\n \"\"\"\n model = ResNet(BasicBlock, [3, 4, 6, 3], **kwargs)\n if pretrained:\n model.load_state_dict(model_zoo.load_url(model_urls['resnet34']))\n return model\n\n\ndef resnet50(pretrained=False, **kwargs):\n \"\"\"Constructs a ResNet-50 model.", "\n Args:\n pretrained (bool): If True, returns a model pre-trained on ImageNet\n \"\"\"\n model = ResNet(Bottleneck, [3, 4, 6, 3], **kwargs)\n if pretrained:\n model.load_state_dict(model_zoo.load_url(model_urls['resnet50']))\n return model\n\n\ndef resnet74(pretrained=False, **kwargs):\n \"\"\" ResNet-74\"\"\"\n model = ResNet(Bottleneck, [3, 4, 14, 3], **kwargs)\n return model\n\n\ndef resnet101(pretrained=False, **kwargs):\n \"\"\"Constructs a ResNet-101 model.", "\n Args:\n pretrained (bool): If True, returns a model pre-trained on ImageNet\n \"\"\"\n model = ResNet(Bottleneck, [3, 4, 23, 3], **kwargs)\n if pretrained:\n model.load_state_dict(model_zoo.load_url(model_urls['resnet101']))\n return model\n\n\ndef resnet152(pretrained=False, **kwargs):\n \"\"\"Constructs a ResNet-152 model.", "\n Args:\n pretrained (bool): If True, returns a model pre-trained on ImageNet\n \"\"\"\n model = ResNet(Bottleneck, [3, 8, 36, 3], **kwargs)\n if pretrained:\n model.load_state_dict(model_zoo.load_url(model_urls['resnet152']))\n return model\n\n\n# ======================\n# Recurrent Gate Design\n# ======================\n\ndef repackage_hidden(h):\n if type(h) == Variable:\n return Variable(h.data)\n else:\n return tuple(repackage_hidden(v) for v in h)\n\n\nclass RNNGate(nn.", "Module):\n \"\"\"given the fixed input size, return a single layer lstm \"\"\"\n def __init__(self, input_dim, hidden_dim, rnn_type='lstm'):\n super(RNNGate, self).__init__()\n self.rnn_type = rnn_type\n self.input_dim = input_dim\n self.hidden_dim = hidden_dim\n\n if self.rnn_type == 'lstm':\n self.rnn = nn.", "LSTM(input_dim, hidden_dim)\n else:\n self.rnn = None\n self.hidden = None\n\n # reduce dim\n self.proj = nn.", "Conv2d(in_channels=hidden_dim, out_channels=1,\n kernel_size=1, stride=1)\n self.prob = nn.", "Sigmoid()\n\n def init_hidden(self, batch_size):\n # Before we've done anything, we dont have any hidden state.", "\n # Refer to the Pytorch documentation to see exactly\n # why they have this dimensionality.", "\n # The axes semantics are (num_layers, minibatch_size, hidden_dim)\n return (autograd.", "Variable(torch.zeros(1, batch_size,\n self.hidden_dim).cuda()),\n autograd.", "Variable(torch.zeros(1, batch_size,\n self.hidden_dim).cuda()))\n\n def repackage_hidden(self):\n self.hidden = repackage_hidden(self.hidden)\n\n def forward(self, x):\n batch_size = x.size(0)\n self.rnn.flatten_parameters()\n\n out, self.hidden = self.rnn(x.view(1, batch_size, -1), self.hidden)\n\n out = out.squeeze()\n proj = self.proj(out.view(out.size(0), out.size(1), 1, 1,)).squeeze()\n prob = self.prob(proj)\n\n disc_prob = (prob > 0.5).float().detach() - prob.detach() + prob\n disc_prob = disc_prob.view(batch_size, 1, 1, 1)\n return disc_prob, prob\n\n\n# =======================\n# Recurrent Gate Model\n# =======================\nclass RecurrentGatedResNet(nn.", "Module):\n def __init__(self, block, layers, num_classes=1000, embed_dim=10,\n hidden_dim=10, gate_type='rnn', **kwargs):\n self.inplanes = 64\n super(RecurrentGatedResNet, self).__init__()\n\n self.num_layers = layers\n self.conv1 = nn.", "Conv2d(3, 64, kernel_size=7, stride=2,\n padding=3, bias=False)\n self.bn1 = nn.", "BatchNorm2d(64)\n self.relu = nn.", "ReLU(inplace=True)\n self.maxpool = nn.", "MaxPool2d(kernel_size=3, stride=2, padding=1)\n\n self.embed_dim = embed_dim\n self.hidden_dim = hidden_dim\n\n # going to have 4 groups of layers. ", "For the easiness of skipping,\n # We are going to break the sequential of layers into a list of layers.", "\n self._make_group(block, 64, layers[0], group_id=1, pool_size=56)\n self._make_group(block, 128, layers[1], group_id=2, pool_size=28)\n self._make_group(block, 256, layers[2], group_id=3, pool_size=14)\n self._make_group(block, 512, layers[3], group_id=4, pool_size=7)\n\n if gate_type == 'rnn':\n self.control = RNNGate(embed_dim, hidden_dim, rnn_type='lstm')\n else:\n print('gate type {} not implemented'.format(gate_type))\n self.control = None\n\n self.avgpool = nn.", "AvgPool2d(7)\n self.fc = nn.", "Linear(512 * block.expansion, num_classes)\n\n for m in self.modules():\n if isinstance(m, nn.", "Conv2d):\n n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels\n m.weight.data.normal_(0, math.sqrt(2. / ", "n))\n elif isinstance(m, nn.", "BatchNorm2d):\n m.weight.data.fill_(1)\n m.bias.data.zero_()\n elif isinstance(m, nn.", "Linear):\n n = m.weight.size(0) * m.weight.size(1)\n m.weight.data.normal_(0, math.sqrt(2. / ", "n))\n\n def _make_group(self, block, planes, layers, group_id=1, pool_size=56):\n \"\"\" Create the whole group \"\"\"\n for i in range(layers):\n if group_id > 1 and i == 0:\n stride = 2\n else:\n stride = 1\n\n meta = self._make_layer_v2(block, planes, stride=stride,\n pool_size=pool_size)\n\n setattr(self, 'group{}_ds{}'.format(group_id, i), meta[0])\n setattr(self, 'group{}_layer{}'.format(group_id, i), meta[1])\n setattr(self, 'group{}_gate{}'.format(group_id, i), meta[2])\n\n def _make_layer_v2(self, block, planes, stride=1, pool_size=56):\n \"\"\" create one block and optional a gate module \"\"\"\n downsample = None\n if stride !", "= 1 or self.inplanes !", "= planes * block.expansion:\n downsample = nn.", "Sequential(\n nn.", "Conv2d(self.inplanes, planes * block.expansion,\n kernel_size=1, stride=stride, bias=False),\n nn.", "BatchNorm2d(planes * block.expansion),\n\n )\n layer = block(self.inplanes, planes, stride, downsample)\n self.inplanes = planes * block.expansion\n\n # this is for having the same input dimension to rnn gate.", "\n gate_layer = nn.", "Sequential(\n nn.", "AvgPool2d(pool_size),\n nn.", "Conv2d(in_channels=planes * block.expansion,\n out_channels=self.embed_dim,\n kernel_size=1,\n stride=1))\n if downsample:\n return downsample, layer, gate_layer\n else:\n return None, layer, gate_layer\n\n def repackage_hidden(self):\n self.control.hidden = repackage_hidden(self.control.hidden)\n\n def forward(self, x):\n \"\"\"mask_values is for the test random gates\"\"\"\n # pdb.set_trace()\n\n batch_size = x.size(0)\n x = self.conv1(x)\n x = self.bn1(x)\n x = self.relu(x)\n x = self.maxpool(x)\n\n # reinitialize hidden units\n self.control.hidden = self.control.init_hidden(batch_size)\n\n masks = []\n gprobs = []\n # must pass through the first layer in first group\n x = getattr(self, 'group1_layer0')(x)\n # gate takes the output of the current layer\n gate_feature = getattr(self, 'group1_gate0')(x)\n mask, gprob = self.control(gate_feature)\n gprobs.append(gprob)\n masks.append(mask.squeeze())\n prev = x # input of next layer\n\n for g in range(4):\n for i in range(0 + int(g == 0), self.num_layers[g]):\n if getattr(self, 'group{}_ds{}'.format(g+1, i)) is not None:\n prev = getattr(self, 'group{}_ds{}'.format(g+1, i))(prev)\n x = getattr(self, 'group{}_layer{}'.format(g+1, i))(x)\n prev = x = mask.expand_as(x)*x + (1-mask).expand_as(prev)*prev\n gate_feature = getattr(self, 'group{}_gate{}'.format(g+1, i))(x)\n mask, gprob = self.control(gate_feature)\n if not (g == 3 and i == (self.num_layers[3]-1)):\n # not add the last mask to masks\n gprobs.append(gprob)\n masks.append(mask.squeeze())\n\n x = self.avgpool(x)\n x = x.view(x.size(0), -1)\n x = self.fc(x)\n\n return x, masks, gprobs, self.control.hidden\n\n\ndef imagenet_rnn_gate_18(pretrained=False, **kwargs):\n \"\"\" Construct SkipNet-18 + SP \"\"\"\n model = RecurrentGatedResNet(BasicBlock, [2, 2, 2, 2],\n embed_dim=10, hidden_dim=10, gate_type='rnn')\n return model\n\n\ndef imagenet_rnn_gate_34(pretrained=False, **kwargs):\n \"\"\" Construct SkipNet-34 + SP \"\"\"\n model = RecurrentGatedResNet(BasicBlock, [3, 4, 6, 3],\n embed_dim=10, hidden_dim=10, gate_type='rnn')\n return model\n\n\ndef imagenet_rnn_gate_50(pretrained=False, **kwargs):\n \"\"\" Construct SkipNet-50 + SP \"\"\"\n model = RecurrentGatedResNet(Bottleneck, [3, 4, 6, 3],\n embed_dim=10, hidden_dim=10, gate_type='rnn')\n return model\n\n\ndef imagenet_rnn_gate_101(pretrained=False, **kwargs):\n \"\"\" Constructs SkipNet-101 + SP \"\"\"\n model = RecurrentGatedResNet(Bottleneck, [3, 4, 23, 3],\n embed_dim=10, hidden_dim=10, gate_type='rnn')\n return model\n\n\ndef imagenet_rnn_gate_152(pretrained=False, **kwargs):\n \"\"\"Constructs SkipNet-152 + SP \"\"\"\n model = RecurrentGatedResNet(Bottleneck, [3, 8, 36, 3],\n embed_dim=10, hidden_dim=10, gate_type='rnn')\n return model\n\n\n# =============================\n# Recurrent Gate Model with RL\n# =============================\n\nclass RNNGatePolicy(nn.", "Module):\n def __init__(self, input_dim, hidden_dim, rnn_type='lstm'):\n super(RNNGatePolicy, self).__init__()\n\n self.rnn_type = rnn_type\n self.input_dim = input_dim\n self.hidden_dim = hidden_dim\n\n if self.rnn_type == 'lstm':\n self.rnn = nn.", "LSTM(input_dim, hidden_dim)\n else:\n self.rnn = None\n self.hidden = None\n\n self.proj = nn.", "Conv2d(in_channels=hidden_dim, out_channels=1,\n kernel_size=1, stride=1)\n self.prob = nn.", "Sigmoid()\n\n def hotter(self, t):\n self.proj.weight.data /= t\n self.proj.bias.data /= t\n\n def init_hidden(self, batch_size):\n # Before we've done anything, we dont have any hidden state.", "\n # Refer to the Pytorch documentation to see exactly\n # why they have this dimensionality.", "\n # The axes semantics are (num_layers, minibatch_size, hidden_dim)\n return (autograd.", "Variable(torch.zeros(1, batch_size,\n self.hidden_dim).cuda()),\n autograd.", "Variable(torch.zeros(1, batch_size,\n self.hidden_dim).cuda()))\n\n def repackage_hidden(self):\n self.hidden = repackage_hidden(self.hidden)\n\n def forward(self, x):\n batch_size = x.size(0)\n self.rnn.flatten_parameters()\n out, self.hidden = self.rnn(x.view(1, batch_size, -1), self.hidden)\n\n out = out.squeeze()\n out = out.view(out.size(0), out.size(1), 1, 1)\n proj = self.proj(out).squeeze()\n prob = self.prob(proj)\n bi_prob = torch.stack([1-prob, prob]).t()\n\n # do action selection in the forward pass\n if self.training:\n # action = bi_prob.multinomial()\n dist = Categorical(bi_prob)\n action = dist.sample()\n else:\n dist = None\n action = (prob > 0.5).float()\n action_reshape = action.view(action.size(0), 1, 1, 1).float()\n return action_reshape, prob, action, dist\n\n\n# ================================\n# Recurrent Gate Model with RL\n# ================================\nclass RecurrentGatedRLResNet(nn.", "Module):\n def __init__(self, block, layers, num_classes=1000, embed_dim=10,\n hidden_dim=10, **kwargs):\n self.inplanes = 64\n super(RecurrentGatedRLResNet, self).__init__()\n\n self.num_layers = layers\n self.conv1 = nn.", "Conv2d(3, 64, kernel_size=7, stride=2,\n padding=3, bias=False)\n self.bn1 = nn.", "BatchNorm2d(64)\n self.relu = nn.", "ReLU(inplace=True)\n self.maxpool = nn.", "MaxPool2d(kernel_size=3, stride=2, padding=1)\n\n self.embed_dim = embed_dim\n self.hidden_dim = hidden_dim\n\n # going to have 4 groups of layers. ", "For the easiness of skipping,\n # We are going to break the sequential of layers into a list of layers.", "\n self._make_group(block, 64, layers[0], group_id=1, pool_size=56)\n self._make_group(block, 128, layers[1], group_id=2, pool_size=28)\n self._make_group(block, 256, layers[2], group_id=3, pool_size=14)\n self._make_group(block, 512, layers[3], group_id=4, pool_size=7)\n\n self.control = RNNGatePolicy(embed_dim, hidden_dim, rnn_type='lstm')\n\n self.avgpool = nn.", "AvgPool2d(7)\n self.fc = nn.", "Linear(512 * block.expansion, num_classes)\n\n self.softmax = nn.", "Softmax()\n\n # save everything\n self.saved_actions = {}\n self.saved_dists = {}\n self.saved_outputs = {}\n self.saved_targets = {}\n\n for m in self.modules():\n if isinstance(m, nn.", "Conv2d):\n n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels\n m.weight.data.normal_(0, math.sqrt(2. / ", "n))\n elif isinstance(m, nn.", "BatchNorm2d):\n m.weight.data.fill_(1)\n m.bias.data.zero_()\n elif isinstance(m, nn.", "Linear):\n n = m.weight.size(0) * m.weight.size(1)\n m.weight.data.normal_(0, math.sqrt(2. / ", "n))\n m.bias.data.zero_()\n\n def _make_group(self, block, planes, layers, group_id=1, pool_size=56):\n \"\"\" Create the whole group\"\"\"\n for i in range(layers):\n if group_id > 1 and i == 0:\n stride = 2\n else:\n stride = 1\n\n meta = self._make_layer_v2(block, planes, stride=stride,\n pool_size=pool_size)\n\n setattr(self, 'group{}_ds{}'.format(group_id, i), meta[0])\n setattr(self, 'group{}_layer{}'.format(group_id, i), meta[1])\n setattr(self, 'group{}_gate{}'.format(group_id, i), meta[2])\n\n def _make_layer_v2(self, block, planes, stride=1, pool_size=56):\n \"\"\" create one block and optional a gate module \"\"\"\n downsample = None\n if stride !", "= 1 or self.inplanes !", "= planes * block.expansion:\n downsample = nn.", "Sequential(\n nn.", "Conv2d(self.inplanes, planes * block.expansion,\n kernel_size=1, stride=stride, bias=False),\n nn.", "BatchNorm2d(planes * block.expansion),\n )\n layer = block(self.inplanes, planes, stride, downsample)\n self.inplanes = planes * block.expansion\n\n gate_layer = nn.", "Sequential(\n nn.", "AvgPool2d(pool_size),\n nn.", "Conv2d(in_channels=planes * block.expansion,\n out_channels=self.embed_dim,\n kernel_size=1,\n stride=1))\n return downsample, layer, gate_layer\n\n def forward(self, x, target_var, reinforce=False):\n batch_size = x.size(0)\n x = self.conv1(x)\n x = self.bn1(x)\n x = self.relu(x)\n x = self.maxpool(x)\n\n # reinitialize hidden units\n self.control.hidden = self.control.init_hidden(batch_size)\n\n masks = []\n gprobs = []\n actions = []\n dists = []\n\n # must pass through the first layer in first group\n x = getattr(self, 'group1_layer0')(x)\n # gate takes the output of the current layer\n gate_feature = getattr(self, 'group1_gate0')(x)\n\n mask, gprob, action, dist = self.control(gate_feature)\n gprobs.append(gprob)\n masks.append(mask.squeeze())\n prev = x # input of next layer\n\n current_device = torch.cuda.current_device()\n actions.append(action)\n dists.append(dist)\n\n for g in range(4):\n for i in range(0 + int(g == 0), self.num_layers[g]):\n if getattr(self, 'group{}_ds{}'.format(g+1, i)) is not None:\n prev = getattr(self, 'group{}_ds{}'.format(g+1, i))(prev)\n x = getattr(self, 'group{}_layer{}'.format(g+1, i))(x)\n prev = x = mask.expand_as(x)*x + (1-mask).expand_as(prev)*prev\n if not (g == 3 and (i == self.num_layers[g] - 1)):\n gate_feature = getattr(self,\n 'group{}_gate{}'.format(g+1, i))(x)\n mask, gprob, action, dist = self.control(gate_feature)\n gprobs.append(gprob)\n masks.append(mask.squeeze())\n actions.append(action)\n dists.append(dist)\n\n x = self.avgpool(x)\n x = x.view(x.size(0), -1)\n x = self.fc(x)\n\n if reinforce:\n softmax = self.softmax(x)\n # action = softmax.multinomial()\n dist = Categorical(softmax)\n action = dist.sample()\n actions.append(action)\n dists.append(dist)\n\n with global_lock:\n self.saved_actions[current_device] = actions\n self.saved_outputs[current_device] = x\n self.saved_targets[current_device] = target_var\n self.saved_dists[current_device] = dists\n\n return x, masks, gprobs, self.control.hidden\n\n\ndef imagenet_rnn_gate_rl_18(pretrained=False, **kwargs):\n \"\"\" Construct SkipNet-18 + HRL.", "\n has the same architecture as SkipNet-18+SP \"\"\"\n model = RecurrentGatedRLResNet(BasicBlock, [2, 2, 2, 2], embed_dim=10,\n hidden_dim=10, gate_type='rnn')\n return model\n\n\ndef imagenet_rnn_gate_rl_34(pretrained=False, **kwargs):\n \"\"\" Construct SkipNet-34 + HRL.", "\n has the same architecture as SkipNet-34+SP \"\"\"\n model = RecurrentGatedRLResNet(BasicBlock, [3, 4, 6, 3], embed_dim=10,\n hidden_dim=10, gate_type='rnn')\n return model\n\n\ndef imagenet_rnn_gate_rl_50(pretrained=False, **kwargs):\n \"\"\" Construct SkipNet-50 + HRL.", "\n has the same architecture as SkipNet-50+SP \"\"\"\n model = RecurrentGatedRLResNet(Bottleneck, [3, 4, 6, 3], embed_dim=10,\n hidden_dim=10, gate_type='rnn')\n return model\n\n\ndef imagenet_rnn_gate_rl_101(pretrained=False, **kwargs):\n \"\"\" Construct SkipNet-101 + HRL.", "\n has the same architecture as SkipNet-101+SP \"\"\"\n model = RecurrentGatedRLResNet(Bottleneck, [3, 4, 23, 3], embed_dim=10,\n hidden_dim=10, gate_type='rnn')\n return model\n\n\ndef imagenet_rnn_gate_rl_152(pretrained=False, **kwargs):\n \"\"\" Construct SkipNet-152 + HRL.", "\n has the same architecture as SkipNet-152+SP \"\"\"\n model = RecurrentGatedRLResNet(Bottleneck, [3, 8, 36, 3], embed_dim=10,\n hidden_dim=10, gate_type='rnn')\n return model\n\n\n" ]

{ "pile_set_name": "Github" }

[ 0, 0.011042944785276074, 0.015873015873015872, 0.009216589861751152, 0, 0.03488372093023256, 0, 0.012195121951219513, 0.02702702702702703, 0, 0.023255813953488372, 0, 0.02631578947368421, 0, 0.003125, 0.00641025641025641, 0.017391304347826087, 0.02564102564102564, 0.022222222222222223, 0.0028653295128939827, 0, 0.009174311926605505, 0, 0, 0.020512820512820513, 0, 0, 0, 0.007352941176470588, 0, 0.011312217194570135, 0.008823529411764706, 0.011764705882352941, 0.008316008316008316, 0.008695652173913044, 0.005917159763313609, 0.011527377521613832, 0.007042253521126761, 0.008, 0.00847457627118644, 0, 0, 0, 0.0038860103626943004, 0.007220216606498195, 0.017391304347826087, 0.02564102564102564, 0.022222222222222223, 0.006097560975609756, 0, 0.009208103130755065, 0, 0.009174311926605505, 0, 0, 0.016260162601626018, 0.01652892561983471, 0.0025380710659898475, 0, 0, 0, 0.007352941176470588, 0, 0, 0, 0, 0.004104368220463207, 0.013888888888888888, 0.008264462809917356, 0.008, 0.0047169811320754715, 0, 0, 0, 0.0018148820326678765, 0.007633587786259542, 0.017391304347826087, 0.02564102564102564, 0.022222222222222223, 0.006097560975609756, 0, 0.015, 0, 0.014285714285714285, 0, 0, 0, 0.016260162601626018, 0.01652892561983471, 0.002430133657351154, 0, 0, 0, 0.007352941176470588, 0, 0, 0, 0.003017729158807997, 0.003278688524590164, 0.003278688524590164, 0.003246753246753247, 0.0032258064516129032, 0 ]

[ "Legal\n\nEnergy Efficient Replacement Windows\n\nKrypton or Argon-Filled, Low-E, Double-Hung Energy Star Units\n\nWindows aren’t what they used to be. ", "Plenty of contractor-grade windows are installed by residential building contractors, but an upgrade to a krypton or argon-filled, low-e, double-hung model will quickly pay for itself.", "\n\nA residential energy audit will quickly point out which windows will benefit most from replacement.", "\n\nWhy Krypton or Argon-Filled Windows?", "\n\nArgon is in fact a very useful gas. ", "It’s used inside incandescent light bulbs to keep oxygen from degrading the tungsten filament. ", "It’s used in processing the silicon to make the semiconductors that make reading this article possible.", "\n\nWith the advent of double-paned windows, manufacturers originally flushed the cavity with nitrogen or simply filled it with air. ", "But they quickly discovered that a slower-moving gas reduces conduction, thus minimizing the transfer of heat and cold from the inside to the outside of the dwelling.", "\n\nBoth krypton and Argon fit the bill. ", "Both are non-toxic which places them in the green building arena. ", "Krypton is the more efficient of the two gases, but more expensive, so it is used in cases where the cavity space between the double panes needs to be reduced.", "\n\nLow-Emittance (Low-E) Coatings as a Radiant Barrier\n\nLow-e coatings are metal or metallic oxide layers that are microscopically thin and almost invisible to the naked eye. ", "The manufacturer applies them to window and skylight glazings. ", "The end result is a reduction of the U-factor by quelling radiant heat flow.", "\n\nIn this respect, low-e coatings, combined with a gas-filled void, function in a manner similar to radiant barrier foil or paint used in conjunction with attic insulation.", "\n\nThere are two categories of coatings which should be taken into account when selecting energy-efficient windows. ", "High solar gain glazing is more effective in winter, but low solar gain (also called sputtered) wins the prize during the summer.", "\n\nThe home’s dominant weather pattern should determine which type will result in lower utility bills.", "\n\nThe Role of Window Frame Material\n\nMany homeowners focus exclusively on the window and its thermal properties. ", "But consider that the frame itself comprises anywhere from 10 to 30 percent of the unit’s surface area. ", "It’s no surprise that the material has an impact on efficiency.", "\n\n• Aluminum: This material is extruded so it’s a good choice for complex arcs and other shapes. ", "A disadvantage is high thermal conductance. ", "It’s also notorious for condensation and even frost inside the home in very cold weather.", "\n\n• Aluminum with a Thermal Break: In an attempt to reduce thermal conductance, the inside of the frame is separated from the outside and a less conductive material is place between the two sides.", "\n\n• Vinyl: Actually this is polyvinyl chloride, or the familiar PVC used in residential plumbing applications. ", "It provides a good insulating factor and the color is permanent since it goes all the way through; there’s no need to paint.", "\n\n• Insulated Vinyl: This is the same as regular vinyl with an insulated core designed to improve thermal performance.", "\n\n• Fiberglass: This material is also known in the industry as glass fiber reinforced polyester. ", "It’s extruded like aluminum, but offers superior thermal qualities.", "\n\n• Wood Window Frames: Wood is the most traditional of all frames. ", "Since it needs to be painted, changing the home’s color scheme is a snap. ", "It can also be fashioned into complex shapes. ", "Routine maintenance is important since wood is susceptible to rotting." ]

{ "pile_set_name": "Pile-CC" }

[ 0.013793103448275862, 0, 0, 0.05263157894736842, 0, 0, 0, 0, 0, 0.05128205128205128, 0, 0.006289308176100629, 0.005747126436781609, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]

[ "socialengage.io is worth $ 430,560.00\n\n1.68Rating by cutestat.net\n\nIt has a alexa rank of #57,611 in the world. ", "It is a domain having .io extension. ", "It is estimated worth of $ 430,560.00 and have a daily income of around $ 598.00. ", "As no active threats were reported recently, socialengage.io is SAFE to browse." ]

{ "pile_set_name": "Pile-CC" }

[ 0, 0, 0, 0 ]

[ "More 'tiny houses' open, helping Tallahassee's affordable housing crisis\n\nShow Caption Hide Caption Tiny home, sweet home Russell Wallace becomes one of the first new residents of The Dwellings, a tiny home community\n\nThe Dwellings — the nation's only tiny house community to fuse social services and shared amenities on a large campus — is eyeing the final stretch to complete its build out.", "\n\nIt's a welcome option to combat Tallahassee's affordable housing crisis, where the median monthly rent exceeds $950 and nearly half of the population struggles to meet its basic needs.", "\n\nWalking into one of the larger units, Debbie Kolfer and John Robinson see a place they can call home. ", "The two relocated from Tampa to Tallahassee to help Kolfer's daughter, a 19-year-old expecting mother in college.", "\n\n\"Watching the shows on TV, I didn’t believe there was a lot you could do with a tiny house but actually coming in\" and seeing a finished one \"blew my mind,\" Robinson said during a ribbon-cutting event and tour Friday.", "\n\nUnits range from 190-420 square feet. ", "They're stocked with modern finishes and specialty refrigerators that land between the size of a mini and full-size refrigerator.", "\n\nQualified residents pay $600-$800 per month, which include utilities, internet, three meals per day, on-site case management and medical services. ", "Residents have access to a laundry room, a dining hall, a community garden and a fire pit, among other amenities.", "\n\n\"This is an affordable housing community, but it's much more,\" said Rick Kearney, a local tech titan who's helped revolutionize Tallahassee's response to the homeless and affordable housing. \"", "It's really a community for people who have had challenges in their lives.\"", "\n\nKearney's philanthropic efforts created Westgate Community, a gated permanent and transitional housing complex, and the Comprehensive Emergency Services Center, a homeless shelter named for him, which opened in 2015. ", "Both efforts, and now The Dwellings, point to the growing need for housing.", "\n\nFor example, he said the city received 15,000 applications for housing vouchers during its last application cycle. ", "Kearney said that speaks to a \"huge need.\"", "\n\n“They come to a place like this, they don’t feel like someone’s unwelcomed guest,” Kearney said. \"", "They can now think about what (they) can do to be more productive ... and help others.”", "\n\nOne of the major challenges is the volume of households that are considered 'cost burdened,' according to a recent city analysis. ", "The term refers to households where more than 30% of income is spent on rent and utilities.", "\n\nTo be 'severely cost burdened' means 50% or more of a person's income is spent on housing. ", "Out of Tallahassee's 75,195 households, 54% are considered cost burdened, more than a third of whom are college students.", "\n\nIn recent years, the city of Tallahassee and Leon County have devoted more resources and attention toward affordable housing, including a need for new initiatives.", "\n\nSome of those efforts include establishing a community land trust, adding inclusionary housing within the massive Canopy development at Welaunee, and expanding the down payment assistance program with the Tallahassee Lenders' Consortium.", "\n\nKearney said there are several dozen people on a waiting list for The Dwellings, which began construction in late 2016 and opened its first units a year later.", "\n\nHe went on to talk about the myriad of volunteer opportunities, from the kitchen to the market to the woodshop for the residents.", "\n\n\"There's nothing like it,\" he said.", "\n\nContact TaMaryn Waters at tlwaters@tallahassee.com or follow @TaMarynWaters on Twitter." ]

{ "pile_set_name": "OpenWebText2" }

[ 0.00510204081632653, 0, 0.019230769230769232, 0.008849557522123894, 0.0045662100456621, 0, 0, 0, 0, 0.005154639175257732, 0, 0.0091324200913242, 0.013333333333333334, 0, 0.023809523809523808, 0.01, 0, 0, 0, 0, 0, 0, 0.008368200836820083, 0.006211180124223602, 0, 0, 0.02247191011235955 ]

[ "@mixin foo { @content }\na { b: c; @include foo {} }\n" ]

{ "pile_set_name": "Github" }

[ 0.057692307692307696 ]

[ "Guilty by insolubility--does a protein's detergent insolubility reflect a caveolar location?", "\nSeveral recent papers have described a simple approach to isolate caveolae and proteins involved in apical sorting in epithelial cells that is based on their detergent insolubility. ", "These publications have excited such diverse fields of cell biology as intracellular protein transport, signal transduction and, in particular, research into caveolae. ", "In this article, Kurzchalia, Hartmann and Dupree argue that a more critical evaluation of this detergent insolubility is needed before a subcellular location or function can be ascribed to a protein." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0, 0.005025125628140704 ]

[ "871 F.2d 118\nU.S.v.", "Vela*\nNO. ", "88-2858\nUnited States Court of Appeals,Fifth Circuit.", "\nMAR 15, 1989\n\n1\nAppeal From: S.D.Tex.", "\n\n\n2\nAFFIRMED.", "\n\n\n\n*\n Fed.", "R.App.", "P. 34(a); 5th Cir.", "R. 34.2\n\n\n" ]

{ "pile_set_name": "FreeLaw" }

[ 0, 0, 0.03773584905660377, 0.02564102564102564, 0, 0, 0.16666666666666666, 0, 0.1 ]

[ "Health service interventions targeting relatives of heart patients: a review of the literature.", "\nRelatives of heart patients experience anxiety, uncertainty, and low quality of life, and the hospitalization of a heart patient is associated with increased risk of death for the partner. ", "Relatives' physical and mental problems may be rectified by activities established by the health services to support relatives and to enable them to handle the situation. ", "This study systematically reviewed the literature on the latest health service interventions targeting relatives of heart patients. ", "The literature about interventions targeting relatives of heart patients has been systematically reviewed to clarify what the health services do for relatives of heart patients and to assess the effects of interventions. ", "We searched Medline, EMBASE, PsycINFO, CINAHL database, CSA and the Cochrane Library from January 2000 to March 2006. ", "Only six scientific articles reported on interventions testing health service activities for relatives of heart patients, and one literature review scrutinized earlier studies within the field. ", "All the interventions indicate positive effects on patients' and/or relatives' health and well-being, in accordance with nurses' assessments. ", "Nevertheless, the few studies and their questionable quality means that full evidence on the effects of the interventions is lacking. ", "Future research should clarify the evidence for health service interventions and draw up guidelines for activities for the relatives of heart patients." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0, 0, 0, 0.025423728813559324, 0, 0, 0, 0 ]

[ "Image caption Could MRI scans be used to predict the chance of developing Alzheimer's\n\nBrain scans may be able to indicate potential Alzheimer's patients years before symptoms appear, according to the results of a small study.", "\n\nResearch published in Neurology showed parts of some patients' brains had shrunk up to a decade before signs of Alzheimer's would otherwise be evident.", "\n\nAlzheimer's Research UK said there was stong evidence that the disease began to develop in mid-life.", "\n\nEarly diagnosis may one day prove vital in enabling effective treatment.", "\n\nSixty-five patients, all of whom started with normal brain function, took part in the study at Massachusetts General Hospital and the Rush University Medical Center in the US.", "\n\nMRI scans were taken and the thickness of regions of the brain were measured.", "\n\nBrain Thinning\n\nTwenty per cent of patients with brains of average thickness went on to develop Alzheimer's.", "\n\nHowever, no patients with thicker brains developed the disease, while 55% of those with thinner ones also contracted it.", "\n\nRegions of the brain showing thinning in Alzheimer's disease Medial temporal lobe\n\nTemporal pole\n\nSuperior frontal gyrus\n\nDr Brad Dickerson, lead author and neuroscientist at the Massachusetts General Hospital, said: \"We used what we know about the signature brain changes seen in patients with Alzheimer's dementia, measured those areas in individuals with no symptoms and eventually determined that those who ultimately developed dementia showed subtle shrinking long before they had any symptoms.\"", "\n\nEarly Diagnosis\n\nIn the UK, one in 14 people over the age of 65 has Alzheimer's disease.", "\n\nResearch suggests that it takes several years before the disease produces noticeable symptoms.", "\n\nIt means any form of treatment might need to start in middle age, so early diagnosis would be vital.", "\n\nRebecca Wood, chief executive of Alzheimer's Research UK, said: \"This adds to growing evidence that the changes associated with Alzheimer's disease begin long before symptoms start to appear.", "\n\n\"However, this is a small study and it needs to be expanded before we can be certain of the method's accuracy in predicting Alzheimer's.", "\n\nThe authors acknowledge that further work involving many more patients is needed." ]

{ "pile_set_name": "OpenWebText2" }

[ 0, 0, 0, 0, 0.011299435028248588, 0, 0, 0, 0.00398406374501992, 0, 0, 0, 0.0051813471502590676, 0, 0 ]

[ "Emma Tillman\n\nEmmaline Fanchon Tillman, ( Faust; November 22, 1892 – January 28, 2007) was an American supercentenarian and, at age 114 years 67 days, the oldest validated living person from the death of 115-year-old Puerto Rican man Emiliano Mercado del Toro on January 24, 2007 until her own death four days later.", "\n\nBiography\n\nTillman was one of 23 children born to former slaves Alphonso Faust and Martha Gibson Faust in Gibsonville, North Carolina. ", "Her maiden name, Faust, had been adopted from the plantation owner who owned her father's family before the Civil War, Cane Faust. ", "The family moved to Glastonbury, Connecticut in 1900, where Tillman became the only African-American attending Glastonbury High School, graduating in 1909 as the first African-American to do so there. ", "Although she enjoyed studying commercial arithmetic, and did the accounts on her father's tobacco farm, she moved to Hartford in 1914, to take up a position as a housekeeper, the only paid employment open to her at that time. ", "She married Arthur Tillman the same year, and they had two daughters. ", "Tillman ran her own baking and catering service for about sixty years, sometimes serving meals for visiting state dignitaries such as Governors Raymond E. Baldwin and Ella T. Grasso, and whose regular customers included Dr. Thomas Hepburn, a noted Hartford Hospital urologist and father to actress Katharine Hepburn, who she served as the family cook for a number of years. ", "Her husband died in 1939. ", "Four of her siblings lived past age 100, including a brother who lived to be 108, a sister who reached 105 and two others who reached 102.", "\n\nThroughout her lifetime, Tillman was involved in various NAACP social programs and the National Council of Negro Women.", "\n\nThe day before her 110th birthday, former Connecticut Governor John G. Rowland proclaimed that her birthday, November 22, would be known within the state as \"Emma Tillman Day\".", "\n\nTillman was a parishioner at the Metropolitan A.M.E. Zion Church for more than 80 years, where she became informally known as the \"mother\" of the church and the A.M.E. Conference as a whole. ", "She was a member of the church choir for more than seventy years, and president of the senior for fifteen years.", "\n\nShe lived independently until the age of 110. ", "On January 18, 2007, she became the oldest living woman following the death of 115-year-old Canadian Julie Winnefred Bertrand, and on January 24, 2007 she became the world's oldest living person with the death of 115-year-old Emiliano Mercado del Toro (a native of Puerto Rico).", "\n\nShe died in an East Hartford nursing home on January 28, 2007, aged 114 years, 67 days. ", "She holds the record for the shortest period spent as the world's oldest person. ", "After her death, Yone Minagawa of Japan became the world's oldest person.", "\n\nOn March 9, 2007, Tillman was discussed as a major subject of a lecture by Felicia Nimue Ackerman, a professor of philosophy at Brown University, titled \"Nature vs. the Tragedy of Emma Faust Tillman's Death\", at the Karbank Symposium in Environmental Philosophy at Boston University. ", "The lecture discussed issues related to environmental philosophy, particularly the value of individual human lives compared to the value of natural environments and their preservation.", "\n\nSee also\n\n Ageing\n Maximum life span\n Senescence\n 100 oldest American people ever\n\nReferences\n\nExternal links\n Oldest people list \n Salvation Army website \n \n\nCategory:1892 births\nCategory:2007 deaths\nCategory:American people of German descent\nCategory:American supercentenarians\nCategory:People from Hartford, Connecticut\nCategory:People from Gibsonville, North Carolina\nCategory:African-American centenarians\nCategory:Female supercentenarians" ]

{ "pile_set_name": "Wikipedia (en)" }

[ 0.006329113924050633, 0.014598540145985401, 0.007633587786259542, 0.014925373134328358, 0, 0.014285714285714285, 0.0106951871657754, 0, 0, 0.024793388429752067, 0.0056179775280898875, 0.010362694300518135, 0, 0, 0.0035971223021582736, 0, 0, 0.0136986301369863, 0.013986013986013986, 0, 0.002242152466367713 ]

[ "Two from the Vault\n\nTwo from the Vault is a three-CD live album by the rock band the Grateful Dead. ", "It was recorded at the Shrine Auditorium in Los Angeles, California on August 24, 1968. ", "The event was left unreleased for nearly 25 years, before being mixed down from the original multi-track reels and released on Grateful Dead Records in 1992.", "\n\nRecording\n\nThe concert was recorded on a then-state-of-the-art, one-inch 8-track tape machine that was supplied by the band's record label, Warner Bros. The record company also insisted on supplying engineers who turned out to be unfamiliar with the close miking technique involved in recording rock music. ", "Consequently, each of the eight tracks contained significant leakage from all of the other instruments in the band, resulting in severe phase cancellation problems.", "\n\nAlmost twenty-four years later, Don Pearson and producer Dan Healy solved this problem by employing a B&K 2032 Fast Fourier transform (FFT) digital spectrum analyzer to measure the delay in time between the different microphones, using the track of bassist Phil Lesh as the time centerpiece. ", " The delay times were fed into a TC1280 stereo digital delay, which, along with careful mixing, resulted in a nearly perfect stereo image.", "\n\nReleases\nAn expanded edition of the album, with a third CD, was released in 2007 featuring the three songs (from August 23, 1968) previously released as bonus tracks on the 2003 reissue of Anthem of the Sun .", "\n\nTwo from the Vault was released by Light in the Attic Records as a four-disc vinyl LP on December 9, 2014.", "\n\nTrack listing\n\nDisc one\n\"Good Morning, Little Schoolgirl\" (Sonny Boy Williamson) – 15:59\n\"Dark Star\" > (Jerry Garcia, Phil Lesh, Ron McKernan, Bob Weir, Mickey Hart, Bill Kreutzmann, Robert Hunter) – 11:20\n\"St. Stephen\" > (Garcia, Lesh, Hunter) – 4:40\n\"The Eleven\" > (Lesh, Hunter) – 14:27\n\"Death Don't Have No Mercy\" (Reverend Gary Davis) – 8:23\n\nDisc two\n\"That's It for the Other One\" > (Garcia, Kreutzmann, Weir) – 15:40\n\"New Potato Caboose\" > (Lesh, Robert Petersen) – 14:16\n\"Turn On Your Lovelight\" (Joseph Scott, Deadric Malone) – 17:13\n\"(Walk Me Out in the) Morning Dew\" (Bonnie Dobson, Tim Rose) – 7:13\nNote: \"Morning Dew\" ends abruptly as the power is turned off due to curfew \n\nDisc three (2007 expanded edition)\n\"Alligator\" > (McKernan, Lesh, Hunter) – 18:43\n\"Caution (Do Not Stop on Tracks)\" > (McKernan, Grateful Dead) – 11:30\n\"Feedback\" (Grateful Dead) – 4:01\nNote: Tracks on this disc, recorded the previous night, also appear as bonus tracks on Anthem of the Sun\n\nPersonnel\n\nGrateful Dead\nJerry Garcia – guitar, vocals\nPhil Lesh – bass, vocals\nRon \"Pigpen\" McKernan – keyboards, harmonica, vocals\nBob Weir – guitar, vocals\nBill Kreutzmann – percussion\nMickey Hart – percussion\n\nProduction\nDan Healy – producer\nJeffrey Norman – engineer\nDon Pearson – engineer\nJoe Gastwirt – mastering\nDick Latvala – tape archivist\nTimothy Harris – cover art\n\nCharts\nAlbum - Billboard\n\nReferences\n\nExternal links\nTwo from the Vault liner notes and technical information at frognet.net\n\nCategory:Grateful Dead live albums\nCategory:1992 live albums\nCategory:Albums produced by Dan Healy (soundman)\nCategory:Grateful Dead Records live albums" ]

{ "pile_set_name": "Wikipedia (en)" }

[ 0.03, 0, 0, 0.0032258064516129032, 0, 0.013605442176870748, 0, 0, 0.009259259259259259, 0.023184868822452714 ]

[ "If anyone can run past Baylor to the national title, it might be Notre Dame and guard Skylar Diggins. / ", "David Butler II, USA TODAY Sports\n\nby Kelly Whiteside, USA TODAY Sports\n\nby Kelly Whiteside, USA TODAY Sports\n\nCan anybody beat Baylor? ", "Expect that to be the dominant theme throughout the NCAA women's basketball tournament, which begins Saturday.", "\n\nAs expected defending national champion Baylor, Notre Dame, Connecticut and Stanford earned No. ", "1 seeds when the 64-team field was announced Monday. ", "The same four teams also earned top seeds last year; it's the first time that has ever happened in consecutive seasons.", "\n\n\"Nothing's different this year,\" Baylor star Brittney Griner said. ", "The game's most imposing player was referring to the team's goal to win another national title, but she could have been talking about the top-seeded teams in the tournament.", "\n\nThe top four teams also advanced to the Final Four last year with Baylor knocking off Notre Dame in the title game.", "\n\nSeven of the top eight seeded teams are the same from 2012; No. ", "2 seed California is the only newcomer.", "\n\n\"We look at previous matchups during the regular season and first and second round matchups in previous years,\" said Carolayne Henry, chair of the Division I women's basketball committee, adding that it's difficult to project further into the bracket.", "\n\nWhile unpredictability may be the hallmark of the men's tournament, sustained dominance has been the theme in the women's game. ", "UConn and Stanford have competed in the past five Final Fours.", "\n\nStanford (31-2) handed Baylor (32-1) its only loss this season, ending the Bears 42-game winning streak in November. ", "The two could meet again in the national semifinal on April 7 in New Orleans.", "\n\nBaylor is hoping to become the fourth different school to win consecutive NCAA championships, joining Southern California, UConn and Tennessee.", "\n\nThe Bears could be challenged along the way by the No.2-seeded Lady Vols, who have made every NCAA tournament since it began in 1982. ", "Though the Vols have been mainstays in the tournament, this will be the first time coach Pat Summitt won't be on the sidelines after retiring last season due to early onset dementia.", "\n\nStanford hopes to finally give its Hall of Fame coach Tara VanDerveer a national championship for the first time in 21 years. ", "However, the Cardinal don't have an easy road since they could face No. ", "2 seed California in the regional final. ", "The Golden Bears ending the Cardinal's 81-game conference winning streak in January.", "\n\nUConn (29-4) has lost to Notre Dame (31-1 )three times this season, including in the Big East championship game last week, and the two rivals could meet again on the other side of the bracket in the national semifinal. ", "If so, it would be the third consecutive meeting in the national semifinal.", "\n\nUConn won't have to travel very far before then. ", "The Huskies play their first two rounds at home and if they advance, they would play the next two rounds in Bridgeport, Conn. The Huskies are hoping for their eighth national title, which would tie Tennessee's record.", "\n\nUConn guard Kelly Faris said the Huskies are looking to redeem themselves after falling short of expectations. ", "The Huskies didn't win the Big East conference regular season or tournament title for the first time in 19 years.", "\n\n\"This is our last chance to respond in the right way, respond how (coach Geno Auriemma) teaches us to, and the way he challenges us to,\" Faris said. \"", "This is our last chance to step up and make something of it.\"", "\n\nThe most intriguing matchup for UConn could be a possible game against Delaware in the regional finals. ", "Senior Elena Delle Donne, the nation's second-leading scorer, had signed with UConn but transferred to Delaware before her freshman season to be closer to her family.", "\n\nNotre Dame is the only No. ", "1 seed not hosting the early rounds. ", "The Irish will have to travel to Iowa City for the first two rounds and then Norfolk if they advance. ", "It was expected the Irish might end up in Columbus, Ohio, a neutral-site floor. (", "Ohio State did not make the tournament.) ", "The Irish could face host Iowa in the second round.", "\n\nOne factor was location, Henry said. \"", "We looked at Notre Dame in Columbus and when we looked at the mileage, Notre Dame is slightly closer to Iowa than it is to Columbus,\" Henry said.", "\n\nAfter losing in the title game the previous two seasons, the Irish hope third time is a charm in Skylar Diggins' senior season. ", "Their conference, the soon-to â??be-torn apart Big East earned eight bids, more than any other conference." ]

{ "pile_set_name": "Pile-CC" }

[ 0.019230769230769232, 0.04411764705882353, 0.00909090909090909, 0.030612244897959183, 0, 0, 0.014492753623188406, 0, 0.008547008547008548, 0, 0, 0.007905138339920948, 0, 0.016129032258064516, 0.008403361344537815, 0, 0.006896551724137931, 0.014705882352941176, 0.01098901098901099, 0.015625, 0, 0, 0.011904761904761904, 0.004524886877828055, 0, 0, 0, 0.008849557522123894, 0, 0.013157894736842105, 0, 0, 0.006024096385542169, 0.034482758620689655, 0, 0.00980392156862745, 0, 0.024390243902439025, 0, 0.025, 0.013793103448275862, 0, 0 ]

[ "Localized hypersensitivity and late coronary thrombosis secondary to a sirolimus-eluting stent: should we be cautious?", "\nThe US Food and Drug Administration recently issued a warning of subacute thrombosis and hypersensitivity reactions to sirolimus-eluting stents (Cypher). ", "The cause and incidence of these events have not been determined. ", "We present findings of a 58-year-old man who died of late stent thrombosis 18 months after receiving 2 Cypher stents for unstable angina. ", "Although angiographic and intravascular ultrasound results at 8 months demonstrated the absence of neointimal formation, vessel enlargement was present. ", "An autopsy showed aneurysmal dilation of the stented arterial segments with a severe localized hypersensitivity reaction consisting predominantly of T lymphocytes and eosinophils. ", "The known pharmacokinetic elution profile of Cypher stents and the presence of polymer fragments surrounded by giant cells and eosinophils suggest that a reaction to the polymer may have caused late stent thrombosis. ", "Careful long-term follow-up of patients with vessel enlargement after Cypher stent placement is recommended." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0.012903225806451613, 0, 0.007246376811594203, 0, 0, 0.004608294930875576, 0.009259259259259259 ]

[ "Paris, le 20 mars 2015 — La levée de boucliers unanime des acteurs de l’Internet, des associations de défense des libertés et des professionnels de la Justice contre le projet de loi sur le renseignement présenté en Conseil des ministres le jeudi 19 mars illustre de façon éclatante les dangers multiples de ce projet et la nécessité absolue de changer d’orientation, quand bien même le gouvernement tente de rattraper sa communication en proposant, a posteriori, un dialogue avec la société civile.", "\n\n\n\nDepuis les premières annonces mardi 17 mars et jusqu’à ce jour, les oppositions au projet de loi relatif au Renseignement se font de plus en plus directes, massives et claires. ", "Les associations de défense des libertés publient les unes après les autres leur opposition au projet de loi, tandis que les avocats et magistrats font part de leurs inquiétudes et que les syndicats professionnels du numérique s’inquiètent de devenir des « auxiliaires de renseignement » en collectant les données de leurs clients.", "\n\nLa Quadrature du Net, comme d’autres associations et institutions ayant ouvertement exprimé leurs désaccords et inquiétudes sur le projet de loi, a été invitée en urgence (jeudi soir) à être présente cet après-midi vendredi 20 mars à Matignon.", "\n\nIl s’agit d’être reçus par des conseillers de Manuel Valls pour nous « présenter les enjeux » du projet de loi et « calmer nos inquiétudes ». ", "Comme auprès des acteurs économiques de l’Internet reçus hier par Axelle Lemaire, le gouvernement tente de désamorcer la pluie de critiques qui s’expriment sur le projet de loi sur le Renseignement.", "\n\nNous n’irons pas à cette réunion. ", "Les concertations après coup sont un simulacre de prise en compte de la société civile, et ne servent qu’à entretenir l’illusion du dialogue. ", "Si le gouvernement se souciait réellement de notre avis, il avait tout le loisir de nous le demander avant, en amont et pendant la rédaction du projet de loi.", "\n\nAujourd’hui, le projet de loi ayant été présenté en Conseil des ministres, la balle est dans le camp des parlementaires : à eux de prendre la mesure de leurs responsabilités face à un texte dangereux et gravement attentatoire aux libertés individuelles et publiques. ", "Nous les exhortons à amender en profondeur ce texte et à se faire les défenseurs des citoyens qu’ils représentent. ", "Nous sommes prêts à discuter des nombreuses modifications qu’il est nécessaire d’apporter à ce projet de loi, à défaut d’avoir été écoutés en amont lorsque nous avons, de multiples fois, mis en garde contre les tentations de surveillance massive.", "\n\n« Ce projet de loi suscite à juste titre une opposition large et des inquiétudes multiples. ", "Le champ de la discussion doit aujourd’hui être public, entre citoyens, parlementaires, gouvernement, presse et société civile. ", "Nous ne servirons pas de faire-valoir associatif à la communication gouvernementale » déclare Adrienne Charmet, coordinatrice des campagnes de La Quadrature du Net" ]

{ "pile_set_name": "OpenWebText2" }

[ 0.008016032064128256, 0, 0.0030211480362537764, 0, 0.013888888888888888, 0.010101010101010102, 0, 0.014084507042253521, 0.012658227848101266, 0.011152416356877323, 0, 0, 0, 0, 0.012269938650306749 ]

[ "The WBS Timeline\n\nThe concept of the WBS timeline centres around the most significant milestones that happened during the School’s 50 year history. ", "We then added some interesting local and global facts alongside those dates in order to give historical context to those special WBS moments.", "\n\n1968\n\nWBS\n\nWits Business School opens its doors. ", "The first programme is a partnership with Stanford University: Executive Development Programme for Senior Managers." ]

{ "pile_set_name": "Pile-CC" }

[ 0, 0, 0, 0.008771929824561403 ]

[ "\nNIST Simulates New DNA Sequencing via Base Attractions through Graphene Nanopore - dpflan\nhttp://www.nist.gov/mml/acmd/nist-simulates-fast-accurate-dna-sequencing-through-graphene-nanopore.cfm\n======\ndpflan\nPrevious approaches relied upon detecting electrical changes in the graphene\nnanopore as the strand passed through; this new approach cleverly uses\ngraphene nanoribbons with copies of nucleotide bases attached which chemically\nmicro-bond to their respective opposite base (A<->T, G<-C) as the strand is\npassed through. ", "This significantly improves accuracy of the detection method.", "\n\n" ]

{ "pile_set_name": "HackerNews" }

[ 0.003795066413662239, 0, 0 ]

[ "“Before a standing army can rule, the people must be disarmed, as they are in almost every country in Europe. ", "The supreme power in America cannot enforce unjust laws by the sword; because the whole body of the people are armed, and constitute a force superior to any band of regular troops.”", "\n\nNoah Webster, An Examination of the Leading Principles of the Federal Constitution , October 10, 1787\n\n“Before a standing army can rule, the people must be disarmed, as they are in almost every country in Europe. ", "The supreme power in America cannot enforce unjust laws by the sword; because the whole body of the people are armed, and constitute a force superior to any band of regular troops.”", "\n\nNoah Webster, An Examination of the Leading Principles of the Federal Constitution , October 10, 1787\n\nSneakySFDude:\n\nWhat country can preserve its liberties if their rulers are not warned from time to time that their people preserve the spirit of resistance. ", "Let them take arms.”", "\n\ntheduke:\n\nHow about the nuts on the crazy side of the 2a that say you need to be armed to the hilt because the gobermit is coming after you.", "\n\nThe shooter was a Bernie Sanders campaign volunteer. ", "He callled for Democratic Socialism - just like Ocasio-Cortez. ", "He was deeply invested in liberal politics and hated Donald Trump and frequently wrote to Democrat senators to cajole and advise them. ", "All his friends say he was a progressive liberal." ]

{ "pile_set_name": "Pile-CC" }

[ 0, 0, 0.004651162790697674, 0, 0.003816793893129771, 0, 0, 0, 0.015873015873015872, 0.007407407407407408, 0 ]

[ "Bed of nails mattress - cot mattress babies r us\n\nhow deep is how long should a pillow top mattress last bed of nails mattress\n\nThe centerpiece of the article is the story of Paul Nungesser, a Columbia University student who accused of rape by Sulkowicz.", "\nFor over 120 years, Shifman Mattress Company has been committed to providing superior craftsmanship, lasting comfort and value.", "\nZinus Memory Foam 12-Inch Green Tea Mattress combines all these aspects to make best mattress store san diego one firm mattress that feels right for your back and lasts you a long time.", "\nIf you have a spouse, finding a mattress that meets both your needs is imperative.", "\nbed of nails mattress When you walk into a local mattress showroom , you won't feel overwhelmed by the expansive selection.", "\nThey have 12 inches of brick hard foam around the circumference leaving only a small area of adjustable mattress.", "\nReuse of crib mattress must be considered carefully because older mattresses might contain toxic substances, such as flame retardants, that have been banned.", "\nSo it's not surprising that more and more Americans are opting for a firmer sleeping surface to carry them comfortably through the night.", "\n\nFor those who want to spend less than Casper / Leesa prices on a mattress, but still want to circumvent the sketchy mattress store's cheapest offerings, Tuft and Needle may be a good alternative.", "\nThe two-inch thick topper graco bassinet mattress oval is a preferred choice for low to medium pressure relief and is best suited for lighter-weight occupants.", "\nThe Visco Gel layer is placed on top coupon code for mattress giant of an advanced ventilated TFK pocket spring core, which eliminates completely any partner disturbance and provides balanced reflexive support in any sleep position.", "\nTwin mattresses - Twin size is generally made to accommodate one child or one adult sleeper.", "\nNo one gives exact specifications on their padding, which makes it difficult to make comparisons between different types of foam and between different mattresses and mattress brands.", "\nUsed in a mattress, Outlast material responds to constantly changing skin temperature and manages heat buildup.", "\nIf you choose to go in the direction of memory foam, I would recommend that you use the minimum thickness that is suitable for your needs and preferences.", "\nThe importance of placing this vertical mattress suture is to evert bed of nails mattress the lid margin once it is tied.", "\nSo whether you're looking for a good value, extra space, or just the most luxurious sleep possible, King mattresses are the largest size commonly manufactured and look at this webpage epitome of comfort in the master bedroom.", "\nSold At: Hobby Lobby Stores nationwide from January 2016 through April 2016 for about $7.", "\nBesides the retailers named above, Boyd Flotation mattresses are also sold online on Amazon , Wayfair , The Waterbed Doctor , Waterbed Heaven Bedrooms , and other retail sites.", "\n\nI called Sleepy's immediately to exchange but was told that as per Tepurpedic rules, we were to keep the mattress for one whole month before returning it.", "\nWe have range of Pocket Sprung Mattresses, Memory bed of nails mattress Foam Mattresses, Open Coil Spring Mattresses, Latex Mattresses, Natural Organic Mattresses to choose from.", "\nWe ended up switching to latex pillows and between the mattress and the pillows, it's like laying on a cloud.", "\nOur advice is offered in good faith but without guarantee, as individual coupon code for mattress giant conditions and product use are beyond our control.", "\nThe wool is a natural thermostat, adjusting throughout the night to changes in temperature and moisture.", "\n\nbed of nails mattress celebrity choice bed mattress\n\nIf such is the case, consider hiring a cleaning service for proper mattress cleaning. ", "The articulating support structure 12 has multiple support sections adjustable to various angles including a foot support member 13. ", "The amazing fiber technology of this mattress topper prevents the unnecessary fusion of the fiber and keeps the mattress in its place. ", "Foam pads are more affordable, quicker to setup, and can be used for multiple purposes - like extra support for a frameless backpack or a seat around camp. ", "If the outlet or person you are buying a mattress from is an expert in the materials, construction and layering used in a mattress, and how to fit a mattress to your unique needs and preferences, this can save you many hours of personal research and even more time spent shopping store to store only to hear stories that create more confusion and frustration than they do to educate and inform. ", "You will, however, be able to raise or lower the temperature of your mattress to up to 30 degrees above and 15 degrees below your ambient room temperature. ", "Premium individually wrapped Sealy Stability coils flex independently of one another to provide targeted support, while the SolidEdge support system with Marvelux high density foam extends the usable sleep surface to the mattress edge. ", "Enjoy a rejuvenating sleep with the Beautyrest Recharge Windchase Luxury Firm Low Profile Mattress Set. ", "A memory foam mattress shouldn't be turned, but instead rotated regularly to keep it in ship shape. ", "Back and other pain relief is a reported benefit of Simmons mattresses, whether it's the Beautyrest or Beautysleep line. ", "I bought the Serta Intricate set for $800.00 after much research and 'trying out' of various mattresses. ", "With over 30,000 products to flick through plus an extra 10,000 online exclusives you'll be spoilt for choice with Argos. ", "ALPS is a company known for the bleach memory foam mattress of its products and the rechargeable mountaineering air mattress does not disappoint. ", "Perhaps, some companies which pickup the old mattresses when they deliver the new ones are into recycling, but I'm not sure we have the institutions here to do that. ", "Though there are several name brand memory foam mattresses in the race for your wallet, the 7-zone memory foam mattress is one of the best memory foam mattresses your money can buy. ", "Saatva owners report the most favorable customer service and support experiences by far. ", "Silentnight are kings of the sleep industry and their continuous spring system, called Miracoil, dramatically improves your sleeping experience by minimising roll-together and keeping your back in a comfortable position. ", "The first couple nights were sooo amazing, but lately, the mattress has been way too soft for me. ", "And if fungi have become established in the mattress during previous use by another baby, the generation of toxic gas commences sooner and in greater volume when the mattress is re-used. ", "The price of the Kingsdown mattress is equivalent to the price of high quality brands of mattresses.", "\n\nwhich simmons mattress is the westin heavenly bed reviews\n\nStraight out of the box, the mattress had a bit of a chemical smell and needed some time to off-gas. ", "The people who sell these mis-represented mattresses used to pay five dollars for five reviews, but now, the same people who work the call centers South are coining cash writing reviews for as little as $1. ", "Tempur-Pedic owner experience data shows that this is a potential weakness of the warranty as softening can be just as common and problematic as sagging. ", "Pillow sales - which really took off when they started selling through Amazon Prime, he said futon bunk beds with mattress included supported two years of high-tech mattress development, and after cycling through a handful of suppliers, they were ready. ", "Having a mattress with dual sides extends it's life since you have the ability to flip it on occasion.", "\n\nluxury down mattress topper\n\nWhere we have customers in the door who hate foam, they love our slumberland twin size mattress Bedding latex beds. ", "These contaminants are eliminated from the room when the cover is removed at patient discharge-a much more effective process than traditional terminal cleaning. ", "The Leesa comes with a fashionable and stylish cover that is available in two designs - Classic Leesa and the Airlifting Limited Edition, each adding an extreme level of style and design. ", "I had a bicycle patch on it and it seemed to work, but now mattress seems to have a slow leak. ", "I didn't want to blow the mattress up myself, so I had to purchase a separate pump as well.", "\n\nair with chambers quotes mattress\n\nSoybeans are a renewable resource and using a percentage of American grown natural and renewable oils means less crude oils are required to build a hypoallergenic and firm foam that parents can depend on for their little ones. ", "These sheets are printed with two tonal colors on a silky-soft and shimmering 300-thread-count sateen cotton. ", "body impressions and mattress two outer support core layers 50, 70 have the same pattern, wherein a substantially different ELD exists on the edges of the foam layer than exists in the center of the foam layer. ", "As a homeowner with 2 tenanted rental properties plus a Holiday Cottage rental, Argos is a huge asset when it comes to home furnishings and appliances. ", "Fitted skirt and patented safety stay on corners and extra deep sides stop the pad from slipping off even the thickest mattress. ", "One good tip for how to clean a mattress is to take the mattress to the garage or basement and stand it against the wall.", "\n\nrv short queen mattress vancouver\n\nDon't rush into a purchase or you may regret it. ", "Yes, I understand that not being an expert on mattress components, I certainly can't tell exactly what a mattress will feel like just looking at specs. ", "Following the same terminology of success his successor, Mr. Follow specific instructions in the manual to install these items once the Pack 'n Play has been correctly set up. ", "Plush mattresses are great for individuals looking for a cushioned, cradling sleep. ", "As a result, these details may not be comprehensive and may not apply to all Helix Sleep models. ", "The technology is fully adjustable and hidden inside the mattress, requires no heavy steel structure under The mattress can be supported by any kind of bed frame, no need to change your bedroom furniture. ", "Winner: Amerisleep has better motion isolation, but Essentia is easier to move around on. ", "An improvement on the standard Pocketed Coil system used in the other ranges of Beautyrest mattresses, the Advanced Pocketed Coil system features three wires woven together synthetic latex mattress toxic every single coil to offer exceptional stability and motion separation.", "\n\nmattress wichita falls qb\n\nThe post-partum sweating is uncomfortable, no doubt; with a few extra blanket and dressing in warm clothes, a.c has helped me get through it at times. ", "A foam mattress will have a tiny bit of give, while a spring mattress will be impossible to squash in. ", "The SafeRest comes with a long, 10-year full replacement warranty, and customer service gets raves for their responsiveness to the scattered durability complaints we saw. ", "The cover on the Milliard 6-Inch Memory Foam Tri-Fold Mattress is a soft jacquard bamboo that provides superior comfort. ", "These are non-pillow top mattresses, but utilize more foam than a Firm Mattress. ", "Memory foam detects this and molds itself around the body accordingly by adding or reducing pressure in the injured area. ", "Only by trying out a mattress for at least 15 minutes in each of your favorite sleep positions can you truly know how comfortable a bed feels. ", "After all, you don't want to buy a bed that you love, only to find that you can't sleep on it after a year or two. ", "We have many different organic mattress toppers: a Sueno natural wool mattress topper, a Sueno Natural Rubber mattress topper, a Green Sleep Natural Rubber Topper, a Savvy Rest natural latex mattress topper, and a Bella Sera natural latex mattress topper. ", "Arthritis can be extremely painful, especially first thing in the morning and these heated pads can help to alleviate some of these symptoms, making mornings much more tolerable. ", "It will fit your baby's toddler bed, so there is no need to purchase another mattress. ", "Motion isolation, noise levels and maintenance are low for this type of mattress. ", "The top cover is dry cleanable to keep it looking new while it prevents the harboring of harmful germs and micro organisms. ", "Among the primary reasons can be that you have access to products that can enhance the condition discount mattress denver bronco enhance of your bed or bed mattress. ", "I'm not calling the company and I'm not going back into a store because that took 1.5 years to get it resolved the first time. ", "Feather mattress pads don't wick moisture and are not recommended for babies: They are often too fluffy for a crib. ", "I am selling a single sized bed with the header, frame, box spring, and foam mattress all for 100 dollars.", "\n\nsimmons back care mattress pad\n\nSure, you can spritz it with some fabric spray to help eliminate odors, beaumont pillowtop mattress king that really doesn't do the job of cleaning it. ", "This mattress boasts 13 inches of the best materials: Rayon from Bamboo mattress ticking, a lavish memory foam pillow top, independently encased coils, and 3 zone foam. ", "As part of this new initiative, UL plans to build a new, centrally located U.S. The topper will adjust to each person separately without translating any bouncing, jostling or waves when one person moves or mounts the bed.", "\n\nmattress used by hilton hotels las vegas\n\nOpen-celled 3.0 lb. ", "The Eve's design with memory foam on top of polyurethane foam creates a best super king size bed split mattress both worlds situation, where you get the benefit of great support and pressure relief from memory foam, with added support from the polyurethane foam below. ", "We had put a deposit on a very very expensive mattress at a big box store for half off the price. ", "Additionally, some feel that it measures wrong and that it is not a 52inch mattress.", "\n\nwaterproof mattress pad not plastic\n\nMy mother as a wedding gift had promised us a mattress when we finally decided to settle down. ", "Such mattresses consist of numerous air cells that can be deflated and inflated on a set schedule, which also helps improve airflow and reduces moisture content between the body and the mattress. ", "I make the topsheets into half-fitted sheets as my husband likes to have a thing is; he pulls it up while asleep and we end up with the crumpled sheet all over the bed. ", "The only mattress I've been on that was more comfy was the Heavenly Mattress at a Westin Hotel. ", "Pure Latex Bliss will provide you with years of comfort with the benefit of a 15-year limited warranty. ", "The never flat pump came on when I was crouched right by it setting up the bed and it was really quiet. ", "The last thing you want to do is to throw your back out lifting your baby from the crib. ", "Once we had removed them from their packages I noticed they come with a velvet like, soft, smooth zipper cover to protect your pillow. ", "This makes it very easy to take it out of the crib for those inconvenient sheet changes. ", "The top cotton fabric is a special velour fabric that has 100-Percent cotton fibers for cool comfort and a durable polyester web for superior performance. ", "The new mattress was a gift, so the payment was made by your girlfriend's mother. ", "You can buy waterbed mattresses with different levels of motion, and support or firmness. ", "Users say the Sunbeam Microplush Heated Throw warms up quickly and is thick and insulated well enough to maintain its warmth even after the 3 hour safety switch turns the heating elements off. ", "If you are seeking for an us latex mattress manufacturers mattress topper at good quality, you should consider the Visco Elastic memory foam mattress topper as a choice. ", "Products Offered: Manufacturers of all kinds of mattress covers, cotton sheets, pillow cases etc. ", "We are meticulously selective when it comes to choosing our suppliers because conventional wool suppliers often rely on techniques we are opposed to. ", "The worst reports link mattress chemicals to asthma, skin issues, fertility problems, hair loss, neurological concerns, and cancer. ", "The challenge is still the shipping time, cost of delivery, and being present when the delivery truck arrives. ", "As of now, there have not been many issues reported about this product, yet one issue found with this mattress is the corners of the mattress getting compressed thus some gaps were found between the mattress and the crib within which it is placed.", "\n\nlatex foam queen mattress topper\n\nThey've been made much tougher compared to regular polyester used in air mattress production, so they don't stretch over time. ", "It is assumed that the box can be opened from the inside, as well as the outside. ", "Introduction in 1995 of the Backcare Mattress - this mattress aided the proper alignment of the spine while carpenter perfect harmony mattress topper occupant slept. ", "Even if you don't end up buying the thing there, it is good to lay on it and feel how it is.", "\n\naffordable wool mattress topper\n\nthe only way to ensure that a mattress with polyfoam uses higher quality polyfoam that won't break down very quickly would be by purchasing from a factory direct local outlet or a better sleep shop who will tell you exactly the type and quality of polyfoam in the mattress, how they make it, and the truth about how long you can reasonably expect it to last. ", "Brand names of Sealy, Tempur-Pedic, Simmons, and Serta have been rated highly by consumers for comfort and for customer service. ", "The incontinent-resistant, easy to clean, nylon straps secure the mattress at the head and foot of the bed. ", "Wool affordable mattress cape cod fl - Wool is used in this mattress as a soft and insulating upholstery layer for added comfort." ]

{ "pile_set_name": "Pile-CC" }

[ 0.011811023622047244, 0.0078125, 0, 0, 0, 0, 0.006329113924050633, 0, 0.015228426395939087, 0, 0.004291845493562232, 0, 0, 0, 0, 0, 0.004424778761061947, 0.011111111111111112, 0.01694915254237288, 0.00641025641025641, 0.01675977653631285, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.00847457627118644, 0, 0, 0.008264462809917356, 0.009523809523809525, 0, 0.00684931506849315, 0, 0, 0, 0.004524886877828055, 0, 0, 0, 0, 0, 0.006493506493506494, 0.003937007874015748, 0, 0, 0, 0.005319148936170213, 0, 0, 0, 0, 0.004739336492890996, 0.006578947368421052, 0, 0, 0, 0.006578947368421052, 0.005681818181818182, 0, 0, 0, 0.011111111111111112, 0.0036363636363636364, 0.005555555555555556, 0, 0.005847953216374269, 0, 0, 0, 0, 0, 0.01171875, 0, 0, 0, 0, 0, 0, 0, 0, 0.005376344086021506, 0.005917159763313609, 0.004524886877828055, 0, 0, 0, 0, 0, 0, 0, 0.020833333333333332, 0.009615384615384616, 0, 0, 0, 0, 0, 0, 0, 0.0051813471502590676, 0.011764705882352941, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.015503875968992248, 0, 0 ]

[ "Pages\n\nTuesday, July 15, 2008\n\nNo Gays Here\n\nCarole van der Berg and Helen Ross were happily planning their wedding when they were informed that the hall that they had chosen had decided to cancel their contract. ", "It seems the owners are Jehovah Witness and are strictly against Gay marriage.", "\n\nAccording toDispatchOnline, Hormann the owner said, \"I like them very much and I have nothing against them but we do not condone same- sex marriages. ", "They are making a big thing out of nothing\"\n\nA big deal out of failing to be overjoyed at being discriminated against...A big deal for not accepting bigotry. ", "I am sick of those who are intolerant claiming that the targets of their discrimination are \"too sensitive\" when we react in anger. ", "To not be validated as a human being to be treated as less thanis a big deal.", "\n\nLet's be honest how many times have we seen Jehovah Witnesses turn to the courts to assert their right to religious freedom. ", "We as a society are expected to tolerate their point of view and yet they cannot find it within themselves to be tolerant of others...in a wordhypocrisy. ", "If you want justice and the tolerance of others you need to extend that same courtesy. ", "If you claim to be asserting a religious position it should be based on more than a verse in Leviticus. ", "How about remembering things like let he is who is without sin cast the first stone, or all are sinners in the eyes of God? ", "You have no right to sit in judgement of others based in religion. ", "It is a disgrace and completely counter to the theology you claim to represent. ", "Actions like this are exactly what gives Christianity a bad name.", "\n\n@Johnie roll Womanist musings is a safe place and your commentary is not appreciated. ", "So do us all a favor and go fuck yourself According to your bible God made people in her own image so I am pretty sure that he would not approve of you attitude towards her creation. ", "So you ignorant fuck spread hate elsewhere you are not welcome here\n\nI know Renee already rightly told you to go fuck yourself, and that you're not welcome here. ", "To that I have nothing to add. ", "Spewing hatred may help you feel self-righteous, but it shouldn't be tolerated in polite society.", "\n\nIn the unlikely event that, though unwelcome, you still lurk here, hoping some spiteful comment may escape moderation, please allow me to indulge you with a question:\n\nWhat's your shirt made out of?", "\n\nGo ahead, check the tag on the back of the shirt you're wearing right now, and see what it's made of. ", "Then go through your closet or your dresser, and find all of your other shirts, and see what they're made of.", "\n\nWhy, you ask?", "\n\nBecause the Holiness Code you appeal to in your homophobic screed ascribes the same moral value to wearing mixed fibers as it does to men having sexual relations with other men. (", "Women aren't even considered in the text, as the ancient Israelites either couldn't conceive of sex without a penis, or simply weren't concerned about it.)", "\n\nIf you're going to use an ancient set of laws to condemn people, if you're going to toss around the dreaded word \"abomination,\" you should follow those laws yourself, to the letter. ", "Each and every single one of them.", "\n\nWhich means you'd better, among other things, find out what your shirt is made of. ", "And if it is a cotton/polyester blend, you'd better get it off fast, because wearing that shit could send you straight to hell.", "\n\nIntense Debate Comments\n\nAbout Me\n\nI am the mother of two darling little boys that fill my life with hope. ", "They have inspired me to help raise awareness of the issues that plague this little blue planet.", "\nIf you are looking for a blog that is all about how wonderful and rosy this world is, turn right because this space will not be for you. ", "I am a committed humanist. ", "I believe in the value of people over commodities. ", "I believe in the human right to food, clothing, shelter, and education. ", "I am pacifist, anti-racist, WOC. ", "My truth may not be your truth, but I intend to speak it nonetheless.", "\nPlease feel free to direct questions or commentary womanistmusings@gmail.com" ]

{ "pile_set_name": "Pile-CC" }

[ 0.014084507042253521, 0.01282051282051282, 0.006578947368421052, 0, 0, 0, 0.007874015748031496, 0, 0, 0, 0, 0, 0, 0, 0.011363636363636364, 0, 0.006172839506172839, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.030303030303030304, 0, 0.012987012987012988 ]

[ "Lifestyle, harassment at work and self-assessed health of female flight attendants, nurses and teachers.", "\nHealth-related lifestyle, harassment at work, and self-assessed health of female flight attendants in comparison to that of female nurses and female primary school teachers were surveyed. ", "A higher proportion of flight attendants than nurses or teachers were smokers, 26% vs. 15% and 17% respectively; and consumed alcohol at least once a week, 40% vs. 21% and 16%. ", "Repeated sexual harassment at work was more common among the flight attendants, 31% vs. 8% and 4%; whereas bullying, physical violence and threats were less prevalent among the flight attendants (12%) than among nurses (19%). ", "Flight attendants were on average somewhat taller, but weighed on average less, 63.8 kg vs. 72.4 kg and 72.7 kg respectively. ", "Repeated exposure to sexual harassment, bullying, violence and threats was related to less physical and psychological well-being in all the groups. ", "Teachers scored on average significantly lower than did the flight attendants on general health and physical well-being, while nurses did not." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0, 0, 0, 0, 0 ]

[ "Q:\n\nTrack amount of clicks user clicks button\n\nI was wandering if there was any way to track the amount of clicks a user clicks a button. ", "\nI am creating a game and the user will only be allowed to take 5 turns. ", "After these turns the user has lost the game.", "\nI need to create maybe an if statement where the amount of clicks the user takes reaches > 5 then the user has lost. ", "Is this possible. ", "\nI appreciate any help on this. ", "Thanks\nEdit: \nButton link2Btn = (Button)findViewById( R.id.answerSelected );\n\nlink2Btn.setOnClickListener( new View.", "OnClickListener()\n{\npublic void onClick(View v)\n{\n i++;\ngetAnswer();\n }\nThe get Answer method works fine except the if i > 5 statement within get Answer which is:\nelse if(i>5){\n correctAnswer.setText(\"You have lost\");\n\nA:\n\nUse a flag variable and make an increment over a button press.", "\nlike\nint i=0;\n\nwhen button pressed\ni++;\n\nNow your condition\nif(i>5){}\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0, 0, 0, 0, 0, 0, 0.008620689655172414, 0.003389830508474576, 0.013888888888888888 ]

[ "Ripple is a real-time global settlement network that offers instant, certain and low-cost international payments. ", "Launched in 2012, ripple “enables banks to settle cross-border payments in real time, with end-to-end transparency, and at lower costs.” ", "Ripple’s consensus ledger (its method of conformation) is unique in that it doesn’t require mining. ", "In this way, ripple sets itself apart from bitcoin and many other altcoins. ", "Since Ripple’s structure doesn't require mining, it reduces the usage of computing power, and minimizes network latency. ", "Ripple believes that “distributing value is a powerful way to incentivize certain behaviors” and thus currently plans to distribute XRP primarily “through business development deals, incentives to liquidity providers who offer tighter spreads for payments, and selling XRP to institutional buyers interested in investing in XRP.” ", "So far, ripple has seen success with this model; it remains one of the most enticing digital currencies among traditional financial institutions looking for ways to revolutionize cross-border payments.", "\nNextcoinmarket Has Influenced Products. ", "Nextcoinmarket Associated With Foreign Exchange, Common Assets And Other Underlying Variables, Involves A High Level Of Risk And The Possibility Of Loss Of Some Or All Of Your Investment.please Consider Carefully Whether Trading Or Investing In Bitcoin Is Appropriate To Your Financial Situation. ", "Only Risk Capital Should Be Used When Trading Or Investing In Bitcoin. ", "You Must Review The Terms Of Service And This Risk Disclosure Prior To Setting Up An Account. ", "Website Is Owned And Operated By Operational Address: 28 Dzavaharlal Neru Str., ", "Entrance Silver Business Center, 1324, Sofia, Bulgaria\nEthereum: an open-source cryptocurrency launched in 2015 and proposed by Vitalik Buterin, Garvin Wood, and Joseph Lubin. ", "It is based on the blockchain model whose main objective is the decentralization of the market. ", "In 2016 there was a fork from which the Classic Ethereum emerged. ", "Currently and since its inception, Ethereum ranks second in tems of the market capitalization.", "\n\nELcurrency are influenced products. ", "ELcurrency associated with foreign exchange, common assets and other underlying variables, involves a high level of risk and a possibility of loss of some or all of your investment.", "Please consider carefully whether trading or investing in bitcoin is appropriate to your financial situation. ", "Only risk capital should be used when trading or investing in bitcoin. ", "You must review Terms of Service and this Risk Disclosure prior to establishing an account.", "\n\nIt’s currently more scalable then pretty much every coin on that list. ", "It OFFERS INSTANT TRANSACTIONS for all syscoin assets. ", "One of the most secure chains out there due to being merge mined with btc. ", "Leading the way in decentralized marketplace, multiple icos upcoming. ", "Real companies with millions of customers are turning to syscoin to use blockchain tech to enhance their business..\nCex.io provides a wide range of services for using bitcoin and other cryptocurrencies. ", "The platform lets users easily trade fiat money with cryptocurrencies and conversely cryptocurrencies for fiat money. ", "For those looking to trade bitcoins professionally, the platform offers personalized and user-friendly trading dashboards and margin trading. ", "Alternatively, CEX also offers a brokerage service which provides novice traders an extremely simple way to buy bitcoin at prices that are more or less in line with the market rate. ", "The Cex.io website is secure and intuitive and cryptocurrencies can be stored in safe cold storage. ", "Check out the Cex.io FAQ\nNOTES: We created this site in 2015, here three years later (in 2018) the market has evolved and changed a considerable amount. ", "Thus, presenting a list of cryptocurrencies went from being a reasonable thing to do to an impossible task for a site that doesn’t make listing cryptocurrencies its main focus. ", "For a list of most of the current cryptocurrencies, you can check out CoinMarketCap.com. ", "Our brief list below will focus only on some top coins that have made it through the years or that are still relevant today and will note some up-and-coming coins.", "\n\nLitecoin was one of the first cryptocurrencies after Bitcoin and tagged as the silver to the digital gold bitcoin. ", "Faster than bitcoin, with a larger amount of token and a new mining algorithm, Litecoin was a real innovation, perfectly tailored to be the smaller brother of bitcoin. “", "It facilitated the emerge of several other cryptocurrencies which used its codebase but made it, even more, lighter“. ", "Examples are Dogecoin or Feathercoin.", "\n\nFor those who really want the real cryptocurrency experience, I think the simplest place to buy, sell, and store coins is Coinbase (and our tutorial below will help you get set up with that), but you can only buy, sell, and store Bitcoin, Ethereum, Litecoin, Bitcoin Cash, and a small (but growing) selection of other coins on Coinbase. ", "However, if you are serious about trading cryptocurrency, you’ll want to also sign up for another exchange like Coinbase’s Coinbase Pro, Bittrex, Binance, or Kraken (and may want to find other solutions for wallets to store your coins in like TREZOR). ", "See our list of exchanges for beginners for a more complete list. ", "TIP: Even if you are going to get fancy with wallets and exchanges, Coinbase is a good starting point because it works as a simple on-ramp / off-ramp for fiat (i.e. you can easily trade dollars for cryptos on Coinbase, and this is not true of most exchanges).", "\n\nThose of you who are looking for cryptocurrency charts live, you have come to the right place! ", "All our cryptocurrency price charts you see on our website are updated in real-time without any delays. ", "On cryptocurrencylist.io you can view charts from any cryptocurrency with your chosen timeframe to examine the price history of this particular crypto.", "\n\nCryptocurrencies are a potential tool to evade economic sanctions for example against Russia, Iran, or Venezuela. ", "In April 2018, Russian and Iranian economic representatives met to discuss how to bypass the global SWIFT system through decentralized blockchain technology.[56] Russia also secretly supported Venezuela with the creation of the petro (El Petro), a national cryptocurrency initiated by the Maduro government to obtain valuable oil revenues by circumventing US sanctions.", "\n\nSo kann man Limits festlegen oder schlichtweg einen geringeren Kapitaleinsatz wählen. ", "Dies ist gerade für Anfänger eine stark zu empfehlende Vorgehensweise. ", "Man hört relativ oft, dass der Krypto Handel eine Unterart des normalen Währungshandels ist. ", "Die Gemeinsamkeiten sind tatsächlich eklatant. ", "Sie fangen schon damit an, dass der User oftmals einen Devisenbroker benötigt, um überhaupt mit digitalen Währungen handeln zu können.", "\nRegulation: One of the most important drivers of the cryptocurrency price is the regulation. ", "In news feed of FXStreet, the regulation follows all the news regarding the new laws and considerations that the main market regulators are creating. ", "Understanding the positions of the individual governments towards different cryptocurrencies is crucial for the trading of the individual investor.", "\nBut I am very confused and not sure who to trust in this crypto world. ", "After reading many comments and news, I found out that most of these brokers are not safe to buy crypto positions from. ", "I live in Norway and I would ask you kindly to advice me which broker I should be able to buy crypto positions from. ", "I signed up with Binance for one week ago, but still did not get any verification yet from them.", "\n\nNow after you bought yourself some Bitcoin, the time has come to choose your exchange platform. ", "This is where you are able to instantly trade one cryptocurrency into another. ", "Take note of the currency trading pairs – each exchange has a list of their own. ", "There are exchanges, where you are able to only exchange Bitcoin to Altcoin, but not Altcoin to any other Altcoin. ", "This hinders the ability to trade fast and flexible. ", "That is why we have made a list of the best crypto-exchanges. ", "On this list you can find the most competitive bitcoin brokers available, offering lowest transaction fees.", "\n\nAus meiner Sicht besonders spannend sind die Kapitel zu den weiteren Anwendungsmöglichkeiten, die sich aus dem Blockchain-Konzept für ein \"automatisiertes Vertrauen\" ergeben. ", "Die Konsequenz, darüber nachzudenken, wofür wir in Zukunft Vertrauensvolle (zentrale) Instanzen benötigen und wo und wie wir Vertrauen algorithmisch (dezentral) abbilden können, finde ich absolut faszinierend. ", "Es geht also nicht nur bei Banken und in der Finanzwelt um die Fragen „zentral oder dezentral“, „Torwächter oder Macht“, sondern auch in vielen anderen Branchen. - ", "Eine absolute Leseempfehlung!", "\nInvestieren ist spekulativ. ", "Bei der Anlage ist Ihr Kapital in Gefahr. ", "Diese Website ist nicht für die Verwendung in Rechtsordnungen vorgesehen, in denen der beschriebene Handel oder die beschriebenen Investitionen verboten sind, und sollte nur von Personen und auf gesetzlich zulässige Weise verwendet werden. ", "Ihre Investition ist in Ihrem Land oder Wohnsitzstaat möglicherweise nicht für den Anlegerschutz geeignet. ", "Führen Sie daher Ihre eigene Due Diligence durch. ", "Diese Website steht Ihnen kostenlos zur Verfügung, wir erhalten jedoch möglicherweise Provisionen von den Unternehmen, die wir auf dieser Website anbieten. ", "Klicken Sie hier für weitere Informationen.", "\nTIP: If you don’t understand the tax implications of trading cryptocurrency tread very carefully. ", "There are some nasty traps you could fall into when trading coins. ", "For one, they are not necessarily considered “like-kind assets.” ", "If that is confusing, then consider sticking with trading USD for coins in Coinbase until you grasp the concept. ", "Learn about cryptocurrency and taxes.", "\nBitcoin ist so etwas wie die Mutter aller Kryptowährungen. ", "Sie gibt es mittlerweile seit ca. ", "neun Jahren und interessanterweise weiß man bis heute nicht, wer ihr eigentlicher Gründer ist. ", "Das hat der Nachfrage aber nicht geschadet. ", "Zum Ende des Jahres 2017 sah es sogar kurz danach aus, als würde er die 17500 € Grenze knacken. ", "Leider hat das nicht geklappt und im April 2018 liegt der Kurs bei ca. ", "5000€.\nWelcome to the sixteenth Coin Report. ", "In today’s report, I will be assessing the fundamental and technical strengths and weaknesses of NKN. ", "This will comprise of an analysis of a number of significant metrics, an evaluation of the project’s community and development and an overview of its price-history. ", "The report will conclude with a grading out of 10. ", "NKN (New Kind of Network – a name that tips its hat to Stephen Wolfram; the man behind A New Kind of Science) was launched in January 2018, but raised funding via an ICO in April 2018, after which its token was created. ", "The ICO sold out in minutes and raised over $12mn, though this figure is difficult to verify as there are discrepancies between sources. ", "The token itself recently underwent a swap to become part of the Ethereum ecosystem; a swap that is still ongoing, until the end of this month. ", "NKN was priced at $0.13 during the ICO. ", "The token itself operates using a Proof-of-Relay consensus mechanism; a novel development...\nSollten diese Recht behalten, so steht auch eine vermehrten Handel an Kryptowährungen nichts mehr im Wege. ", "Möglicherweise ist der Zeitpunkt für einen Einstieg in das Krypto Trading daher günstig. ", "Mit dem richtigen Anbieter findet man auch Crypto Trading deutsch vor, was insbesondere Händler aus der Bundesrepublik begrüßen dürften. ", "Wer also Kryptowährungen handeln möchte, findet online weitreichende Möglichkeiten und auf unserer Startseite weitere interessante Infos.", "\nKurs-Update: Bitcoin, Ethereum, Ripple und Co. Märkte vor 1 Tag Die Gesamtmarktkapitalisierung aller Kryptowährungen beläuft sich auf 144 Mrd. ", "US-Dollar, wovon Bitcoin einen Marktanteil von 50 Prozent für sich beansprucht. ", "Das Handelsvolumen betrug in den letzten 24 Stunden 40 Mrd. ", "US-Dollar. ", "Durchschnittlich haben sich die Kurse der Kryptowährungen innerhalb der Top 10 um 1 Prozent geändert.", "\nThank you for this guide. ", "Hopefully there are no stupid questions here – but a quick clarification would be helpful. ", "This and some of your other guides make reference to “requesting a transaction” at the very beginning of the process. ", "What does that mean? ", "Is is simply the request to purchase bitcoin in exchange for USD or whatever medium of exchange? ", "Thank you in advance!", "\nIf you happen to own a business and if you’re looking for potential new customers, accepting cryptocurrencies as a form of payment may be a solution for you. ", "The interest in cryptocurrencies has never been higher and it’s only going to increase. ", "Along with the growing interest, also grows the number of crypto-ATMs located around the world. ", "Coin ATM Radar currently lists almost 1,800 ATMs in 58 countries.", "\nIt’s currently more scalable then pretty much every coin on that list. ", "It OFFERS INSTANT TRANSACTIONS for all syscoin assets. ", "One of the most secure chains out there due to being merge mined with btc. ", "Leading the way in decentralized marketplace, multiple icos upcoming. ", "Real companies with millions of customers are turning to syscoin to use blockchain tech to enhance their business..\nDon’t Monopolize the Conversation. ", "We appreciate passion and conviction, but we also believe strongly in giving everyone a chance to air their thoughts. ", "Therefore, in addition to civil interaction, we expect commenters to offer their opinions succinctly and thoughtfully, but not so repeatedly that others are annoyed or offended. ", "If we receive complaints about individuals who take over a thread or forum, we reserve the right to ban them from the site, without recourse.", "\nBytom wants to improve income asset management by making it easy to transfer these assets via smart contacts. ", "They also want to digitize the management of non-public securities and options, in addition to increasing asset securitization. ", "Asset securitization is simply the ability to register and tokenize any asset from the atomic world, and easily put it on the blockchain.", "\nBefore we take a closer look at some of these alternatives to bitcoin, let’s step back and briefly examine what we mean by terms like cryptocurrency and altcoin. ", "A cryptocurrency, broadly defined, is virtual or digital money which takes the form of tokens or “coins.” ", "While some cryptocurrencies have ventured into the physical world with credit cards or other projects, the large majority remain entirely intangible. ", "The “crypto” in cryptocurrencies refers to complicated cryptography which allows for a particular digital token to be generated, stored, and transacted securely and, typically, anonymously. ", "Alongside this important “crypto” feature of these currencies is a common commitment to decentralization; cryptocurrencies are typically developed as code by teams who build in mechanisms for issuance (often, although not always, through a process called “mining”) and other controls. ", "Cryptocurrencies are almost always designed to be free from government manipulation and control, although as they have grown more popular this foundational aspect of the industry has come under fire.", "\nSchließlich liegt jeder digitalen Währung eine eigene ganz besondere Funktionsweise zu Grunde, welches im Laufe der Zeit auch über deren Erfolg und Wertsteigerung entscheiden dürfte. ", "Außerdem informieren sich kluge Trader ohnehin immer über die zu handelnden Produkte, um dann vor keine Überraschungen gestellt zu werden, wenn es um den konkreten Handel geht.", "\n\n“While it’s still fairly new and unstable relative to the gold standard, cryptocurrency is definitely gaining traction and will most certainly have more normalized uses in the next few years. ", "Right now, in particular, it’s increasing in popularity with the post-election market uncertainty. ", "The key will be in making it easy for large-scale adoption (as with anything involving crypto) including developing safeguards and protections for buyers/investors. ", "I expect that within two years, we’ll be in a place where people can shove their money under the virtual mattress through cryptocurrency, and they’ll know that wherever they go, that money will be there.” – ", "Sarah Granger, Author, and Speaker.", "\n\nMarkets are dirty. ", "But this doesn‘t change the fact that cryptocurrencies are here to stay – and here to change the world. ", "This is already happening. ", "People all over the world buy Bitcoin to protect themselves against the devaluation of their national currency. ", "Mostly in Asia, a vivid market for Bitcoin remittance has emerged, and the Bitcoin using darknets of cybercrime are flourishing. ", "More and more companies discover the power of Smart Contracts or token on Ethereum, the first real-world application of blockchain technologies emerge.", "\n\nMostly due to its revolutionary properties cryptocurrencies have become a success their inventor, Satoshi Nakamoto, didn‘t dare to dream of it. ", "While every other attempt to create a digital cash system didn‘t attract a critical mass of users, Bitcoin had something that provoked enthusiasm and fascination. ", "Sometimes it feels more like religion than technology.", "\n\nEthereum is more than a peer-to-peer currency created by Vitalik Buterin; it operates as an infrastructure. ", "The technology launched during 2015 with its first offering of ether, the Ethereum altcoin, raising $18.5 million. ", "The centralized platform provides cryptocurrency, but it also allows the blockchain to be used for developing a variety of applications, such as contracts and crowdsourcing.", "\nCryptocurrencies are encrypted digital currencies which are transferred between peers. ", "They are decentralized, meaning not governed by any bank or government institution. ", "They are a sequence of encrypted codes transmitted and stored over a network. ", "All transactions are confirmed and stored on a public ledger. ", "The system uses other complex techniques to certify and validate the record keeping process. ", "Lack of regulation for cryptocurrencies mean that they are highly volatile by nature, and an investment with this can make a lot of money fast, and at the same time it can turn and one can lose money fast. ", "The reason it is not yet accepted by a lot of businesses is partly due to the lack of regulation. ", "There is a set amount of digital coins that can be created and which was outlined from the beginning, after that number is reached no further coins can be produced. ", "The reality is such, that Bitcoin and digital currencies prices rise and drop for various reasons such as media and bad press, news events, and government statements, more people are using it and for this reason the price is rising. ", "Their unpredictability makes it exciting for most traders. ", "Moving forward there are discussions on how to manage the currencies and that in itself can swing the price.", "\n\n2) Pseudonymous: Neither transactions nor accounts are connected to real-world identities. ", "You receive Bitcoins on so-called addresses, which are randomly seeming chains of around 30 characters. ", "While it is usually possible to analyze the transaction flow, it is not necessarily possible to connect the real world identity of users with those addresses.", "\n\nZcash, a decentralized and open-source cryptocurrency launched in the latter part of 2016, looks promising. “", "If bitcoin is like http for money, zcash is https,\" is one analogy zcash uses to define itself. ", "Zcash offers privacy and selective transparency of transactions. ", "Thus, like https, zcash claims to provide extra security or privacy where all transactions are recorded and published on a blockchain, but details such as the sender, recipient, and amount remain private. ", "Zcash offers its users the choice of “shielded” transactions, which allow for content to be encrypted using advanced cryptographic technique or zero-knowledge proof construction called a zk-SNARK developed by its team. ", "As of February 9, 2019, Zcash had a market cap of $291.25 million and a value per token of $49.84.", "\nThere is great joy in trying to buy into an average position on the top coins (but not Bitconnect https://cryptocurrencyfacts.com/2017/11/05/is-bitconnect-a-scam/) and then incrementally taking profits. ", "There is nothing but sorrow ahead for those who go chasing unreal (and let’s stress that term in every sense) gains from things like “the Billion Coin.”", "\n\nVerification Requirements – The vast majority of the Bitcoin trading platforms both in the US and the UK require some sort of ID verification in order to make deposits & withdrawals. ", "Some exchanges will allow you to remain anonymous. ", "Although verification, which can take up to a few days, might seem like a pain, it protects the exchange against all kinds of scams and money laundering.", "\n\nCryptocurrencies are digital gold. ", "Sound money that is secure from political influence. ", "Money that promises to preserve and increase its value over time. ", "Cryptocurrencies are also a fast and comfortable means of payment with a worldwide scope, and they are private and anonymous enough to serve as a means of payment for black markets and any other outlawed economic activity.", "\n\nThe total market capitalization of cryptocurrencies is growing rapidly, recently surpassing more $800 billion and hitting an all-time high - with a more than 850 percent increase since the start of the year. ", "When most people think of cryptocurrencies they think of Bitcoin, which was the first to market and is the current leader. ", "But it’s not the only game in town. ", "There are around 1,100 different cryptocurrencies to date. ", "Not sure which is best? ", "Check out this list of top cryptocurrencies to find out more about each, how cryptocurrency works, and where to buy cryptocrurency.", "\nIf your focus is to conduct crypto-to-crypto trading, Binance is one of the best options. ", "Ranked as one of the most popular cryptocurrency exchanges worldwide, they provide you with impressive offerings along with an extremely low trading fee. ", "Although the Binance platform is a young entrant into the market, it is rapidly growing, and holds a huge selection of altcoins with Bitcoin, Ethereum, and Tether pairings.", "\nNextcoinmarket Has Influenced Products. ", "Nextcoinmarket Associated With Foreign Exchange, Common Assets And Other Underlying Variables, Involves A High Level Of Risk And The Possibility Of Loss Of Some Or All Of Your Investment.please Consider Carefully Whether Trading Or Investing In Bitcoin Is Appropriate To Your Financial Situation. ", "Only Risk Capital Should Be Used When Trading Or Investing In Bitcoin. ", "You Must Review The Terms Of Service And This Risk Disclosure Prior To Setting Up An Account. ", "Website Is Owned And Operated By Operational Address: 28 Dzavaharlal Neru Str., ", "Entrance Silver Business Center, 1324, Sofia, Bulgaria\nTrading foreign exchange on margin carries a high level of risk and may not be suitable for all investors. ", "The high degree of leverage can work against you as well as for you. ", "Before deciding to trade foreign exchange you should carefully consider your investment objectives, level of experience and risk appetite. ", "The possibility exists that you could sustain a loss of some or all of your initial investment and therefore you should not invest money that you cannot afford to lose. ", "You should be aware of all the risks associated with foreign exchange trading and seek advice from an independent financial advisor if you have any doubts.", "\nAls besonders attraktiv gelten auch mobile Handelsmöglichkeiten. ", "Sie runden das Angebot eines guten Cryptotraders elegant ab und sorgen für zeitliche und örtliche Flexibilität des Händlers. ", "Am Besten dürfte der mobile Handel mithilfe von spezieller Software, bzw. ", "Apps funktionieren. ", "Man sollte darauf achten, ob es entsprechende Anwendungen für Android und Ios Geräte gibt. ", "Ist dies der Fall, so kann man sich seiner mobilen Endgeräte bedienen um Trades einfach und schnell zu absolvieren.", "\nAm angenehmsten ist wohl das Vorhandensein eines Live Chats. ", "Aber auch kostenlose Hotlines sind direkte Kontaktmöglichkeiten, welche hoffentlich kurze Wartezeiten aufweisen. ", "Schließlich sollte auch eine E-Mail Adresse bzw. ", "ein Kontaktformular vorzufinden sein, welches den schriftlichen Weg ermöglicht. ", "Sind all diese Formen gegeben, so bleibt kaum Grund zur Klage und man darf sich über einen starken Support freuen.", "\nElcurrency is a cryptocurrency trading platform where you can trade on different kinds of digital currencies and generate as much profit as you can with the help of our experience and services. ", "Join us in our journey and be part of our prosperous team. ", "Our company was founded at the early stage of cryptocurrency industry, and has grown together with it.", "\n\nCryptocurrencies hold the promise of making it easier to transfer funds directly between two parties in a transaction, without the need for a trusted third party such as a bank or credit card company; these transfers are facilitated through the use of public keys and private keys for security purposes. ", "In modern cryptocurrency systems, a user's \"wallet,\" or account address, has the public key, and the private key is used to sign transactions. ", "Fund transfers are done with minimal processing fees, allowing users to avoid the steep fees charged by most banks and financial institutions for wire transfers.", "\n\nEthereum: Is probably the third most important coin [here in 2018, I’d argue that it is the second most important after Bitcoin]. ", "Ethereum doesn’t have the longevity at the top like Litecoin, but it has some unique features and a market cap that make it a real contender. ", "Most ICOs (Initial Coin Offerings) use Ethereum. ", "It has a less intimidating cost that Bitcoin and has the second highest market cap. ", "On that note, Ethereum classic is also notable. ", "Ethereum is a spin-off (aka “hard fork”) from what isn’t today called Ethereum classic (like how our next up coin, Bitcoin cash, is a spin-off of Bitcoin.) ", "NOTE: Ethereum is a fork of another relevant coin called Ethereum Classic.", "\n\nHowdy, Welcome to popular Cryptocurrency blog 'CoinSutra'. ", "I'm Harsh Agrawal, a tech enthusiast & Digital nomad from New Delhi, India.", "I started CoinSutra to help users around the globe to learn about popular Cryptocurrencies.", "Here at CoinSutra I write about Bitcoin Wallet, Cryptocurrency wallets, Online Privacy & Security, VPN experiences & making money from Crypto.", "\nActive traders looking to speculate on Bitcoin over the short or medium term may find that trading CFD/derivatives on Bitcoin using an online forex broker will provide them with 24hour trading, potentially lower margin, and the ability to go either long or short. ", "Because of counter-party risk, choosing a broker is just as important as finding one with the best trading tools or commission rates.", "\nTIP: The top coins by market cap are usually the best bets. ", "Most of the top coins today were toward the top of the list back in 2015. ", "There are a few coins high on the list today that you should be wary of, for example, Bitconnect. ", "There are a few new ones, like the Bitcoin forks, Bitcoin Cash and Bitcoin Gold, that are new to the list. ", "Meanwhile, there are a few that didn’t stand up to the test of time, like Peercoin. ", "As a rule of thumb, historically speaking, the higher the coin is on the list, the more solid it is an investment over time.", "\nTaking the first option listed above, which is to buy the underlying, you become the direct holder of the digital asset. ", "Upon purchase, the cryptocurrency is sent to your bitcoin address or account (wallet) with the exchange. ", "From there, you can transfer the crypotocurrency to any bitcoin address or wallet address using your private key that verifies you control ownership of the asset.", "\nRipple: Similarly to Bitcoin and Ethereum, Ripple is a decentralized payment system based on open-source. ", "In fact, the project came years before its competitors, specifically in 2004 at the hands of Ryan Fugger. ", "However, it was not until 2016 when he obtained the license to operate it. ", "Unlike the previous ones, Ripple does not have mining operations nor will new currencies be created in the future. ", "It is the third largest cryptocurrency in terms of market capitalization.", "\nThe benefit of a USD wallet on Coinbase is that you can put money in that and then, once the deposit clears, use it to buy coins immediately moving forward. ", "If you try to buy directly with your bank account, the transaction can take about a week. ", "Given this it is smart to fund your USD wallet or buy USDC and then use that moving forward to buy crypto. ", "You’ll still need to wait for the deposit to clear, but once it is cleared with your bank you can use the funds. ", "You can buy coins on Coinbase.com via your USD wallet (just toggle to USD wallet instead of bank account when making a purchase), although you’ll still pay the broker fee, and you can buy coins on Coinbase Pro using USD or USDC for low or no fees (remember, no fees for limit orders, low fees for market orders).", "\n\nFor those who really want the real cryptocurrency experience, I think the simplest place to buy, sell, and store coins is Coinbase (and our tutorial below will help you get set up with that), but you can only buy, sell, and store Bitcoin, Ethereum, Litecoin, Bitcoin Cash, and a small (but growing) selection of other coins on Coinbase. ", "However, if you are serious about trading cryptocurrency, you’ll want to also sign up for another exchange like Coinbase’s Coinbase Pro, Bittrex, Binance, or Kraken (and may want to find other solutions for wallets to store your coins in like TREZOR). ", "See our list of exchanges for beginners for a more complete list. ", "TIP: Even if you are going to get fancy with wallets and exchanges, Coinbase is a good starting point because it works as a simple on-ramp / off-ramp for fiat (i.e. you can easily trade dollars for cryptos on Coinbase, and this is not true of most exchanges).", "\n\nELcurrency are influenced products. ", "ELcurrency associated with foreign exchange, common assets and other underlying variables, involves a high level of risk and a possibility of loss of some or all of your investment.", "Please consider carefully whether trading or investing in bitcoin is appropriate to your financial situation. ", "Only risk capital should be used when trading or investing in bitcoin. ", "You must review Terms of Service and this Risk Disclosure prior to establishing an account.", "\n\nNow after you bought yourself some Bitcoin, the time has come to choose your exchange platform. ", "This is where you are able to instantly trade one cryptocurrency into another. ", "Take note of the currency trading pairs – each exchange has a list of their own. ", "There are exchanges, where you are able to only exchange Bitcoin to Altcoin, but not Altcoin to any other Altcoin. ", "This hinders the ability to trade fast and flexible. ", "That is why we have made a list of the best crypto-exchanges. ", "On this list you can find the most competitive bitcoin brokers available, offering lowest transaction fees.", "\n\nIf I like an ICO, read the white paper, research the founders, google it for hours, and come away with a good vibe, I will put a small amount of Ether (or whatever currency) on it and see where it goes as a long term play. ", "90% of the time it just ties up money that would have been better spent holding the coin. ", "Further, you can often buy the token cheaper on the open market using a platform like EtherDelta. ", "Further, you can often wait until it is listed on a major exchange and then buy it.", "\nThose of you who are looking for cryptocurrency charts live, you have come to the right place! ", "All our cryptocurrency price charts you see on our website are updated in real-time without any delays. ", "On cryptocurrencylist.io you can view charts from any cryptocurrency with your chosen timeframe to examine the price history of this particular crypto.", "\nLitecoin: Litecoin is probably the second most important digital coin [true in 2015, it is still relevant today]. ", "It had the third-highest market cap as of June 2015, but today it sits closer to 7. ", "Despite the decline, CPU mining is still sort of possible, people know what a Litecoin is, it uses essentially the same technology of Bitcoin, and it costs about 1/50th – 1/100th of what Bitcoin does (depending on the day). ", "A Litecoin is a lot like a Bitcoin before the whole ‘Silk Road‘ controversy, or as some people would say “a Litecoin is like a Bitcoin except with a value closer to what a reasonable person would expect a digital coin to have in a rational market.”", "\nRipple: Similarly to Bitcoin and Ethereum, Ripple is a decentralized payment system based on open-source. ", "In fact, the project came years before its competitors, specifically in 2004 at the hands of Ryan Fugger. ", "However, it was not until 2016 when he obtained the license to operate it. ", "Unlike the previous ones, Ripple does not have mining operations nor will new currencies be created in the future. ", "It is the third largest cryptocurrency in terms of market capitalization.", "\nAs a cryptocurrency attracts more interest, mining becomes harder and the amount of coins received as a reward decreases. ", "For example, when Bitcoin was first created, the reward for successful mining was 50 BTC. ", "Now, the reward stands at 12.5 Bitcoins. ", "This happened because the Bitcoin network is designed so that there can only be a total of 21 mln coins in circulation.", "\nNxt: Not only does this nifty coin sport a name similar to Steve Job’s other company; it uses a cool and different algorithm for producing coins. ", "This algorithm – an implementation of a proof-of-stake scheme rather than proof-of-work – may be less burdensome on the environment and has long-term potential. ", "It may be worth a tad less than the other coins we recommend; it is worth about a penny on the dollar on a good day. ", "However, less cost per coin means you have less to lose if the coin value deflates. ", "Nxt is like Namecoin. ", "It had a super cool code but didn’t though perform at the same level as other cryptos (until late 2017 where it saw a notable price hike). ", "It is still priced very low in USD.", "\nMonero (XMR) was created in April 2014 and focuses on privacy, decentralization and scalability. ", "It is a secure, private and untraceable currency system. ", "Monero uses a special kind of cryptography to ensure that all of its transactions remain 100% unlinkable and untraceable. ", "The word ‘Monero’ comes from the language Esperanto where it literally means ‘coin’.", "\nThose of you who are looking for cryptocurrency charts live, you have come to the right place! ", "All our cryptocurrency price charts you see on our website are updated in real-time without any delays. ", "On cryptocurrencylist.io you can view charts from any cryptocurrency with your chosen timeframe to examine the price history of this particular crypto.", "\nIMPORTANT DISCLAIMER: All content provided herein our website, hyperlinked sites, associated applications, forums, blogs, social media accounts and other platforms (“Site”) is for your general information only, procured from third party sources. ", "We make no warranties of any kind in relation to our content, including but not limited to accuracy and updatedness. ", "No part of the content that we provide constitutes financial advice, legal advice or any other form of advice meant for your specific reliance for any purpose. ", "Any use or reliance on our content is solely at your own risk and discretion. ", "You should conduct your own research, review, analyse and verify our content before relying on them. ", "Trading is a highly risky activity that can lead to major losses, please therefore consult your financial advisor before making any decision. ", "No content on our Site is meant to be a solicitation or offer.", "\nBitcoin continues to lead the pack of cryptocurrencies, in terms of market capitalization, user base and popularity. ", "Nevertheless, virtual currencies such as ethereum and ripple, which are being used more for enterprise solutions, are becoming popular, while some altcoins are being endorsed for superior or advanced features vis-à-vis bitcoins. ", "Going by the current trend, cryptocurrencies are here to stay but how many of them will emerge as leaders amid the growing competition within the space will only be revealed with time.", "\nCoinSwitch supports most of the wallets. ", "If your wallet address is recognized as invalid, there is always a reason: You confused Dash (DASH) with Dashcoin DSH). ", "These are two different coins with two different amounts. ", "You confused Factom (FCT) with Fantomcoin (FCN). ", "Different currencies as well. ", "Typo or character missing. ", "Make sure that the address you specify matches the actual address of your wallet.", "\nIn early 2018, Bloomberg News reported the largest cryptocurrency exchanges based on the volume and estimated revenues data collected by CoinMarketCap.[29] Similar statistics was reported on Statista in a survey by Encrybit to understand cryptocurrency exchange problems. ", "According to the survey, the top three cryptocurrency exchanges are Binance, Huobi, and OKEX. ", "Other data points in the survey included the problems that cryptocurrency traders experience with cryptocurrency exchanges and the expectation of traders. ", "Security and high trading fees are the top concerns.[30][31] The exchanges are all fairly new and privately held. ", "Several do not report basic information such as the names of the owners, financial data, or even the location of the business.[32] A study by Mistertango reported that 88% of cryptocurrency exchanges are in favor of regulation for operational certainly and price stability.[33]\n\nConsider signing up for another exchange and trading cryptocurrency for cryptocurrency (and then transferring that back into Bitcoin, Bitcoin Cash, Ethereum, and Litecoin, and then back into Coinbase, and then back into USD). ", "Don’t forget to record your transactions for the tax man and to brush up on the tax implications (I can’t stress that enough).", "\nIn 2006, US-based digital currency exchange business GoldAge Inc., a New York state business, was shut down by the US Secret Service after operating since 2002.[8] Business operators Arthur Budovsky and Vladimir Kats were indicted \"on charges of operating an illegal digital currency exchange and money transmittal business\" from their apartments, transmitting more than $30 million to digital currency accounts.[5] Customers provided limited identity documentation, and could transfer funds to anyone worldwide, with fees sometimes exceeding $100,000.[5] Budovsky and Kats were sentenced in 2007 to five years in prison \"for engaging in the business of transmitting money without a license, a felony violation of state banking law\", ultimately receiving sentences of five years probation.[9]\nCryptocurrencies hold the promise of making it easier to transfer funds directly between two parties in a transaction, without the need for a trusted third party such as a bank or credit card company; these transfers are facilitated through the use of public keys and private keys for security purposes. ", "In modern cryptocurrency systems, a user's \"wallet,\" or account address, has the public key, and the private key is used to sign transactions. ", "Fund transfers are done with minimal processing fees, allowing users to avoid the steep fees charged by most banks and financial institutions for wire transfers.", "\nJust figure out, what coin do you want to buy and seek for a stable wallet version. ", "As a rule, each coin has its official wallet client. ", "Note that each wallet has its unique address or a tag with a private key that is required to restore your wallet if lost. ", "CoinSwitch never asks your private keys. ", "Store them in a safe place and never show anyone. ", "Once private keys are stolen, your cryptocurrency wallet with all the coins will be lost forever.", "\nInstead, block producers generate blocks and are rewarded in EOS tokens based on their production rates. ", "EOS includes a complex system of rules to govern this process, with the idea being that the network will ultimately be more democratic and decentralized than those of other cryptocurrencies. ", "As of October 5, 2018, EOS had a market cap of $2.49 billion and a per token value of $2.74.", "\nThere is lots of value created by ‘pump- & dumpers’ so watch out! ", "Always set a goal, which you want to achieve, e.g. 2% or 35 USD per day. ", "If you don’t check you exchanges daily, then the best thing you could probably do is add a limit order. ", "A limit order is executed, when a specific price is reached. ", "For example, if you buy Ether for 0.05 BTC. ", "You can make a limit sell order for 0.075 BTC. ", "This means that you will automatically sell your Ether when the value is higher than 0.075 BTC. ", "This assures, that you don’t miss any big price movements- it can always fall back to 0.05 before you can see the profit opportunity.", "\nHi, I live in the US and bought 2 BTC last week at Coinbase on Mar 15 and haven’t seen the Bitcoins appear in my wallet yet (today is 20th Mar). ", "Even after providing personal details and being promised to buy/sell instantly I don’t seen that happening any time soon. ", "And to top it all they charged me 1.49% in fees (can provide screenshots of the transactions).", "\nBitcoin: Bitcoin is an easy pick. ", "It was the first major usable cryptocurrency; it has the highest market cap; its coins trade at the highest cost of all cryptocurrencies (about USD 225 as of June 2015, but as high as $5,000 during early September 2017). ", "Despite the big increase in price, Bitcoin seems to be the best choice for anyone entering the cryptocurrency space. ", "It is the most familiar and invested-in coin. ", "Primarily Bitcoin is the reason anyone is talking about cryptocurrency in the first place. ", "You might not want to start a CPU-based Bitcoin mining company in 2017 or start buying coins for $5,000 each (which is fine since you can buy fractions of a coin) but it’s still going to take 1st place on our list.", "\nAn increase in cryptocurrency mining increased the demand of graphics cards (GPU) in 2017.[37] Popular favorites of cryptocurrency miners such as Nvidia's GTX 1060 and GTX 1070 graphics cards, as well as AMD's RX 570 and RX 580 GPUs, doubled or tripled in price – or were out of stock.[38] A GTX 1070 Ti which was released at a price of $450 sold for as much as $1100. ", "Another popular card GTX 1060's 6 GB model was released at an MSRP of $250, sold for almost $500. ", "RX 570 and RX 580 cards from AMD were out of stock for almost a year. ", "Miners regularly buy up the entire stock of new GPU's as soon as they are available.[39]" ]

{ "pile_set_name": "Pile-CC" }

[ 0, 0, 0, 0, 0, 0.006060606060606061, 0, 0, 0, 0.014084507042253521, 0.010638297872340425, 0, 0.017045454545454544, 0, 0.015151515151515152, 0.010638297872340425, 0, 0.0055248618784530384, 0, 0, 0.01098901098901099, 0, 0.01818181818181818, 0, 0, 0, 0, 0, 0.005494505494505495, 0, 0, 0, 0.011235955056179775, 0, 0, 0.005917159763313609, 0, 0, 0.0058997050147492625, 0.01984126984126984, 0, 0.011583011583011582, 0, 0, 0, 0, 0.005420054200542005, 0.011363636363636364, 0.028169014084507043, 0.021505376344086023, 0.02127659574468085, 0.014925373134328358, 0, 0.006666666666666667, 0, 0, 0, 0, 0.010416666666666666, 0, 0, 0, 0, 0, 0, 0, 0.011299435028248588, 0.009523809523809525, 0.024390243902439025, 0, 0, 0.023809523809523808, 0.016666666666666666, 0.018691588785046728, 0.04, 0.019230769230769232, 0.046511627906976744, 0.010101010101010102, 0, 0, 0.017699115044247787, 0.02702702702702703, 0.03333333333333333, 0.029411764705882353, 0.021052631578947368, 0.022727272727272728, 0.020833333333333332, 0.028169014084507043, 0.022222222222222223, 0.00980392156862745, 0, 0, 0.013636363636363636, 0.0072992700729927005, 0.006944444444444444, 0.05, 0.005, 0.011235955056179775, 0.021897810218978103, 0.014598540145985401, 0.006944444444444444, 0.0125, 0, 0, 0.019801980198019802, 0, 0, 0, 0, 0.010309278350515464, 0, 0, 0, 0, 0, 0, 0.01818181818181818, 0, 0, 0, 0, 0, 0, 0.009009009009009009, 0, 0, 0, 0, 0, 0, 0, 0, 0.016304347826086956, 0.005681818181818182, 0, 0, 0, 0, 0.02857142857142857, 0, 0, 0, 0, 0, 0.013245033112582781, 0.00684931506849315, 0.006134969325153374, 0, 0.00909090909090909, 0.008695652173913044, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.01020408163265306, 0.004901960784313725, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.01098901098901099, 0, 0.011627906976744186, 0, 0, 0.014084507042253521, 0.010638297872340425, 0, 0, 0, 0, 0, 0, 0.045454545454545456, 0.016, 0, 0, 0.03296703296703297, 0.034782608695652174, 0.03225806451612903, 0.008849557522123894, 0, 0.025, 0.02631578947368421, 0, 0, 0, 0, 0, 0, 0, 0.007042253521126761, 0.04081632653061224, 0, 0, 0.00641025641025641, 0.013513513513513514, 0.01639344262295082, 0.02666666666666667, 0.02197802197802198, 0.035211267605633804, 0.0037735849056603774, 0, 0.01639344262295082, 0, 0.01020408163265306, 0.028037383177570093, 0.011904761904761904, 0, 0, 0, 0, 0.009345794392523364, 0.009433962264150943, 0, 0, 0, 0.006329113924050633, 0, 0.009345794392523364, 0, 0.016025641025641024, 0.0058997050147492625, 0.01984126984126984, 0, 0.011583011583011582, 0, 0.0055248618784530384, 0, 0, 0.01098901098901099, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.01020408163265306, 0, 0, 0, 0, 0, 0, 0, 0, 0.009345794392523364, 0.009433962264150943, 0, 0, 0, 0, 0.011111111111111112, 0.024390243902439025, 0, 0.006802721088435374, 0, 0, 0, 0, 0, 0.02857142857142857, 0.02040816326530612, 0, 0.00819672131147541, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.023809523809523808, 0.008333333333333333, 0, 0.02040816326530612, 0, 0, 0, 0.01098901098901099, 0.010638297872340425, 0, 0, 0.0019801980198019802, 0, 0.004553734061930784, 0, 0, 0, 0, 0, 0.024390243902439025, 0, 0, 0.009433962264150943, 0.005235602094240838, 0.010869565217391304, 0, 0, 0, 0, 0, 0, 0, 0, 0.00684931506849315, 0, 0, 0, 0, 0, 0, 0, 0.004672897196261682, 0.010810810810810811, 0, 0.014285714285714285, 0.011363636363636364 ]

[ "ISLAMABAD: A man accused of blasphemy won freedom on Friday after spending over nine years in jail.", "\n\nThe Supreme Court absolved the 58-year-old of the blasphemy charges for want of evidence.", "\n\nPolice arrested him on Sept 29, 2008 for allegedly desecrating pages of the Holy Quran in a mosque in the Sadiq Gunj area of Bahawalnagar and registered a case under Section 295-B of the Pakistan Penal Code (PPC).", "\n\nSuffering from hearing and speech impairments, Akhtar Mohammad had reported the incident to Imam Hafiz Mohammad Munir who informed the local panchayat, where the suspect was beaten up and later handed over to police.", "\n\nThe police filed a charge sheet against him before the Bahawalnagar district and sessions judge who sentenced the man to life imprisonment in 2009.", "\n\nThe Lahore High Court upheld the sentence in 2014. ", "The Supreme Court took up the case the same year and provided the accused a state counsel as he could not afford a lawyer.", "\n\nDuring the hearing on Friday, the defence counsel pointed out that the prosecution evidence produced against the accused was not in accordance with the Evidence Act. ", "He argued that the star witness in the case was a person suffering from hearing and speech impairments and, therefore, could not even be termed a witness under the Evidence Act.", "\n\nThe counsel contended that investigation into the incident did not fulfil the mandatory requirements of the Criminal Procedure Code.", "\n\nAfter hearing the arguments, a Supreme Court bench comprising Justice Dost Mohammad and Justice Qazi Faez Isa absolved the accused of the blasphemy charges with the observation that the person who falsely accused others of desecration of the Holy Quran and revered personalities in fact committed blasphemy.", "\n\nPublished in Dawn, December 30th, 2017" ]

{ "pile_set_name": "OpenWebText2" }

[ 0, 0.01098901098901099, 0.009302325581395349, 0.009174311926605505, 0, 0.018867924528301886, 0.00819672131147541, 0, 0, 0, 0.009708737864077669, 0 ]

[ " Opinions of the United\n2000 Decisions States Court of Appeals\n for the Third Circuit\n\n\n2-28-2000\n\nWoodwind Estates v. Gretkowski\nPrecedential or Non-Precedential:\n\nDocket 99-3280\n\n\n\n\nFollow this and additional works at: http://digitalcommons.law.villanova.edu/thirdcircuit_2000\n\nRecommended Citation\n\"Woodwind Estates v. Gretkowski\" (2000). ", "2000 Decisions. ", "Paper 37.", "\nhttp://digitalcommons.law.villanova.edu/thirdcircuit_2000/37\n\n\nThis decision is brought to you for free and open access by the Opinions of the United States Court of Appeals for the Third Circuit at Villanova\nUniversity School of Law Digital Repository. ", "It has been accepted for inclusion in 2000 Decisions by an authorized administrator of Villanova\nUniversity School of Law Digital Repository. ", "For more information, please contact Benjamin.Carlson@law.villanova.edu.", "\n\fFiled February 28, 2000\n\nUNITED STATES COURT OF APPEALS\nFOR THE THIRD CIRCUIT\n\nNo. ", "99-3280\n\nWOODWIND ESTATES, LTD.,", "\n\n Appellant\n\nv.\n\nW. J. GRETKOWSKI; LARRY SEBRING;\nJAMES DECKER, Individually, and in their capacity as\nSupervisors of Stroud Township; W. TAYLOR WENCK;\nEDWARD CRAMER; FRANK HERTING;\nJOAN KEIPER, Individually, and as members of the\nPlanning Commission of Stroud Township;\nSTROUD TOWNSHIP, MONROE COUNTY, PENNSYLVANIA\n\nOn Appeal from the United States District Court\nfor the Middle District of Pennsylvania\n(D.C. Civil No. ", "97-cv-00472)\nDistrict Judge: Honorable James F. McClure, Jr.\n\nArgued December 7, 1999\n\nBefore: SLOVITER, ROTH and COWEN, Circuit Judg es\n\n(Filed: February 28, 2000)\n\n Marshall E. Anders, Esq. (", "Argued)\n 802 Main Street\n Stroudsburg, PA 18360\n\n Counsel for Appellant\n\f Eugene F. Hickey, II, Esq. (", "Argued)\n Schneider, Gelb, Goffer & Hickey\n 400 Spruce Street, Suite 500\n Scranton, PA 18503\n\n Counsel for Appellees\n\nOPINION OF THE COURT\n\nCOWEN, Circuit Judge.", "\n\nThis appeal is a civil rights action under 42 U.S.C.S 1983\nbrought by plaintiff Woodwind Estates, Ltd. (Woodwind)\nagainst defendants Stroud Township (the \"Township\") and\nindividual officers of the Township. ", "The suit emanated from\nthe failure of the Township to approve development plans\nfor specific property. ", "The central issue on appeal is whether\nthe District Court properly granted defendants' motion for\njudgment as a matter of law on Woodwind's S 1983\nsubstantive due process claim. ", "Because we find that the\nDistrict Court erred in granting this motion, we will reverse\nand remand for further proceedings.", "\n\nWoodwind is a Pennsylvania limited partnership which at\nall times relevant to this action sought to build a\nsubdivision development on seventy-five acres in Stroud\nTownship, Pennsylvania. ", "In August 1995, Woodwind was\nawarded approximately $1.1 million in federal low income\nhousing tax credits by the Pennsylvania Housing Finance\nAgency (\"PHFA\") for developing an \"affordable housing\"\nproject. ", "The project was to consist of one hundred single\nfamily homes for low income families. ", "In order to retain the\nfederal income tax credits, PHFA required Woodwind to\ncomplete the project by December 31, 1997. ", "Woodwind\nsought to obtain subdivision approval for the project\npursuant to the Township's Subdivision and Land\nDevelopment Ordinance (\"SALDO\").", "\n\nWoodwind's first step in the approval process was the\nsubmission of a preliminary development plan (\"the Plan\").", "\nOn March 18, 1996, it submitted the Plan and supporting\ninformation to the Township. ", "Pursuant to the Township's\nordinance the Plan was evaluated initially by the Stroud\n\n 2\n\fTownship Planning Commission. ", "The Planning Commission\nwas to issue an advisory opinion to the Board of\nSupervisors which in turn would make the final decision\nregarding approval.", "\n\nAt a meeting on March 27, 1996, the Planning\nCommission first considered the Plan. ", "At the meeting, the\nattorney for the Planning Commission advised the\nCommission that the Plan satisfied the criteria for approval\nas a subdivision.", "\n\nThe March 27, 1996, meeting also was attended by a\ncitizens group known as the \"Concerned Neighbors of\nWoodwind Estates\" who opposed the project because they\ndid not want low-income residents living in the\nneighborhood. ", "The citizens group was represented at the\nmeeting by a private attorney, Marc Wolfe. ", "On behalf of the\ncitizens group, Wolfe urged the Planning Commission to\ndeny approval for the Preliminary Plan insisting instead\nthat Woodwind seek approval for the project as a planned\nunit development (\"PUD\"). ", "The requirements for approval of\na PUD are more onerous, stringent, and time-consuming\nthan the requirements for subdivision approval.", "\n\nDuring the course of the meeting on March 27, 1996,\nmembers of the Planning Commission echoed the concerns\nof the citizens group about the income-level and the\nsocioeconomic background of prospective tenants from the\nWoodwind project, and the potential adverse economic\neffects of the project on local property values. ", "None of these\nconcerns, however, are conditions for subdivision approval\nunder the Township's ordinance.", "\n\nOn March 27, 1996, the Planning Commission advised\nWoodwind that it would not review the Preliminary Plan\nbecause it was an \"incomplete submission\" lacking certain\ntechnical information. ", "App. ", "at 690. ", "Woodwind thereafter\nsubmitted a revised Preliminary Plan (\"the Revised Plan\")\nwhich contained the requested information.", "\n\nAt a meeting on April 24, 1996, the attorney for the\nPlanning Commission advised the Commission that the\nRevised Plan met the criteria for subdivision approval.", "\nNevertheless, attorney Wolfe urged the Planning\nCommission to deny approval for the Revised Plan, again\n\n 3\n\finsisting that Woodwind obtain approval for the project as\na PUD rather than as a subdivision.", "\n\nThe Planning Commission took no action on the Revised\nPlan for approximately six months. ", "Finally, on October 30,\n1996, the Planning Commission voted 4-2 to recommend to\nthe Board of Supervisors to deny approval of the Revised\nPlan.", "\n\nThe Board of Supervisors adopted the recommendation of\nthe Planning Commission voting unanimously to deny\napproval. ", "Shortly after the vote, the Board of Supervisors\nissued a written notice denying approval, which notice gave\nno reason to Woodwind for the denial. ", "Attorney Wolfe\ncontacted a member of the Board of Supervisors and\nadvised him that the Board's denial letter was legally\ninadequate because it did not state any reason for the\ndecision.", "\n\nFollowing the above conversation, Wolfe himself drafted a\nsecond letter of denial which he sent to the Board of\nSupervisors but not to Woodwind setting forth the alleged\nviolations or shortcomings of the Revised Plan. ", "Quite\nobviously Woodwind was not even in a position to respond\nto Wolfe's letter. ", "Relying heavily upon significant portions of\nthe denial letter drafted by Wolfe, the attorney for the Board\nof Supervisors subsequently sent a letter dated November\n27, 1996, notifying Woodwind of the reasons for the denial.", "\nThe letter included as the primary reason for the denial the\nexact same reason which Wolfe previously had proposed in\nhis draft: \"The Board of Supervisors considers the above\napplication to constitute a Planned Unit Development in\nthat the project includes residential units located on a tract\nof land at least 50 acres in size which is planned for\ndevelopment in its entirety under single ownership or\ncontrol.\" ", "App. ", "at 707.", "\n\nAfter the Revised Plan was denied, Woodwind determined\nthat it was impossible to complete the project by the\nDecember 1997 deadline. ", "When Woodwind could not meet\nthe deadline, PHFA subsequently withdrew financing and\nthe project was canceled.", "\n\nWoodwind initiated this action by filing a complaint in\nUnited States District Court for the Middle District of\n\n 4\n\fPennsylvania alleging that the defendants unlawfully denied\nits application for a planned real estate subdivision. ", "Named\nas defendants were Stroud Township and seven individual\ndefendants who are members of the Planning Commission\nand Board of Supervisors. ", "Woodwind brought suit under 42\nU.S.C. SS 1983 (Count 1), 1981 (Count 2), 1985 (Count 3),\nand various supplemental state law claims (Counts 4-6).", "\n\nBefore submitting the case to the jury, the District Court\ngranted defendants' motion for judgment as a matter of law\nunder Rule 50(a) of the Federal Rules of Civil Procedure.", "\nThe District Court subsequently issued a memorandum\nexplaining its dismissal of Woodwind's S 1983 substantive\ndue process claim.", "\n\nWoodwind appeals only the dismissal of its S 1983\nsubstantive due process claim. ", "We will reverse and remand\nfor further proceedings.", "\n\nWe exercise plenary review of an order granting or\ndenying a motion for judgment as a matter of law and apply\nthe same standard as the District Court. ", "See Lightning\nLube, Inc. v. Witco Corp., 4 F.3d 1153, 1166 (3d Cir. ", "1993).", "\n\"The question is not whether there is literally no evidence\nsupporting the party against whom the motion is directed\nbut whether there is evidence upon which the jury could\nproperly find a verdict for that party.\" ", "Patzig v. O'Neill, 577\nF.3d 841, 846 (3d Cir. ", "1978) (citation omitted) (quotation\nomitted). ", "Such a motion should be granted only if, viewing\nthe evidence in the light most favorable to the nonmovant\nand giving it the advantage of every fair and reasonable\ninference, there is insufficient evidence from which a jury\nreasonably could find liability. ", "See Lightning Lube, 4 F.3d at\n1166.", "\n\nSubstantive due process \"is an area of the law`famous\nfor its controversy, and not known for its simplicity.' \"", "\nDeBlasio v. Zoning Bd. ", "of Adjustment, 53 F.3d 592, 598 (3d\nCir. ", "1995) (quoting Schaper v. City of Huntsville, 813 F.2d\n709, 716 (5th Cir. ", "1987)). ", "On the one hand, federal courts\nare reluctant to sit as appeal boards for disputes between\nland developers and a Township's planning body. ", "On the\nother hand, developers have a due process right to be free\nfrom \"arbitrary and irrational zoning actions.\" ", "Arlington\n\n 5\n\fHeights v. Metropolitan Housing Corporation, 429 U.S. 252,\n263 (1973). ", "During the past decade this court has been\ncalled upon quite frequently to grapple with the obvious\ntension between these two principles in a line of\nsubstantive due process cases. ", "See, e.g., Blanche Road\nCorp. v. Bensalem Twp., ", "57 F.3d 253 (3d Cir. ", "1995);\nDeBlasio v. Zoning Bd. ", "of Adjustment, 53 F.3d 592 (3d Cir.", "\n1995); Parkway Garage v. Philadelphia, 5 F.3d 685 (3d Cir.", "\n1993); Midnight Sessions, Ltd. v. City of Philadelphia, 945\nF.2d 667 (3d Cir. ", "1991); Bello v. Walker, 840 F.2d 1124 (3d\nCir. ", "1988).", "\n\nThe Fourteenth Amendment provides, in part, that\"no\nState [shall] deprive any person of life, liberty, or property\nwithout the due process of law . . . .\" ", "U.S. Const. ", "amend\nXIV, S 1. ", "To prevail on a substantive due process claim\nunder S1983, a plaintiff must establish as a threshold\nmatter that he has a protected property interest to which\nthe Fourteenth Amendment's due process protection\napplies.", "\n\nWoodwind argues that it has a protected property right in\nthe approval of its development plans. ", "Woodwind contends\nthat so long as the development plans met all the\nrequirements of the Township's subdivision ordinance it\nhad an absolute right to approval of the plans as\nsubmitted. ", "The defendants argue that Woodwind has no\nprotected property interest in the approval. ", "Even though\nWoodwind's plans indisputably met the requirements of the\nTownship's subdivision ordinance, according to the\ndefendants they nonetheless had the discretion (based\nupon some unspecified authority) to deny approval of the\nplans as submitted.", "\n\nThis court has recognized \"that the issue of whether and\nwhen state-created property interests invoke substantive\ndue process concerns has not been decided by the Supreme\nCourt.\" ", "Deblasio, 53 F.3d at 598. ", "In this circuit, \" `not all\nproperty interests worthy of procedural due process\nprotection are protected by the concept of substantive due\nprocess.' \" ", "Id. at 598 (quoting Reich v. Beharry, 883 F.2d\n239, 244 (3d Cir. ", "1989)). ", "In Independent Enterprises, Inc. v.\nPittsburgh Water & Sewer Authority, 103 F.3d 1165 (3d Cir.", "\n1997), we stated that \"a substantive due process claim\n\n 6\n\fgrounded in an arbitrary exercise of governmental authority\nmay be maintained only where the plaintiff has been\ndeprived of a `particular quality of property interest,\" id. at\n1179, and further explained that \"all of these cases\ninvolv[ing] zoning decisions, building permits, or other\ngovernmental permission required for some intended use of\nland owned by the plaintiffs,\" id. at n. 12, implicated the\nkind of property interest protected by substantive due\nprocess. ", "It follows that the holder of a land use permit has\na property interest if a state law or regulation limits the\nissuing authority's discretion to restrict or revoke the\npermit by requiring that the permit issue as a matter of\nright upon compliance with terms and conditions\nprescribed by the statute or ordinance. ", "See, e.g.,\nBituminous Materials, Inc. v. Rice County, Minnesota , 126\nF.3d 1068, 1070 (8th Cir. ", "1997).", "\n\nIn this matter, the procedures for approval of subdivision\nplans are set forth in the Township's subdivision ordinance.", "\nIn the section captioned \"Specific Procedures For Plan\nSubmission and Approval,\" the ordinance specifically\nprovides as follows: \". . . ", "the Commission shall determine\nthe extent to which the [subdivision] plan complies with the\nOrdinance and shall recommend to the Board of\nSupervisors that the plan be approved entirely, that it be\nconditionally approved, or that it be disapproved.\" ", "App. ", "at\n503. ", "Under the ordinance, the plan submitted must be\napproved when it complies with all objective criteria for a\nsubdivision.1 In light of the fact that the plan which\nWoodwind submitted indisputably satisfied all of the\nrequirements for approval under the ordinance, and\nbecause the ordinance substantially limits the Township's\ndiscretion regarding approval, we conclude that Woodwind\n_________________________________________________________________\n\n1. ", "Likewise, Pennsylvania courts have stated that\"a subdivision plan\nmust be approved if it complies with [the] applicable regulations . . . .\"", "\nAnderson v. Board of Supervisors of Price Twp., ", "Monroe County, Pa., 437\nA.2d 1308, 1309 (Pa. Cmwlth. ", "1981); See also Pace Resources, Inc. v.\nShrewsbury Twp. ", "Planning Commission, 492 A.2d 818, 821 (Pa. Cmwlth.", "\n1985); Goodman v. Board of Commissioners of the Township of Whitehall,\n411 A.2d 838, 841 (Pa. Cmwlth. ", "1980).", "\n\n 7\n\fhas a protected property interest under the Fourteenth\nAmendment.2\n\nAlong with establishing a protected property interest,\nWoodwind also must demonstrate that it was the victim of\n\"a governmental action [that] was arbitrary, irrational, or\ntainted by improper motive\" in order to show a substantive\ndue process violation under S 1983. ", "Bello v. Walker, 840\nF.2d 1124, 1129 (3d Cir. ", "1988); Accord Parkway Garage v.\nPhiladelphia, 5 F.3d 685, 692 (3d Cir. ", "1993) (a violation of\nsubstantive due process rights is shown where the\ngovernment's actions in a particular case were \"in fact\nmotivated by bias, bad faith or improper motive\"); Blanche\nRoad Corp. v. Bensalem Twp., ", "57 F.3d 253, 263 (3d Cir.", "\n1995) (same). ", "Evidence that the government acted\nimproperly for \"reasons unrelated to the merits of the\napplication for the permits\" may support a finding that the\ngovernment arbitrarily or irrationally abused its power in\nviolation of substantive due process. ", "Bello, 840 F.2d at\n1129; See also Pace Resources Inc. v. Shrewsbury Twp.,", "\n808 F.2d 1023, 1035 (3d Cir. ", "1987) (irrationality not shown\nabsent proof that government took actions against\ndeveloper \"for reasons unrelated to land use planning\"). ", "In\ndisputed factual situations, the determination of the\nexistence of improper motive or bad faith is properly made\nby the jury as the finder of fact. ", "See, e.g. , Bello, 840 F.2d\nat 1130; Midnight Sessions Ltd. v. City of Philadelphia, 945\nF.2d 667, 683 (3d Cir. ", "1991).", "\n_________________________________________________________________\n\n2. ", "According to the defendants, the District Court correctly dismissed the\ncase \"on the ground that the developer failed to establish a\nconstitutionally protected property right.\" ", "Defendants' Br. ", "at 8. ", "If the\nDistrict Court dismissed the substantive due process claim on that\nbasis, it was in error. ", "Woodwind in fact had a protected property interest\nfor the reasons explained above.", "\n\nIt is not entirely clear to us, however, whether the District Court\ngranted judgment as a matter of law on that ground. ", "The District Court\ninitially stated that it was \"[a]ssuming a protected property interest\"\nalthough subsequently it appears to have taken the view that Woodwind\nhad no such protected property interest. ", "App. ", "at 9-10. ", "In any event, the\nDistrict Court was in error by refusing to submit Woodwind's substantive\ndue process claim to the jury on the issue of improper motive or bad\nfaith.", "\n\n 8\n\fApplying this standard, we have not hesitated to vacate\na grant of summary judgment or a judgment as a matter of\nlaw where the evidence at least plausibly showed that the\ngovernment took actions against the developer for\nindefensible reasons unrelated to the merits of the zoning\ndispute. ", "One example is Bello v. Walker, 840 F.2d 1124 (3d\nCir. ", "1988). ", "Bello involved a substantive due process claim\nunder S 1983 brought by a developer against the municipal\ncouncil. ", "The plaintiff in Bello presented evidence that the\ndefendants \"improperly interfered with the process by\nwhich the municipality issued building permits, and that\nthey did so for partisan political or personal reasons\nunrelated to the merits of the application for the permits.\"", "\nId. at 1129. ", "The defendants in Bello presented evidence\nthat the building permits at issue were denied for legitimate\nzoning reasons thus presenting an arguably rational\nground for the decision. ", "Id. at 1130. ", "Because there was a\ngenuine factual dispute over whether the defendants had\ndenied the permit based upon an improper motive, we\nvacated the grant of summary judgment by the District\nCourt. ", "See also Deblasio v. Zoning Bd. ", "of Adjustment, 53\nF.3d 592, 600-01 (3d Cir. ", "1995) (vacating summary\njudgment where genuine dispute as to whether denial of\npermit by governmental decision maker was motivated by\nimproper personal financial reasons); Parkway Garage, Inc.\nv. City of Philadelphia, 5 F.3d 685, 696-99 (3d Cir. ", "1993)\n(vacating directed verdict where jury could reasonably infer\nthat denial of permit by governmental decision maker was\nmotivated by improper economic reasons).", "\n\nIn Blanche Road Corp. v. Bensalem Twp., ", "57 F.3d 253 (3d\nCir. ", "1995), the plaintiff presented evidence that the\ndefendants had engaged in a pattern of improperly refusing\nto release and issue certain permits as part of an effort \"to\ndelay and ultimately to shut down\" the proposed\nsubdivision development. ", "57 F.3d at 260. ", "We held that the\nintentional blocking or delaying of the issuance of permits\nfor reasons unrelated to the merits of the permit application\nviolates principles of substantive due process and is\nactionable under S 1983. ", "Id. at 268-69. ", "We vacated the\norder of the District Court granting the defendants' motion\nfor a directed verdict under Rule 50(a) and ordered a new\ntrial.", "\n\n 9\n\fThis case is similar to Bello and Blanche Road. ", "Woodwind\npresented the following evidence at trial: (1) the defendants\nhad no legitimate basis under the ordinance for demanding\ninformation about the socioeconomic background and\nincome-levels of prospective tenants as a condition of\nsubdivision approval; (2) the defendants denied approval for\nthe plan by adopting significant portions of a letter drafted\nby the private attorney for the citizens group which\nvigorously opposed the development for improper reasons;\nand (3) the defendants intentionally blocked or delayed the\nissuance of the permit for subdivision approval because\nthey were aware that by doing so the developer would be\nunable to meet the building deadline for financing the\nproject. ", "All of this in combination could provide a jury with\na basis from which it could reasonably find that the\ndecision of the defendants to deny approval was made in\nbad faith or was based upon an improper motive. ", "See, e.g.,\nBello, 840 F.2d at 1130.", "\n\nOur conclusion here is bolstered by our narrow scope of\nreview. ", "On a motion for a judgment as a matter of law, the\nnon-moving party is entitled to the benefit of all reasonable\ninferences. \" ", "`The trial judge, in his review of the evidence,\nand this court, in its own appellate review, must expose the\nevidence to the strongest light favorable to the party against\nwhom the motion is made and give him every advantage of\nevery fair and reasonable inference.' \" ", "Parkway Garage, 5\nF.3d at 698 (quoting Fireman's Fund Ins. ", "Co. v. Videofreeze\nCorp., 540 F.2d 1171, 1178 (3d Cir. ", "1976)). ", "As we stated in\nFireman's Fund,\"we cannot say (as a matter of law) that the\nrecord is deficient of that `minimum quantum of evidence\nfrom which a jury might reasonably afford relief '.\" ", "540 F.2d\nat 1178. ", "Woodwind adduced sufficient evidence to overcome\nthe motion. ", "The District Court erred in dismissing the\nS 1983 substantive due process claim. ", "Accordingly, we will\nremand for further proceedings.", "\n\nNext, the supervisor defendants contend that their Rule\n50(a) motion should be upheld on the alternative ground\nthat they are entitled to qualified immunity for their\ndecision to deny Woodwind's application for subdivision\napproval. ", "According to the supervisors, they are entitled to\nqualified immunity simply because they were relying upon\n\n 10\n\fthe recommendation of the planning commission and the\ntownship solicitor. ", "We disagree.", "\n\nThe test for determining whether government officials are\nentitled to qualified immunity for their actions, as set forth\nin Harlow v. Fitzgerald, 457 U.S. 800 (1982) is that\n\"government officials performing discretionary functions\ngenerally are shielded from liability for civil damages insofar\nas their conduct does not violate clearly established\nstatutory or constitutional rights of which a reasonable\nperson would have known.\" ", "457 U.S. at 818. ", "In the instant\ncase, however, when the evidence is viewed in the light\nmost favorable to the plaintiff, it is clear that the supervisor\ndefendants could not have reasonably believed that their\nconduct did not violate plaintiff 's rights. ", "Under the local\nordinance, the Woodwind plan as submitted must have\nbeen approved as a subdivision because it satisfied all of\nthe objective criteria. ", "Yet the supervisor defendants denied\napproval for the subdivision plan. ", "The supervisor\ndefendants have not shown that their interpretation or\nunderstanding of the ordinance was reasonable or that\nPennsylvania law on the subject was unclear. ", "Accordingly,\nthe defense of qualified immunity is not available to the\nsupervisor defendants in the instant matter. ", "See, e.g.,\nBlanche Rd. ", "Corp. v. Bensalem Twp., ", "57 F.3d 253, 269 (3d\nCir. ", "1995) (no qualified immunity for township supervisors\nwhere they could not reasonably have believed that their\nconduct did not violate plaintiff 's constitutional rights).3\n\nThe Township also contends that its Rule 50(a) motion\nshould be upheld on the alternative ground that there was\ninsufficient evidence from which a jury couldfind a\nTownship policy sanctioning conduct that violated\nplaintiff 's constitutional rights. ", "In order to establish\nTownship liability under S 1983, \"a plaintiff must show that\nan official who has the power to make policy is responsible\nfor either the affirmative proclamation of a policy or\n_________________________________________________________________\n\n3. ", "The District Court did not specifically address the issue of whether\nthe\nplanning commissioners were entitled to qualified immunity.", "\nNonetheless, we conclude that the planning commissioners are not\nentitled to qualified immunity for similar reasons as those discussed\nabove.", "\n\n 11\n\facquiescence in a well-settled custom.\" ", "Blanche Rd. ", "Corp. v.\nBensalem Twp., ", "57 F.3d 253, 269 n. 16 (3d Cir. ", "1995). ", "It\nis hornbook law that \"actions by those with final authority\nfor making a decision in the municipality constitute official\npolicy for purposes of S 1983.\" ", "Erwin Chemerinsky, Federal\nJurisdiction, S 8.5 at 479 (3d ed. ", "1999).", "\n\nApplying this test in Blanche Road, we held that there\nwas sufficient evidence of an official policy to establish\nTownship liability under S 1983 because the Township\nsupervisors had final, unreviewable authority for making\nthe decision to deny the permits sought by the plaintiff. ", "57\nF.3d at 269 n.16. ", "Similarly, here the supervisor defendants\nindisputably had final, unreviewable authority for making\nthe decision. ", "Accordingly, Woodwind's evidence, if believed,\nis sufficient to establish Township liability under S 1983.", "\n\nFor the above reasons, we will reverse the District Court's\norder, granting defendants' Rule 50(a) motion, and we will\nremand for further proceedings consistent with this\nopinion.", "\n\nA True Copy:\nTeste:\n\n Clerk of the United States Court of Appeals\n for the Third Circuit\n\n 12\n\f" ]

{ "pile_set_name": "FreeLaw" }

[ 0.005738880918220947, 0, 0, 0.01568627450980392, 0.007042253521126761, 0.013888888888888888, 0.023529411764705882, 0.03125, 0.02570093457943925, 0.03015075376884422, 0.015748031496062992, 0.016216216216216217, 0.004784688995215311, 0, 0.0056179775280898875, 0.00819672131147541, 0, 0.0048543689320388345, 0, 0.008333333333333333, 0, 0, 0, 0.006666666666666667, 0.013513513513513514, 0.011764705882352941, 0.013605442176870748, 0.0045045045045045045, 0.011764705882352941, 0.009433962264150943, 0.007462686567164179, 0.003115264797507788, 0, 0.005291005291005291, 0, 0, 0.008333333333333333, 0.012345679012345678, 0.004273504273504274, 0.01098901098901099, 0.014084507042253521, 0.01694915254237288, 0.006802721088435374, 0.010810810810810811, 0.00909090909090909, 0, 0.008928571428571428, 0.007246376811594203, 0, 0, 0, 0.009174311926605505, 0, 0.014084507042253521, 0, 0.01694915254237288, 0.015503875968992248, 0, 0, 0.006535947712418301, 0.029411764705882353, 0, 0, 0.043478260869565216, 0, 0, 0.02857142857142857, 0, 0.041666666666666664, 0, 0.013513513513513514, 0, 0, 0, 0.008547008547008548, 0, 0.041666666666666664, 0.047619047619047616, 0.03333333333333333, 0.02857142857142857, 0.03389830508474576, 0.012658227848101266, 0.02127659574468085, 0, 0, 0.08333333333333333, 0.0625, 0.004608294930875576, 0, 0, 0, 0, 0.0055248618784530384, 0.07692307692307693, 0, 0.015384615384615385, 0, 0.02127659574468085, 0, 0, 0.010416666666666666, 0, 0, 0, 0.008032128514056224, 0, 0, 0, 0, 0.04081632653061224, 0, 0.03571428571428571, 0, 0.019417475728155338, 0, 0, 0, 0.014084507042253521, 0.009259259259259259, 0, 0, 0, 0.0273972602739726, 0, 0, 0, 0.026785714285714284, 0, 0, 0.005649717514124294, 0, 0, 0.01020408163265306, 0, 0.00819672131147541, 0.0049504950495049506, 0, 0, 0.006024096385542169, 0, 0.01818181818181818, 0, 0, 0.0036101083032490976, 0, 0.005494505494505495, 0, 0, 0.03125, 0.022727272727272728, 0.008130081300813009, 0, 0.047619047619047616, 0, 0, 0.0625, 0, 0, 0.014388489208633094, 0.011904761904761904, 0, 0, 0.02857142857142857, 0, 0, 0, 0.01694915254237288, 0.01818181818181818, 0, 0.005376344086021506, 0.05555555555555555, 0, 0.012345679012345678, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.041666666666666664, 0, 0.0023584905660377358, 0, 0.007575757575757576, 0, 0, 0.08333333333333333, 0.041666666666666664, 0, 0, 0, 0.03225806451612903, 0, 0, 0, 0, 0, 0.0055248618784530384, 0.007194244604316547 ]

[ "Von Cathrin Kahlweit, London\n\nDer britische Premierminister Boris Johnson ist am Sonntagabend in ein Londoner Krankenhaus eingeliefert worden. ", "Johnson hatte sich vor zehn Tagen wegen einer Infektion mit dem Coronavirus aus der Öffentlichkeit zurückgezogen und in einer Wohnung in der Downing Street selbst isoliert. ", "Eigentlich hatte er sich nach sieben Tagen wieder selbst aus der häuslichen Quarantäne entlassen wollen, obwohl grundsätzlich vierzehn Tage angeraten werden.", "\n\nOffenbar geht es ihm aber nach wie vor schlecht und er hat weiter Fieber. ", "Auf einem Video, das Johnson vergangene Woche selbst aufgenommen hatte, um von seiner Krankheit zu berichten und gleichzeitig die Bürger aufzufordern, daheimzubleiben, hatte er mitgenommen ausgesehen. ", "Am Donnerstagabend hatte er sich kurz vor Nummer 10 Downing Street gezeigt, um am landesweiten Applaus für Mitarbeiter des Gesundheitswesens teilzunehmen, hatte sich aber danach umgehend in seine Wohnung zurückgezogen.", "\n\nAktuelles zum Coronavirus - zweimal täglich per Mail oder Push-Nachricht Alle Meldungen zur aktuellen Lage in Deutschland und weltweit sowie die wichtigsten Nachrichten des Tages - zweimal täglich mit SZ Espresso. ", "Unser Newsletter bringt Sie morgens und abends auf den neuesten Stand. ", "Kostenlose Anmeldung: sz.de/espresso. ", "In unserer Nachrichten-App (hier herunterladen) können Sie den Espresso oder Eilmeldungen auch als Push-Nachricht abonnieren.", "\n\nAuch die Freundin von Johnson, Carrie Symonds, die mit dem ersten Kind des Paares schwanger ist, hat sich mit dem Virus infiziert; sie gab das am Wochenende in den sozialen Medien bekannt. ", "Sie habe aber nur milde Symptome, sagte sie.", "\n\nJohnson soll sich auf Anraten seiner Ärzte für Untersuchungen ins Krankenhaus begeben haben; dies sei \"vorsorglich\" geschehen. ", "Er soll nach wie vor starke Krankheitssymptome wie Fieber und Husten haben. ", "Johnson soll mindestens über Nacht in der Klinik bleiben. ", "Er hält weiterhin Kontakt zum Kabinett, wie die BBC meldet, und er hatte auch noch am Samstag mit dem neuen Chef der Labour-Partei, Keir Starmer, telefoniert. ", "Der für das Wohnungswesen zuständige Staatssekretär Robert Jenrick betonte: \"Er wird weiterhin über das Geschehen informiert und in der Regierungsverantwortung bleiben.\" ", "Als Vertreter für den Fall, dass der Premier sein Amt nicht mehr ausüben kann, ist Außenminister Dominic Raab bestellt worden.", "\n\nAm Sonntag gab es im Vereinigten Königreich 48 406 gemeldete Corona-Fälle, 4943 Menschen sind bisher an dem Virus gestorben." ]

{ "pile_set_name": "OpenWebText2" }

[ 0.02097902097902098, 0.03468208092485549, 0.01910828025477707, 0, 0.014925373134328358, 0.01834862385321101, 0.009259259259259259, 0.04225352112676056, 0.02631578947368421, 0.008, 0.015706806282722512, 0.022727272727272728, 0.031007751937984496, 0, 0.05172413793103448, 0.031446540880503145, 0.023529411764705882, 0.031746031746031744, 0.015873015873015872 ]

[ "Anthaxia simiola\n\nAnthaxia simiola is a species of metallic wood-boring beetle in the family Buprestidae. ", "It is found in North America.", "\n\nReferences\n\nFurther reading\n\n \n \n \n\nCategory:Buprestidae\nCategory:Articles created by Qbugbot\nCategory:Beetles described in 1884" ]

{ "pile_set_name": "Wikipedia (en)" }

[ 0.018867924528301886, 0, 0.007692307692307693 ]

[ "package Chapter_10;\n\n/**\n * Chapter 10 Exercise 20:\n *\n * (Approximate e) Programming Exercise 5.26 approximates e using the following\"\n * \"series:\"\n * \"1 1 1 1 1\"\n * \"e = 1 + 1! ", "+ 2! ", "+ 3! ", "+ 4! ", "+ c + i!\"", "\n * \"In order to get better precision, use BigDecimal with 25 digits of precision in\"\n * \"the computation. ", "Write a program that displays the e value for i = 100, 200, . . . , ", "and 1000.\"", "\n *\n * Created by Luiz Arantes Sa on 9/7/14.", "\n */\npublic class Exercise_20 {\n}\n" ]

{ "pile_set_name": "Github" }

[ 0, 0, 0, 0, 0, 0.008547008547008548, 0, 0, 0, 0 ]

[ "Twenty-two head coaches in the FBS share the responsibility of leading the programs for which they played. ", "There's joy in the experience, immense pride and always, at some point, a dash of pain.", "\n\n\"You can't call it work,\" said Appalachian State coach Scott Satterfield, who quarterbacked the Mountaineers from 1992 to '95.", "\n\nThe highs are higher and the lows are lower when they happen at your alma mater.", "\n\n\"You don't ever escape it,\" Texas Tech's Kliff Kingsbury said. \"", "Win, lose or draw, a lot of people in your life know the score. ", "They know how it went, and they want to talk about it.\"", "\n\nSaid Wisconsin coach Paul Chryst: \"There's just a lot more layers.\"", "\n\nIn relationship-status terms, well, it's complicated.", "\n\n\"It's just different than being at another school,\" Oklahoma State coach Mike Gundy said.", "\n\nSaturday at Michigan Stadium, two of the most high-profile college players to return as coach meet in the Big Ten opener for both. ", "Nebraska, winless under Scott Frost in his first season, visits Michigan and Jim Harbaugh, known in his fourth year more for six losses in seven games against the Wolverines' chief rivals than for his 30 victories.", "\n\nJim Harbaugh has felt the pride of coaching at the alma mater but the frustration of struggling against the Wolverines' rivals. ", "Rey Del Rio/Getty Images\n\nBoth former quarterbacks achieved head-coaching success elsewhere before answering the call from home -- Harbaugh with the University of San Diego, Stanford and the San Francisco 49ers, and Frost for the past two seasons at UCF. ", "They endure intense scrutiny but said that the potential reward of restoring championship-level play at their proud programs is worth the pressure and weight of increased expectations.", "\n\n\"It's heightened,\" Harbaugh said this week when asked whether the moments mean more to him because it's Michigan.", "\n\nLast year, before UCF embarked on a 13-0 season that vaulted Frost to the top of the Group of 5 coaching hierarchy, his mother, Carol Frost -- a pioneer for female athletes and coaches in Nebraska some five decades ago -- granted a lengthy podcast interview to Omaha World-Herald writer Dirk Chatelain.", "\n\nHe asked if she could see her son someday in charge of the Cornhuskers.", "\n\n\"It would be fun to envision,\" Carol Frost said, \"if you could ensure that he would win 10 games every year. ", "Because nine is never enough.\"", "\n\nWhen Frost took the job in December after Nebraska slipped to 4-8 in 2017, its worst finish since 1961, the coach sternly asked media members to steer clear of his parents, wife and infant son. ", "No family appeared with Frost at his introductory media conference, unusual under normal circumstances.", "\n\nClearly, though, this was an unusual moment. ", "And it continues for the native son, who was QB on former coach Tom Osborne's final two teams, including the unbeaten 1997 group that shared a national title with Michigan.", "\n\nFrost is 0-2 this month after his much-hyped opener against Akron was canceled because of thunderstorms. ", "Nebraska started slowly in five-point home losses to Colorado and Troy, committing untimely penalties and special-teams errors while relying on quarterbacks in both games who were making their major-college debuts.", "\n\nBig Ten Network cameras captured Frost last week on the sideline amid the storm of three Nebraska turnovers for a second straight week, 10 flags and a 58-yard punt-return TD by Troy's Cedarius Rookard.", "\n\nThe 43-year-old Frost appeared not angry or impatient; rather, he looked a special kind of miserable.", "\n\n\"I'd say dissatisfaction,\" Frost said.", "\n\nAre those emotions magnified because he's living it at Nebraska? ", "Yes and no, said Frost, who heads to Michigan unsure whether true freshman QB Adrian Martinez is ready to return from a one-week absence after he suffered a leg injury in Week 2. ", "Walk-on Andrew Bunch played in his place against Troy.", "\n\n\"I want it to work more, and I know it's going to work,\" he said. \"", "It hasn't happened as quickly as we wanted. ", "We've had some pretty tough breaks in the first few weeks. ", "But that's the way the sport goes.\"", "\n\nIn the next breath, Frost said, he wouldn't feel different in this spot as coach of the Green Bay Packers or at Malcolm (Nebraska) High. ", "That's the school from which Osborne -- then an assistant coach under Bob Devaney -- recruited Frost's father, Larry, to Nebraska as a halfback in 1966.", "\n\nScott Frost's roots run deep in Lincoln after he and his father were both recruited there by former coach Tom Osborne. ", "Stephen Dunn/Getty Images\n\nYou see, part of what complicates matters so much for the likes of Frost and Harbaugh is just how deep the roots extend.", "\n\nDevaney, the figurative godfather of Nebraska coaches, arrived at Nebraska from Wyoming in 1962 and coached his first road game as a big underdog at Michigan. ", "The Huskers won 25-13. ", "Five weeks later began the school's NCAA-record sellout streak, now at 363 games.", "\n\nA decade after the landmark win at Michigan, Devaney had won back-to-back national championships. ", "Nebraska's fandom, of course, salivates over the parallels, but Frost isn't making plans for a parade.", "\n\n\"This will get worse before it gets better,\" he said Saturday in the aftermath of the defeat to Troy, \"but it's always darkest before the dawn.\"", "\n\nSuch realism is essential. ", "Ask Mark Richt at Miami, Kirby Smart at Georgia and David Shaw at Stanford.", "\n\nNot every day is rosy.", "\n\n\"It means so much,\" said Northwestern coach Pat Fitzgerald, an All-American linebacker at the school in 1995 and '96. \"", "But sometimes everybody thinks you know who they are. ", "So building relationships is equally, if not more, important.\"", "\n\nIt's personal, said Central Michigan's John Bonamego, 21-21 in his fourth season after an 0-3 start this fall.", "\n\n\"This is not just a job for me,\" he said. \"", "This is a passion of mine.\"", "\n\nSo it goes for Harbaugh, who started for 2½ seasons at Michigan and led the Wolverines to a share of the 1986 Big Ten title and a Fiesta Bowl win against Nebraska the season prior.", "\n\nLike Frost, his ties to the program run deep. ", "Harbaugh's father, Jack, coached at Michigan from 1973 to '79. ", "Jim began high school in Ann Arbor. ", "And of course, like Frost, Harbaugh, 54, emulated his college coach, Bo Schembechler.", "\n\nLegendary Michigan coach Bo Schembechler was a mentor and is still an inspiration to Jim Harbaugh. ", "AP Photo/Reed Saxon\n\n\"It's been a great thrill,\" Harbaugh said this week of coaching at Michigan, \"one of the great thrills of my life to be here. ", "Love it. ", "I love coaching. ", "I love football. ", "I love coaching at my alma mater. [", "It's] been wonderful.\"", "\n\nAccording to Harbaugh's former Michigan teammate, John Ghindia, reality guides his coaching principles and helps Harbaugh navigate complicated matters.", "\n\n\"If he was unrealistic, it would be more frustrating,\" Ghindia said. \"", "Fans seem to say, 'Well, these teams win in three years.' ", "Hey, if [left tackle Grant Newsome] doesn't get hurt in 2016, they win three more games that year. ", "It made that big of a difference.", "\n\n\"When you're Scott Frost or Jim Harbaugh, you've got to wear blinders. ", "You can't listen to the outside noise.\"", "\n\nYou'll hear no such public talk from Harbaugh, whose team rose to No. ", "3 in the College Football Playoff rankings in 2016 before dropping three of its last four games. ", "At this point, the retort would likely focus on his marks against Ohio State (0-3), Michigan State (1-2) and Notre Dame, 0-1 after a Sept. 1 loss in South Bend.", "\n\nPerhaps in part because he knows of the scrutiny over his every word, Harbaugh rarely offers much insight his emotions over coaching at Michigan.", "\n\n\"He knows when you win at Michigan,\" Ghindia said, \"there's no better feeling. ", "And as far as he's concerned, he's getting closer every day.\"", "\n\nEvery day presents another opportunity for the light to shine brighter. ", "And to burn more deeply." ]

{ "pile_set_name": "OpenWebText2" }

[ 0.009345794392523364, 0, 0.0234375, 0, 0.030303030303030304, 0, 0, 0.014492753623188406, 0, 0.01098901098901099, 0, 0.014018691588785047, 0.015384615384615385, 0.0196078431372549, 0, 0, 0.01644736842105263, 0.0136986301369863, 0.009009009009009009, 0, 0.00510204081632653, 0.009708737864077669, 0, 0.011627906976744186, 0.018691588785046728, 0.004672897196261682, 0.019704433497536946, 0.009708737864077669, 0.025, 0, 0.0111731843575419, 0.018518518518518517, 0, 0, 0, 0, 0.014388489208633094, 0.03289473684210526, 0.01652892561983471, 0.013605442176870748, 0.012422360248447204, 0, 0.012345679012345678, 0.01, 0.00980392156862745, 0, 0, 0.04, 0, 0.01652892561983471, 0, 0, 0.017857142857142856, 0, 0, 0, 0.020833333333333332, 0.031746031746031744, 0.027777777777777776, 0.023529411764705882, 0.0297029702970297, 0.006802721088435374, 0, 0, 0, 0, 0, 0.013071895424836602, 0.013888888888888888, 0, 0.010101010101010102, 0, 0.0273972602739726, 0, 0, 0.010309278350515464, 0.0125, 0.013605442176870748, 0.012345679012345678, 0, 0, 0 ]

[ "Q:\n\nIn Islam, is sexual intercourse compulsory on the wedding night and does the nikah become void if it does not happen?", "\n\nIn islam if two individuals are getting married by choice but live in different countries and are very shy by personality and both decide that they just want to spend time with each other and get more comfortable before such an intimate thing takes place would this mean nikah is void? ", "Also what if both the man and women decide to consumate the marriage months or even a year after the nikah? ", "Would this break the nikah? ", "As both individuals are very shy and are kind of comfortable with each other but hardly speak for each other and so would just want to spend time with each other before anything this intimate happens.", "\nEDIT: I asked two questions not one.", "\n\nIf sex does not happen on the first night, is the nikah void?", "\nIf the couple both the man and woman decide to allow it to happen naturally meaning in a couple of months or even a year would this make the nikah valid? ", "\n\nA:\n\nThere is no rule that sexual intercourse should happen in first night. ", "If the woman is having her periods then sexual intercourse should not happen on first night. ", "There can be other reasons which prevent them, so it is up to the couple to decide on this.", "\nIn Islam there is nothing that say sexual intercourse between couples should happen at any particular time. ", "But only rules on when not to have sexual intercourse, like during day time of ramadan, during periods etc. ", "So couples can decide any time which is not prohibited times.", "\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0.008264462809917356, 0.003472222222222222, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.009174311926605505, 0, 0, 0 ]

[ " NOT RECOMMENDED FOR FULL-TEXT PUBLICATION\n File Name: 11a0066n.06\n\n No. ", "08-6136\n FILED\n UNITED STATES COURT OF APPEALS Feb 01, 2011\n FOR THE SIXTH CIRCUIT LEONARD GREEN, Clerk\n\n\nUNITED STATES OF AMERICA, )\n )\n Plaintiff-Appellee, ) ON APPEAL FROM THE\n ) UNITED STATES DISTRICT\nv. ) COURT FOR THE MIDDLE\n ) DISTRICT OF TENNESSEE\nPIERRE M. RODRIGUEZ, )\n ) OPINION\n Defendant-Appellant. ", " )\n )\n )\n\nBEFORE: MOORE and STRANCH, Circuit Judges; COHN, District Judge.*", "\n\n COHN, District Judge. ", "This is a criminal case. ", "Defendant-Appellant Pierre M. Rodriguez\n\n(“Rodriguez”) appeals his conviction and sentence of 385 months after a jury found him guilty of\n\ntwo counts of conspiracy to commit robbery, two counts of robbery in violation of 18 U.S.C. § 1951,\n\nand two counts of brandishing a firearm during and in relation to a crime of violence in violation of\n\n18 U.S.C. § 924(c). ", "Rodriguez argues that the district court erred by prohibiting his lawyer from\n\ncross-examining a government witness about his prior theft conviction and allowing fingerprint\n\nevidence in the absence of the original source of the fingerprints. ", "He also argues that there was\n\ninsufficient evidence to support his conviction. ", "For the reasons set forth below, Rodriguez’s\n\nconviction will be affirmed.", "\n\n\n\n\n *\n The Honorable Avern Cohn, United States District Judge for the Eastern District of\nMichigan, sitting by designation.", "\n\fUnited States v. Rodriguez\nNo. ", "086136\n\n I. BACKGROUND\n\n Rodriguez’s case involves two robberies. ", "The first robbery occurred at Las Americas\n\nMusicas, a music store, and the second robbery occurred at Russell’s Market, a convenience store.", "\n\n The Las Americas Musicas robbery occurred on the evening of February 14, 2006. ", "Two men\n\nentered the store carrying guns and demanded that the pregnant clerk on duty, Maria Fragosa\n\n(“Fragosa”), give them the money in the register; Fragosa complied.", "\n\n The Russell’s Market robbery took place on March 23, 2006. ", "Two men entered the store in\n\nthe early morning hours. ", "One man pointed a handgun at the store owner, Russell Reed (“Reed”), and\n\ndemanded he turn over the money in the cash register; Reed complied. ", "The two men left the store\n\nat which time Reed grabbed a gun, went outside, and saw the two men leaving the parking lot in a\n\nblack pickup truck. ", "Reed fired the gun at, but did not hit, the truck.", "\n\n Rodriguez was arrested less than two months later after being implicated by another robbery\n\nsuspect. ", "After the arrest, the lead investigator on the case, Charlie Wilder (“Wilder”), interrogated\n\nRodriguez. ", "Wilder asked Rodriguez questions in English. ", "Rodriguez responded to Wilder’s\n\nquestions in Spanish, Rodriguez’s native language. ", "After the interrogation, Wilder transcribed\n\nRodriguez’s confession and Rodriguez signed it.1 Rodriguez was later identified by Reed during a\n\npreliminary court appearance, which Rodriguez asserts was orchestrated by Wilder.", "\n\n In May, 2008, Rodriguez went to trial on a six-count indictment charging him with the\n\nRussell’s Market and Las Americas Musicas robberies, conspiracy to commit both robberies, and\n\n\n 1\n Rodriguez does not challenge his confession on appeal; thus, it is not addressed in this\nopinion.", "\n\n -2-\n\fUnited States v. Rodriguez\nNo. ", "086136\n\nbrandishing a firearm during both robberies. ", "Prior to the start of trial, the government moved to\n\nexclude Rodriguez’s planned cross-examination of Wilder about a past misdemeanor theft\n\nconviction, for the purpose of impeaching Wilder’s credibility. ", "The government objected on the\n\nground that the prior conviction was not admissible under Fed. ", "R. Evid. ", "609. ", "Rodriguez responded\n\nthat, while the conviction was “generally inadmissible” because it was over ten years old, the court\n\nshould allow the cross-examination because “in the interest of justice [sic] the probative value of the\n\nconviction [was] supported by specific facts and circumstances substantially outweigh[ing] its\n\nprejudicial effect.” ", "The government relied on a decision from this Court, United States v. Rattigan,\n\nNo. ", "91-00106, 1993 WL 190910, *5 (6th Cir. ", "June 2, 1993), to argue that theft is not a crime of\n\ndishonesty. ", "The district court judge agreed with the government and granted the motion, stating, “if\n\nthe Sixth Circuit finds that robbery is not a crime of dishonesty, or false statement, how can I\n\nconclude that it’s in the interest of justice?” ", "The district court judge further stated,\n\n under Rule 609(a), I don’t think it’s a felony; it’s a misdemeanor. ", "I think it is beyond\n the time limits. ", "I also find that, given that the Ra[tt]igan case finds that the crime of\n robbery is not a crime of dishonesty or false statement, that the motion in limine to\n exclude this reference should be granted.", "\n\n Several witnesses testified at Rodriguez’s trial. ", "Reed identified Rodriguez as one of the two\n\nmen who committed the Russell’s Market robbery and testified that Rodriguez stole money from him\n\nat gunpoint. ", "Two sheriff’s deputies testified that Rodriguez confessed to the Russell’s Market\n\nrobbery. ", "The deputies said Rodriguez’s confession included details related to the brandishing of the\n\ngun, the other individuals involved, and the get-away cars used during the commission of the crime.", "\n\nFragosa testified that two men in ski masks robbed her at gunpoint, and the second robber, Carlos\n\n\n -3-\n\fUnited States v. Rodriguez\nNo. ", "086136\n\nHernandez (“Hernandez”), testified that he and Rodriguez committed the robbery together with\n\nassistance from two other individuals who waited in a getaway car.", "\n\n An officer on the scene of the Las Americas Musicas robbery testified that he viewed a\n\nsurveillance video at the store, which depicted one of the robbers picking up a CD during the\n\nrobbery, but not taking it. ", "After seeing the video, the officer testified that he directed a crime scene\n\ndetective to the area where the CD was located. ", "The crime scene detective testified that he was able\n\nto lift a fingerprint from the CD, which he placed onto a fingerprint card, and which was later entered\n\ninto evidence by the government. ", "A fingerprint expert testified that she compared the fingerprint\n\ncard to Rodriguez’s fingerprints and the two fingerprints matched.", "\n\n Prior to the fingerprint expert’s testimony, Rodriguez objected on chain of custody grounds\n\nthat the government failed to link the CD from which the fingerprints were lifted to the exact CD\n\ntouched by one of the robbers. ", "The district court overruled Rodriguez’s objection, concluding that\n\n“there ha[d] been a sufficien[t] showing of the chain of custody.”", "\n\n At the close of the government’s case-in-chief, Rodriguez moved for a judgment of acquittal,\n\nwhich was denied. ", "As previously stated, at the end of trial, the jury found Rodriguez guilty on two\n\ncounts of conspiracy to commit robbery, two counts of robbery in violation of 18 U.S.C. § 1951, and\n\ntwo counts of brandishing a firearm during and in relation to a crime of violence in violation of 18\n\nU.S.C. § 924(c). ", "Rodriguez did not renew the Rule 29 motion after the verdict was read. ", "Rodriguez\n\nwas sentenced to a term of imprisonment of 385 months.", "\n\n\n\n\n -4-\n\fUnited States v. Rodriguez\nNo. ", "086136\n\n II. ", "DISCUSSION\n\n A. Wilders’ Prior Theft Conviction\n\n Rodriguez challenges the district court’s refusal to allow admission of evidence on or cross\n\nexamination of Detective Wilder regarding a 1992 misdemeanor theft conviction: theft of two horses\n\nvalued at less than $500. ", "Rodriguez contends that the district court erred in denying him the\n\nopportunity to impeach Wilder’s credibility under Fed. ", "R. Evid. ", "609 and pursuant to his Sixth\n\nAmendment right of confrontation. ", "Rodriguez’s claims are without merit.", "\n\n Because Wilder’s theft conviction was a misdemeanor, Rule 609(a)(2) applies to its\n\nadmission, as does the staleness provision of Rule 609(b). ", "A district court’s evidentiary rulings are\n\nordinarily reviewed by this Court for an abuse of discretion. ", "United States v. Johnson, 581 F.3d 320,\n\n332 (6th Cir. ", "2009). ", "This standard applies to a district court’s Rule 609(b) decisions. ", "See, e.g.,\n\nUnited States v. Franco, 484 F.3d 347, 353 (6th Cir. ", "2007).", "\n\n Fed. ", "R. Evid. ", "609(b) limits, for the purpose of impeachment, the admissibility of prior\n\nconviction evidence “more than 10 years old.” ", "Fed. ", "R. Evid. ", "609(b). ", "That is, “Rule 609(b) creates,\n\nin effect, a rebuttable presumption that convictions over ten years old are more prejudicial than\n\nhelpful and should be excluded.” ", "United States v. Sims, 588 F.2d 1145, 1150 (6th Cir. ", "1978).", "\n\nParticularly, after reviewing the legislative history of Rule 609(b), this Court has held that “evidence\n\nof convictions more than ten years old should be admitted very rarely and only in exceptional and\n\ncircumstances.” ", "United States v. Sloman, 909 F.2d 176, 181 (6th Cir. ", "1990) (internal quotations\n\nomitted).", "\n\n\n\n -5-\n\fUnited States v. Rodriguez\nNo. ", "086136\n\n In this case, the district court found that Wilder’s 14-year-old conviction was inadmissible\n\nfor staleness under Rule 609(b). ", "Thus, even if we assume that the theft evidence was admissible\n\nunder Rule 609(a)(2) as a crime involving an element of dishonesty or false statement,2 the district\n\ncourt did not abuse its discretion by excluding the conviction under Rule 609(b).", "\n\n Rodriguez also challenges the exclusion of Wilder’s theft conviction as a violation of his\n\nSixth Amendment right to confront his accusers. ", "However, Rodriguez did not make a Confrontation\n\nClause objection to the district court’s evidentiary rulings at the trial level, and thus we review his\n\nargument now for plain error. ", "See United States v. Deitz, 577 F.3d 672, 683 (6th Cir. ", "2009).", "\n\n “[T]he Confrontation Clause affords the right to impeach a witness with his criminal record,\n\nsubject to the trial court’s discretion to impose reasonable limitations to prevent harassment and\n\nannoyance of the witness.” ", "Vasquez v. Jones, 486 F.3d 135, 143 (6th Cir. ", "2007) (discussing\n\nDelaware v. Van Arsdall, 475 U.S. 673 (1986) and Davis v. Alaska, 415 U.S. 308 (1974)). ", "On\n\nreview, “[t]he key issue is whether the jury had enough information to assess the defense’s theory\n\nof the case despite the limits placed on cross-examination.” ", "United States v. Holden, 557 F.3d 698,\n\n704 (6th Cir. ", "2009) (citing Boggs v. Collins, 26 F.3d 728, 739 (6th Cir. ", "2000)).", "\n\n\n\n 2\n Fed. ", "R. Evid. ", "609(a)(2) determinations as to whether a specific conviction constitutes a\ncrime of dishonesty or false statement are reviewed de novo because Rule 609(a)(2) removes\nfrom the district court any discretion in admitting evidence of a prior conviction with an element\nof dishonesty or false statement. ", "See United States v. Morrow, 977 F.2d 222, 228 (6th Cir.", "\n1992) (en banc)). ", "Instead, under the rule, evidence that a witness has been convicted of a crime\nwith an element of dishonesty or false statement “shall be admitted” without any balancing of\nprejudice or other factors. ", "Doe v. Sullivan County, 956 F.2d 545, 551 (6th Cir. ", "1992); United\nStates v. Peatross, 377 F. App’x 477, 489 (6th Cir. ", "2010). ", "It is unnecessary to address this issue\nbecause the evidence was properly excluded under Rule 609(b).", "\n\n -6-\n\fUnited States v. Rodriguez\nNo. ", "086136\n\n We must be mindful to view this standard through the lens of plain error. ", "In particular, in\n\norder to show plain error, the district court’s error must have, inter alia, “affected the appellant’s\n\nsubstantial rights, i.e., affected the outcome of the district court’s proceedings.” ", "Puckett v. United\n\nStates, 192 S. Ct. ", "1423, 1425 (2009) (internal citations omitted). ", "Given the strength of the\n\nprosecution’s evidence with regard to the Russell’s Market robbery, in particular Rodriguez’s\n\nconfession and Mr. Reed’s positive identification of Rodriguez as the assailant, it is difficult to\n\nimagine how Rodriguez’s ability to cross-examine Wilder on a 14-year-old misdemeanor theft\n\nconviction would have changed the outcome of the trial. ", "Thus, we find no violation of Rodriguez’s\n\nConfrontation Clause right that rises to a level of plain error.", "\n\n For these reasons, we find no error in the district court’s exclusion of Wilder’s conviction\n\nevidence for impeachment purposes.", "\n\n B. Fingerprint Evidence\n\n Rodriguez next asserts that the district court abused its discretion by allowing fingerprint\n\ntestimony on the ground that the government failed to produce the original source of the fingerprints,\n\na CD, from which the fingerprints were lifted. ", "Instead, the fingerprint expert testified based on a\n\nfingerprint card, which Rodriguez says broke the chain of custody and tainted the admissibility of\n\nthe fingerprint evidence. ", "Rodriguez’s argument is without merit.", "\n\n A district court's evidentiary decisions are reviewed for an abuse of discretion. ", "Rush v. Ill.\n\nCent. ", "R.R. Co., 399 F.3d 705, 715 (6th Cir. ", "2005). ", "Particularly, fingerprint evidence is admissible\n\nwhen the possibilities of misidentification or alteration are “eliminated, not absolutely, but as a\n\nmatter of reasonable probability.” ", "United States v. McFadden, 458 F.2d 440, 441 (6th Cir. ", "1972)\n\n -7-\n\fUnited States v. Rodriguez\nNo. ", "086136\n\n(internal citation omitted). ", "Absent a clear abuse of discretion, challenges to the chain of custody do\n\nnot go to the admissibility of the evidence. ", "United States v. Levy, 904 F.2d 1026, 1030 (6th Cir.", "\n\n1990). ", "Rather, where direct testimony establishes that the defendant handled the item from which\n\nfingerprints were obtained, whether the government has proved an adequate chain of custody goes\n\nto the weight of the evidence. ", "United States v. Crite, No. ", "92-6015, 1993 WL 445084, *6 (6th Cir.", "\n\nNov. 2, 1993) (citing McFadden, supra). ", "In other words, “the verdict depend[s] on whether the jury\n\nbelieve[s] the United States or the defendant.” ", "Id. “This matter of credibility is for the jury, not the\n\nappellate court.” ", "Id. (internal citations omitted).", "\n\n Here, direct testimony reasonably established that one of the robbers touched a CD and that\n\nthe fingerprints lifted from a CD matched Rodriguez’s fingerprints. ", "Further, an adequate chain of\n\ncustody was established in that fingerprints were lifted from a CD at the crime scene onto a\n\nfingerprint card, which was entered into evidence and evaluated by a fingerprint expert, who\n\nconcluded that the fingerprint on the card matched that of Rodriguez. ", "Thus, Rodriguez has not\n\nproffered evidence to show a clear abuse of discretion and the district court did not err in finding that\n\nRodriguez’s challenge went to the weight of the evidence and not admissibility. ", "In other words, it\n\nwas up to the jury to decide whether it believed the government or Rodriguez, not this Court.", "\n\n Rodriguez also challenges the fingerprint evidence on lack of foundation grounds. ", "This\n\nclaim is without merit. ", "Under Fed. ", "R. Evid. ", "703, “[t]he facts or data in the particular case upon\n\nwhich an expert bases an opinion or inference may be those perceived by or made known to the\n\nexpert at or before the hearing.” ", "Further, under Fed. ", "R. Evid. ", "901(b)(1), evidence is properly\n\nauthenticated when a witness testifies that “a matter is what it is claimed to be.”", "\n\n -8-\n\fUnited States v. Rodriguez\nNo. ", "086136\n\n As stated above, the fingerprint expert’s testimony was based on the fingerprint card, which\n\nwas given to her prior to trial, and which contained fingerprints lifted from the crime scene. ", "Thus,\n\nthe fingerprint card was properly relied on and authenticated under Fed. ", "R. Evid. ", "703 and Fed. ", "R.\n\nEvid. ", "901. ", "Accordingly, the district court did not err by admitting the fingerprint evidence.", "\n\n C. Sufficiency of the Evidence\n\n Finally, Rodriguez argues that the evidence is insufficient to support his conviction on the\n\nground that there was no physical evidence tying him to the robberies. ", "Rodriguez also asserts that\n\nthe victim identifications were insufficient to establish his guilt. ", "Rodriguez is incorrect.", "\n\n “This Court reverses a judgment for insufficiency of the evidence only if [the] judgment is\n\nnot supported by substantial and competent evidence upon the record as a whole.” ", "United States v.\n\nBarnett, 398 F.3d 516, 521-22 (6th Cir. ", "2005) (quoting United States v. Stone, 748 F.2d 361, 363\n\n(6th Cir.1984)) (internal quotation marks omitted). “", "However, when the defendant moves for\n\njudgment of acquittal at the close of the government's case-in-chief, and defense evidence is\n\nthereafter presented but the defendant fails to renew the motion at the close of all of the evidence,\n\nhe waives objection to the denial of his earlier motion, absent a showing of a manifest miscarriage\n\nof justice.” ", "United States v. Price, 134 F.3d 340, 350 (6th Cir. ", "1998); see also United States v.\n\nRoberge, 565 F.3d 1005, 1008 (6th Cir. ", "2009). “", "A miscarriage of justice exists only if the record\n\nis devoid of evidence pointing to guilt.” ", "Id.\n\n Here, the Russell’s Market robbery victim, Reed, identified Rodriguez as one of the two men\n\nthat entered his store and robbed him at gunpoint. ", "The Las Americas Musicas robbery victim,\n\nFragosa, testified that two men wearing ski masks robbed her at gunpoint, and Hernandez testified\n\n -9-\n\fUnited States v. Rodriguez\nNo. ", "086136\n\nthat he and Rodriguez committed the robbery together. ", "Finally, Rodriguez’s fingerprints matched\n\nfingerprints lifted from the Las Americas Musicas crime scene.", "\n\n Based on the described evidence, Rodriguez has not shown a manifest miscarriage of justice.", "\n\n III. ", "CONCLUSION\n\n For all of the above reasons, we AFFIRM Rodriguez’s conviction.", "\n\n\n\n\n - 10 -\n\f" ]

{ "pile_set_name": "FreeLaw" }

[ 0, 0.007322175732217573, 0.018604651162790697, 0, 0, 0.0027548209366391185, 0, 0, 0.013513513513513514, 0, 0, 0, 0.0070921985815602835, 0.011363636363636364, 0.011834319526627219, 0, 0, 0.006993006993006993, 0, 0, 0, 0.009523809523809525, 0, 0.023809523809523808, 0.017857142857142856, 0, 0, 0, 0.014563106796116505, 0.010526315789473684, 0.1111111111111111, 0, 0, 0.023529411764705882, 0, 0, 0.00423728813559322, 0.008547008547008548, 0, 0.004672897196261682, 0.01694915254237288, 0, 0.010869565217391304, 0.005208333333333333, 0.0053475935828877, 0, 0.004545454545454545, 0, 0, 0.007575757575757576, 0, 0.007407407407407408, 0.008264462809917356, 0, 0, 0, 0, 0, 0.0032679738562091504, 0.016129032258064516, 0.1111111111111111, 0, 0, 0.006578947368421052, 0, 0.01818181818181818, 0, 0.014925373134328358, 0, 0, 0.07142857142857142, 0.1111111111111111, 0, 0.2, 0.1111111111111111, 0, 0, 0, 0, 0.008968609865470852, 0.018867924528301886, 0, 0, 0.014084507042253521, 0.004048582995951417, 0.006711409395973154, 0.005434782608695652, 0.017857142857142856, 0, 0, 0.021739130434782608, 0.018691588785046728, 0, 0.018518518518518517, 0.03389830508474576, 0, 0.03571428571428571, 0.1111111111111111, 0.0033444816053511705, 0, 0, 0, 0, 0.030303030303030304, 0, 0.009900990099009901, 0, 0, 0, 0.02631578947368421, 0, 0.008086253369272238, 0.009345794392523364, 0.0072992700729927005, 0, 0.005555555555555556, 0, 0, 0.05, 0.02631578947368421, 0, 0, 0, 0, 0, 0, 0.019230769230769232, 0, 0, 0, 0, 0.023809523809523808, 0, 0, 0, 0.005847953216374269, 0.0034602076124567475, 0.0047169811320754715, 0.017699115044247787, 0, 0, 0.09090909090909091, 0.1111111111111111, 0, 0.05, 0.1111111111111111, 0, 0, 0, 0.0125, 0.1111111111111111, 0.07692307692307693, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.01282051282051282, 0.008888888888888889, 0, 0.01904761904761905, 0.01, 0, 0, 0 ]

[ "Neolepetopsidae\n\nNeolepetopsidae is a family of small deep sea sea snails or true limpets, marine gastropod mollusks in the clade Patellogastropoda (according to the taxonomy of the Gastropoda by Bouchet & Rocroi, 2005).", "\n\nThis family has no subfamilies.", "\n\nEtymology \nThe name of the family Neolepetopsidae is composed of the prefix neo, which means \"new\", and the word Lepetopsidae, which is the name of an extinct family of true limpets, from which the species within Neolepetopsidae probably evolved.", "\n\nTaxonomy \nTwo Neolepetosidae species Eulepetopsis vitrea and Paralepetopsis floridensis were genetically analyzed by Harasewych & McArthur (2000), who confirmed placement of Neolepetopsidae within Acmaeoidea/Lottioidea based on analysis of partial 18S rDNA.", "\n\nNeolepetopsidae belongs to superfamily Neolepetopsoidea according to the taxonomy of the Gastropoda by Bouchet & Rocroi, 2005).", "\n\nNeolepetopsoidea was synonymized with Lottioidea so Neolepetopsidae was moved to superfamily Lottioidea in World Register of Marine Species.", "\n\nDistribution \nThe distribution of the Neolepetopsidae includes the Northeastern Pacific, the western Pacific (Paralepetopsis rosemariae) and the Mid-Atlantic Ridge.", "\n\nHabitat \nThese limpets live in the deep sea. ", "Their habitat includes hydrothermal vents, whalebone (baleen) and whale-fall habitats.", "\n\nGenera \nThere are currently known 3 genera and altogether 12 species in Neolepetopsidae:\n Neolepetopsis McLean, 1990 - type genus\n Neolepetopsis gordensis McLean, 1990 - type species\n Neolepetopsis densata McLean, 1990\n Neolepetopsis nicolasensis McLean, 2008 - from whalebone\n Neolepetopsis verruca McLean, 1990\n Neolepetopsis occulta McLean, 1990\n Eulepetopsis McLean, 1990 - monotypic genus\n Eulepetopsis vitrea McLean, 1990\n Paralepetopsis McLean, 1990\n Paralepetopsis floridensis McLean 1990 - type species\n Paralepetopsis clementensis McLean, 2008 - from whalebone\n Paralepetopsis ferrugivora Warén & Bouchet, 2001\n Paralepetopsis lepichoni Warén & Bouchet, 2001 - from Nankai Trench\n Paralepetopsis rosemariae Beck, 1996\n Paralepetopsis sasakii Warén & Bouchet, 2009\n Paralepetopsis tunnicliffae McLean, 2008 - from hydrothermal vent\n\nReferences\n\nFurther reading \n Sasaki T. & Warén A. (2007). \"", "Anatomy of Eulepetopsis vitrea McLean, 1990 (Patellogastropoda: Neolepetopsidae)\". ", "In: Jordaens, K., N. Van Houtte, J. Van Geothem, & T. Backeljau, editors. ", "Abstracts, World congress of malacology. ", "Antwerp, Belgium. ", "July 15–20, 2007195.", "\n\nCategory:Neolepetopsidae" ]

{ "pile_set_name": "Wikipedia (en)" }

[ 0.013574660633484163, 0, 0.008064516129032258, 0.011583011583011582, 0.007751937984496124, 0.028169014084507043, 0.006024096385542169, 0, 0, 0.012168141592920354, 0, 0.02702702702702703, 0, 0, 0, 0 ]

[ "Halo Effects in Trait Assessment Depend on Information Valence: Why Being Honest Makes You Industrious, but Lying Does Not Make You Lazy.", "\nWe propose stronger halo effects in trait assessments from positive information relative to negative information. ", "Due to positive information's higher similarity, positive information should foster both indirect (from a global impression to traits) and direct halo effects (from traits to traits). ", "Negative information's relative distinctiveness should foster only direct halo effects, leading to weaker halo effects overall. ", "Four experiments support these predictions using agency traits and communion traits and behaviors. ", "Further supporting the predictions, halo effects from positive information were visible both within (i.e., from communion/agency information to communion/agency traits) and across (i.e., from agency/communion information to communion/agency traits) these fundamental dimensions of social perception. ", "Halo effects from negative information were visible only within dimensions. ", "The study thereby explains why halo effects from negative information are usually weaker; it supports different processes underlying halo effects; and it provides a case in person perception where positive information has more impact than negative information." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0.0072992700729927005, 0, 0, 0, 0, 0, 0.013157894736842105, 0 ]

[ "How Your Address Affects Your Car Insurance Premiums\n\nDid you know that where you live affects your car insurance premiums? ", "This fact emphasizes the need for accuracy when providing your personal information for a quote as well as the importance of updating your information if you happen to move. ", "Here are some of the reasons why your address matters to car insurance providers.", "\n\nLocal Traffic\n\nWhere you live gives car insurance companies an idea of what your commute looks like. ", "If you live in an area known for heavy traffic or frequent accidents (perhaps there’s a particularly dangerous intersection just a few blocks from your house), then your car insurance rates could be higher than someone who lives in the country or an area where there is very little traffic. ", "Insurance companies periodically look at the number of accidents reported by police in your area as well. ", "If the number of reported accidents has gone up or down, your car insurance premiums may be raised or lowered to reflect this.", "\n\nIf your commute to work each day takes longer than 15 minutes, that can raise your car insurance premium as well. ", "The thought behind this is that the longer you are on the road, the more risk you are exposed to, which leads to higher rates.", "\n\nCrime Rates\n\nCrime in your neighborhood also plays a role in how much you pay in car insurance premiums. ", "Again, your address is used to get crime reports from local police forces so that the insurance company can assess the risk. ", "If there are a higher-than-average car break-ins, thefts, or vandalism in your area, you could be paying more than someone who lives in a more secure neighborhood. ", "You can see a full report on the top stolen cars in the United States on the National Insurance Crime Bureau’s (NICB) website.", "\n\nThe Bottom Line\n\nCar insurance companies are all about assessing risk. ", "The best way for you to lower your car insurance premiums is to shop around and see what each company has to offer. ", "Different car insurance companies may not assign the same level of risk to your neighborhood as others and could offer you a better rate.", "\n\nLearn More About Auto Insurance\n\nWhen it comes to auto insurance, there are a lot of variables, a lot of unfamiliar terms, and a whole lot of peculiarities – particularly when you consider that each state has its own insurance market. ", "Our Auto Insurance Explainers can help sort it out for you, and make more educated decisions about your car insurance." ]

{ "pile_set_name": "Pile-CC" }

[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.007936507936507936, 0, 0, 0, 0, 0 ]

[ "The invention relates to a method of manufacturing a semiconductor device comprising a MOS transistor, for brevity hereinafter generally referred to as MOST (=Metal Oxide Semiconductor Transistor), in which method a gate oxide and a gate electrode are formed on a surface of a semiconductor body of silicon, a source region and a drain region being formed in the semiconductor body on either side of the gate electrode, a part of said source and drain regions bordering on the edge of the gate electrode, for brevity hereinafter generally referred to as LDD (=Lightly Doped Drain) region, being provided with a lower doping concentration, and on either side of the gate electrode, the surface of the semiconductor body being provided with a dielectric layer, the part of which bordering on the surface of the semiconductor body being obtained by thermal oxidation of the semiconductor body, and the more heavily doped parts of the source region and of the drain region being formed by providing a spacer on either side of the gate electrode and subsequently doping the semiconductor body with suitable doping atoms by means of an ion implantation. ", "In practice, generally both the source region and the drain region are provided with an LDD region.", "\nSuch a method is used, in particular, in the manufacture of ICs (=Integrated Circuits) wherein MOS, CMOS or BICMOS circuits are incorporated. ", "The MOS transistor serves, for example, as a switch but may alternatively be embodied so as to be a memory element. ", "The LDD regions preclude, or at least limit, the development of hot charge carriers in that they limit the size of the maximum lateral electric field. ", "The importance hereof increases steadily as the dimensions of the MOS transistors decrease continually.", "\nA method of the type mentioned in the opening paragraph is known from United States patent specification U.S. Pat. ", "No. ", "5,702,986, published on Dec. 30, 1997. ", "The manufacture of a MOST in a silicon substrate is described by means of FIG. ", "1 of said patent specification, in which a more heavily doped part of the source and drain regions is formed by doping the substrate with doping atoms by means of an ion implantation after the gate electrode formed is provided with a spacer on both sides. ", "Prior to this process step, the surface of the semiconductor body is provided, on either side of the gate electrode formed, with a dielectric layer, in this case a silicon dioxide layer, whose formation is partly due to reoxidizing the semiconductor surface after the formation of the gate electrode, and the silicon substrate is doped with doping atoms on either side of the gate electrode by means of an ion implantation at a lower implantation energy and/or implantation flux, thereby forming the LDD regions. ", "After both implantations, a temper step takes place in which, at the location of both the more heavily and the more lightly doped parts of the source and drain regions, the crystal damage in the silicon substrate is repaired, and in which step the doping atoms are rendered electrically active.", "\nA drawback of the known method is that the MOST manufactured thereby still suffers from the above-mentioned \"short-channel\" effects, which manifest themselves, inter alia, in that the threshold voltage of the MOST manufactured decreases substantially for very short lengths of the gate electrode, which is undesirable." ]

{ "pile_set_name": "USPTO Backgrounds" }

[ 0.0017421602787456446, 0, 0.013986013986013986, 0.008620689655172414, 0, 0.009708737864077669, 0.008620689655172414, 0, 0, 0.012658227848101266, 0, 0.001949317738791423, 0, 0 ]

[ "Q:\n\nIs Grothendieck a computer?", "\n\nI can't resist asking this companion question to the one of Gowers. ", "There, Tim Dokchitser suggested the idea of Grothendieck topologies as a fundamentally new insight. ", "But Gowers' original motivation is to probe the boundary between a human's way of thinking and that of a computer. ", "I argued, therefore, that Grothendieck topologies might be more natural to computers, in some sense, than to humans. ", "It seems Grothendieck always encouraged people to think of an object in terms of the category that surrounds it, rather than its internal structure. ", "That is, even the most lovable mathematical structure might be represented simply as a symbol $A$, and its special properties encoded in arrows $A\\rightarrow B$ and $C\\rightarrow A$, that is, a grand combinatorial network. ", "I'm tempted to say that the idea of a Grothendieck topology is something of an obvious corollary of this framework. ", "It's not something I've devoted much thought to, but it seems this is exactly the kind of reasoning more agreeable to a computer than to a woolly, touchy-feelly thinker like me.", "\nSo the actual question is, what other mathematical insights do you know that might come more naturally to a computer than to a human? ", "I won't try here to define computers and humans, for lack of competence. ", "I don't think having a deep knowledge of computers is really a prerequisite for the question or for an answer. ", "But it would be nice if your examples were connected to substantial mathematics. ", "\nI see that this question is subjective (but not argumentative in intent), so if you wish to close it on those grounds, that's fine.", "\nAdded, 11 December: Being a faulty human, I had an inexplicable attachment to the past tense. ", "But, being weak-willed on top of it all, I am bowing to peer pressure and changing the title.", "\n\nA:\n\nA simplicial set is surely an idea which would be more natural to a computer. ", "Breaking a shape up into simplices is still something a human would do, because simplices are contractible geometric objects whose gluings one can explicitly describe. ", "But to pass from this to finite strings with face and degeneracy maps, and then to base your theory on that, is pure computer-thought... and, like any good computer idea, extremely pretty.", "\n\nA:\n\nI think the human/computer dichotomy you set up should be extended to\na human/mathematician/computer trichotomy, just because a substantial\nportion of \"mathematical maturity\" is about learning to think like a computer,\nin your sense.", "\nAnyway I've just put that in place to try and shore up my example. ", "It seems\nthat humans read \"let's say we have X and Y...\" and automatically take the\nextra step of assuming X and Y are unequal. ", "Computers wouldn't bother to\ntake this extra step. ", "Mathematicians, or at least I myself, split into \nthe X=Y and X$\\neq $Y cases but try to obviate that split when writing\ndown a proof.", "\n\nA:\n\nI disagree with the premise of this question.", "\nConventional computers follow a program written by a human. ", " I think, for example, Daniel Moskovich's answer about simplicial sets is something that a human programming a computer (or a computer scientist) would think of when trying to program a computer.", "\nFormalisms like these are things that we humans think of when programming a computer. ", " Hence we have a tendency to think of the as \"more mechanical\", or \"more like a computer\", etc.; ", "but I think it's a mistake to think that this is something that a computer \"would come up with\" on its own. ", " Really it's us humans that come up with them, just when we are thinking in terms of computing.", "\nThere are computers which can be thought of as actually \"thinking\" in a way similar to humans (as opposed to just following a program), e.g. IBM's Watson computer. ", " They need some large data set to learn from, though (just like we do), and if this large data set is all of the mathematics created by humans, then I think the mathematics produced by the computer would look a lot like things \"a human would think of\"!", "\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0.03225806451612903, 0.014084507042253521, 0.02, 0.008695652173913044, 0.008547008547008548, 0.006711409395973154, 0.004464285714285714, 0.008620689655172414, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.005128205128205128, 0, 0, 0, 0, 0.012121212121212121, 0, 0 ]

[ "Na+,K(+)-ATPase activity is reduced in hippocampus of rats submitted to an experimental model of depression: effect of chronic lithium treatment and possible involvement in learning deficits.", "\nThis study was undertaken to verify the effects of chronic stress and lithium treatments on the hippocampal Na+,K(+)-ATPase activity of rats, as well as to investigate the effects of stress interruption and post-stress lithium treatment on this enzyme activity and on spatial memory. ", "Two experiments were carried out; in the first experiment, adult male Wistar rats were divided into two groups: control and submitted to a chronic variate stress paradigm, and subdivided into treated or not with LiCl. ", "After 40 days of treatment, rats were killed, and Na+,K(+)-ATPase activity was determined. ", "In the second experiment, rats were stressed during 40 days, and their performance was evaluated in the Water Maze task. ", "The stressed group was then subdivided into four groups, with continued or interrupted stress treatment and treated or not with lithium for 30 additional days. ", "After a second evaluation of performance in the Water Maze, rats were killed and Na+,K(+)-ATPase activity was also measured. ", "Results showed an impairment in Na+,K(+)-ATPase activity and in Water Maze performance of chronically stressed rats, which were prevented by lithium treatment and reversed by lithium treatment and by stress interruption. ", "These results suggest that the modulation of Na+,K(+)-ATPase activity may be one of the mechanisms of action of lithium in the treatment of mood disorders." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0.009174311926605505, 0, 0, 0, 0.008, 0, 0 ]

[ "Vascular changes in hypertension.", "\nHypertension can directly damage blood vessels, and leads to renal failure, intracranial bleeds, and lacunar infarctions. ", "Of equal importance is the effect of hypertension on the development of atherosclerosis. ", "Specific changes in both the microvasculature and macrovasculature vary depending on the degree and rapidity of blood pressure elevation. ", "Changes in the intima and media can lead to significant narrowing of vessels and ischemia in various tissues. ", "In addition, changes in small-resistance vessels contribute to changes in peripheral-vasculature resistance and thus affect blood pressure regulation. ", "Treatment of moderate to severe elevation in blood pressure clearly results in a decrease in the incidence of stroke. ", "However, evidence that treating mild hypertension reduces coronary events is less convincing. ", "Antihypertensive therapy may result in partial regression of vascular changes, especially fibrinoid necrosis seen in malignant hypertension, but more work needs to be done to clearly define the roles of specific drugs in preventing or regressing hypertensive vascular disease." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0, 0, 0, 0, 0, 0, 0 ]

[ "Superintendent Jeremy Biehl says the school board decided Wednesday night to send preliminary layoff slips to all 34 members of the teaching staff, including librarians and counselors. ", "Biehl says the action was taken because of the uncertainty of the state budget bill." ]

{ "pile_set_name": "OpenWebText2" }

[ 0.005405405405405406, 0 ]

[ "High comorbidity index is not associated with high morbidity and mortality when employing constrained arthroplasty as a primary treatment for intertrochanteric fractures in elderly patients.", "\nThe aim of our study is to investigate the results of constrained total hip arthroplasty as a primary treatment of intertrochanteric fractures (ITF) in elderly patients with high comorbidities. ", "Total hip replacement (THR) with a retentive cup was performed on 73 patients with ITF over the age of 54 years who had high comorbidities and a Charlson score above five. ", "Short- and long-term complications were determined by follow-up. ", "Bivariate analysis was conducted in order to determine the possible determinants of mortality and factors associated with comorbidity as measured by the Charlson comorbidities index. ", "Patient demographics that consisted of females (58.9%) (p < 0.04) with the mean age of both males and females demonstrated no statistical significance. ", "The mean hospitalization time and weight bearing time were 11 and 2.67 days, respectively. ", "Only 4.1% needed re-intervention due to re-fracture and none due to prosthesis failure. ", "There was a statistical significance between the comorbidity index and the mortality rate. ", "However, no statistical significance was identified between the comorbidity index and the functional status after constrained THR. ", "High comorbidity index is not associated with high morbidity and mortality when employing constrained arthroplasty as a primary treatment for ITF in elderly patients." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0.005813953488372093, 0, 0, 0, 0, 0, 0, 0.007633587786259542, 0 ]

[ "SpaceX Manned Dragon Spacecraft Arriving at the International Space Station for Automatic Docking\n\nPicture: SpaceXs Crew Dragon slowly approached the International Space Station on the morning of March 3, local time.", "\n\nNetease Technologies News, March 4, according to TheVerge, on the second day of launch into orbit, SpaceXs Crew Dragon docked with the International Space Station, which is a key part of the companys current test flight. ", "When the manned dragon spacecraft begins to transport astronauts to and from the International Space Station, this docking operation will become a routine operation in the future.", "\n\nSpaceXs Falcon 9 rocket launched from the Kennedy Space Center in Florida aboard a manned dragon spacecraft around 3 a.m. local time on Tuesday.", "\n\nOn the morning of the 3rd, SpaceX sent the manned dragon spacecraft to different locations outside the space station to test the docking capability of the spacecraft. ", "With propellers, the spacecraft regularly approached the International Space Station and then remained in position for two and a half hours. ", "The spacecraft even retreated once to test its ability to retreat in an emergency. ", "Then, when the manned dragon spacecraft reached a distance of about 20 meters, the final instructions were sent to the capsule to dock with the International Space Station.", "\n\nAt 5:51 p.m. local time on the 3rd, the capsule was attached to the docking port, and various hooks were deployed outside the interface to ensure the stability of the capsule. ", "We can confirm that the hard capture mission has been completed, said Anne McClain, NASA astronaut and current ISS worker. ", "David Saint-Jacques, a Canadian astronaut who also works on the International Space Station, said: Congratulations again to teams around the world who helped make this mission a success.", "\n\nPhoto: Anne McLean and David St. Jacques are ready to open the hatch of the International Space Station to welcome the arrival of the Manned Dragon Spacecraft.", "\n\nIts a new technology for SpaceX, which has never automatically docked a spacecraft to a space station before. ", "Since May 2012, SpaceX has been launching cargo dragon spacecraft to the International Space Station, but all of these vehicles have docked in mooring mode. ", "Astronauts on the International Space Station capture the capsule with robotic arms, and then transfer the Dragon spacecraft to the available docking ports. ", "Now, the new Manned Dragon Spacecraft has shown that it no longer needs this traditional model.", "\n\nThe docking marked a major success for SpaceX and NASAs Commercial Astronaut Program. ", "Manned Dragon Spacecraft is one of the two private vehicles being developed by the United States to launch NASA astronauts from the United States to the International Space Station. ", "Docking is a key process to ensure that passengers arrive at the space station. ", "The latest demonstration shows that the Manned Dragon Spacecraft has the ability to deliver astronauts safely to the International Space Station.", "\n\nAs the Manned Dragon Spacecraft connects to the International Space Station, the three people stationed above will open the hatch door of the capsule, enter it and greet the intelligent dummy Ripley. ", "They will also open 180 kilograms of cargo on board the Manned Dragon Spacecraft and conduct a series of tests on the spacecraft in the coming days. ", "Its purpose is to observe how the spacecraft responds to the space environment, and whether its design is safe and as expected to withstand the test.", "\n\nOn Friday, SpaceX will take the capsule back to the United States and try to land in the Atlantic with a parachute. ", "The landing will mark the final milestone of the test mission and the end of the current flight. ", "After that, NASA will evaluate the performance of the manned dragon spacecraft during this flight and ultimately decide when to carry two astronauts on the next test flight. ", "SpaceX manned trials are scheduled for this summer, but NASA and SpaceX still have a lot of work to do before the flight. (", "small)" ]

{ "pile_set_name": "Pile-CC" }

[ 0.013888888888888888, 0.013452914798206279, 0.00558659217877095, 0, 0, 0.0070921985815602835, 0, 0.005813953488372093, 0, 0.024390243902439025, 0.010752688172043012, 0.018633540372670808, 0, 0.006369426751592357, 0.012738853503184714, 0.010526315789473684, 0, 0.016483516483516484, 0, 0.006896551724137931, 0.01485148514851485, 0.006711409395973154, 0, 0, 0, 0.005747126436781609, 0.008130081300813009, 0 ]

[ "LAKE BUENA VISTA, Fla. — Giancarlo Stanton was introduced by the New York Yankees moments ago, donning the pinstripes for the first time and meeting a mass of media. ", "The takeaway from the presser — apart from how happy he seems to be about joining the Yankees — is how poorly the Miami Marlins handled his departure.", "\n\nEarlier in the day Derek Jeter — who, despite being in charge of the Marlins baseball operations department is not here for the Winter Meetings — had a telephone press conference in which he seemed to bristle at the bad press he’s received since the Stanton trade was announced and tried to tell his side of the story of the deal. ", "Here’s his side of the story:\n\nJeter said he related to Stanton that #Marlins wanted him \"to be with us,\" but that Stanton wanted to move along. #", "yankees\n\nDuring his press conference just now Stanton was asked if Jeter persuaded him to stay. ", "He paused for several seconds when he was asked that. ", "Then he answered:\n\nWe had a meeting, yes. ", "We spoke about the direction of the team. ", "I wanted us to go forward and have — and advance with the pitching staff. ", "I thought our lineup was legit and we needed help with our pitchers, and we needed to add rather than subtract. ", "The way they wanted to go was to subtract, so I let that be known that I didn’t want to be part of another rebuild, another losing season, and that’s almost a guaranteed losing season taking away what I thought was a great lineup. ", "So, yes, I didn’t want to be a part of the rebuild.", "\n\nIt was quite clear from both his words and his demeanor that Stanton gave the Marlins a reasonable set of circumstances that would make him want to stay. ", "It was likewise clear that the Marlins did nothing to persuade him to stay.", "\n\nThey also bungled the trade.", "\n\nStanton has taken a lot of heat for rejecting deals to the Giants and the Cardinals. ", "Stanton said, though, that he and his agent presented the Marlins with a list of teams to whom he’d accept a trade beforehand and the Giants and Cardinals were not on the list. ", "Stanton:\n\nNow, I gave my list of teams prior to, and they went to San Francisco and Cardinals and struck deals with them. ", "So I was open to listen to them, but those were not my teams.", "\n\nThe Marlins nonetheless struck those deals and presented them to Stanton. ", "Stanton said that he had a lot of respect for those organizations so he took meetings with them but that they were never on his list and they just weren’t deals that were going to work out.", "\n\nNews of those deals, it’s worth remembering, came out primarily from a reporter based in south Florida, suggesting that they were leaked by the Marlins. ", "So not only was the club pursuing deals they should have known Stanton was going to reject, they made his act of rejecting them an awkward, public act which made Stanton into the bad guy in some quarters. ", "Once those deals feel apart, of course, the Yankees quickly swooped in and got him for a good second baseman who will soon be traded and some magic beans.", "\n\nStanton came off really well in his presser. ", "He answered every questions thrown at him. ", "He praised the Cardinals and the Giants and explained his decision to join the Yankees without delving into the sorts of cliches often used when talking about the Yankees, focusing less on their history and brand and more on their good young players and their current state of competitiveness. ", "He also said, when asked whether he’ll play right field given that Aaron Judge is there, that he doesn’t care where he played as long as he can help the team win. ", "It was a master class in how to introduce yourself to the New York media.", "\n\nContrast that with what’s coming out of Miami. ", "And ask yourself if we should maybe rethink what we talk about when we talk about Yankees, current and former, and the idea of “class.”" ]

{ "pile_set_name": "Pile-CC" }

[ 0.018072289156626505, 0.006666666666666667, 0.003003003003003003, 0.0136986301369863, 0.020833333333333332, 0, 0, 0, 0, 0, 0, 0, 0.00641025641025641, 0, 0, 0.022988505747126436, 0.011299435028248588, 0.00819672131147541, 0, 0.02631578947368421, 0.005291005291005291, 0, 0.00975609756097561, 0.006493506493506494, 0, 0.023255813953488372, 0.01020408163265306, 0.006134969325153374, 0, 0, 0.007407407407407408 ]

[ "704\t0\t0\t0\t0\t0\t3\t7804\t-\tchr10.1029.19.wu1\t723\t6\t710\tchr10\t135413628\t49700428\t49708936\t4\t160,239,119,186,\t13,173,412,531,\t49700428,49704721,49706828,49708750,\n271\t0\t0\t0\t0\t0\t2\t12285\t-\tchr1.1756.9.uw1\t277\t3\t274\tchr1\t245522847\t84549829\t84562385\t3\t103,79,89,\t3,106,185,\t84549829,84559517,84562296,\n" ]

{ "pile_set_name": "Github" }

[ 0 ]

[ "![](", "menthealthlond70646-0023){#sp1 .19}\n\n![](", "menthealthlond70646-0024){#sp2 .20}\n\n![](", "menthealthlond70646-0025){#sp3 .21}\n\n![](", "menthealthlond70646-0026){#sp4 .22}\n" ]

{ "pile_set_name": "PubMed Central" }

[ 0, 0.024390243902439025, 0, 0, 0 ]

[ "/*\n * Copyright (C) 2011 Nokia Corporation and/or its subsidiary(-ies)\n *\n * This library is free software; you can redistribute it and/or\n * modify it under the terms of the GNU Library General Public\n * License as published by the Free Software Foundation; either\n * version 2 of the License, or (at your option) any later version.", "\n *\n * This library is distributed in the hope that it will be useful,\n * but WITHOUT ANY WARRANTY; without even the implied warranty of\n * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. ", " See the GNU\n * Library General Public License for more details.", "\n *\n * You should have received a copy of the GNU Library General Public License\n * along with this library; see the file COPYING.LIB. ", " If not, write to\n * the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,\n * Boston, MA 02110-1301, USA.", "\n */\n\n#include \"config.h\"\n#include \"WebHitTestResultData.h\"\n\n#include \"WebCoreArgumentCoders.h\"\n#include <WebCore/Document.h>\n#include <WebCore/Frame.h>\n#include <WebCore/FrameView.h>\n#include <WebCore/HitTestResult.h>\n#include <WebCore/Node.h>\n#include <WebCore/RenderObject.h>\n#include <WebCore/URL.h>\n#include <wtf/text/WTFString.h>\n\nusing namespace WebCore;\n\nnamespace WebKit {\n\nWebHitTestResultData::WebHitTestResultData()\n{\n}\n\nWebHitTestResultData::WebHitTestResultData(const WebCore::HitTestResult& hitTestResult)\n : absoluteImageURL(hitTestResult.absoluteImageURL().string())\n , absolutePDFURL(hitTestResult.absolutePDFURL().string())\n , absoluteLinkURL(hitTestResult.absoluteLinkURL().string())\n , absoluteMediaURL(hitTestResult.absoluteMediaURL().string())\n , linkLabel(hitTestResult.textContent())\n , linkTitle(hitTestResult.titleDisplayString())\n , isContentEditable(hitTestResult.isContentEditable())\n , elementBoundingBox(elementBoundingBoxInWindowCoordinates(hitTestResult))\n , isScrollbar(hitTestResult.scrollbar())\n , isSelected(hitTestResult.isSelected())\n , isTextNode(hitTestResult.innerNode() && hitTestResult.innerNode()->isTextNode())\n , isOverTextInsideFormControlElement(hitTestResult.isOverTextInsideFormControlElement())\n , allowsCopy(hitTestResult.allowsCopy())\n , isDownloadableMedia(hitTestResult.isDownloadableMedia())\n , imageSize(0)\n{\n}\n\nWebHitTestResultData::WebHitTestResultData(const WebCore::HitTestResult& hitTestResult, bool includeImage)\n : absoluteImageURL(hitTestResult.absoluteImageURL().string())\n , absolutePDFURL(hitTestResult.absolutePDFURL().string())\n , absoluteLinkURL(hitTestResult.absoluteLinkURL().string())\n , absoluteMediaURL(hitTestResult.absoluteMediaURL().string())\n , linkLabel(hitTestResult.textContent())\n , linkTitle(hitTestResult.titleDisplayString())\n , isContentEditable(hitTestResult.isContentEditable())\n , elementBoundingBox(elementBoundingBoxInWindowCoordinates(hitTestResult))\n , isScrollbar(hitTestResult.scrollbar())\n , isSelected(hitTestResult.isSelected())\n , isTextNode(hitTestResult.innerNode() && hitTestResult.innerNode()->isTextNode())\n , isOverTextInsideFormControlElement(hitTestResult.isOverTextInsideFormControlElement())\n , allowsCopy(hitTestResult.allowsCopy())\n , isDownloadableMedia(hitTestResult.isDownloadableMedia())\n , imageSize(0)\n{\n if (!", "includeImage)\n return;\n\n if (Image* image = hitTestResult.image()) {\n RefPtr<SharedBuffer> buffer = image->data();\n if (buffer) {\n imageSharedMemory = WebKit::SharedMemory::allocate(buffer->size());\n memcpy(imageSharedMemory->data(), buffer->data(), buffer->size());\n imageSize = buffer->size();\n }\n }\n}\n\nWebHitTestResultData::~WebHitTestResultData()\n{\n}\n\nvoid WebHitTestResultData::encode(IPC::ArgumentEncoder& encoder) const\n{\n encoder << absoluteImageURL;\n encoder << absolutePDFURL;\n encoder << absoluteLinkURL;\n encoder << absoluteMediaURL;\n encoder << linkLabel;\n encoder << linkTitle;\n encoder << isContentEditable;\n encoder << elementBoundingBox;\n encoder << isScrollbar;\n encoder << isSelected;\n encoder << isTextNode;\n encoder << isOverTextInsideFormControlElement;\n encoder << allowsCopy;\n encoder << isDownloadableMedia;\n encoder << lookupText;\n encoder << dictionaryPopupInfo;\n\n WebKit::SharedMemory::Handle imageHandle;\n if (imageSharedMemory && imageSharedMemory->data())\n imageSharedMemory->createHandle(imageHandle, WebKit::SharedMemory::Protection::ReadOnly);\n encoder << imageHandle;\n encoder << imageSize;\n\n bool hasLinkTextIndicator = linkTextIndicator;\n encoder << hasLinkTextIndicator;\n if (hasLinkTextIndicator)\n encoder << linkTextIndicator->data();\n\n platformEncode(encoder);\n}\n\nbool WebHitTestResultData::decode(IPC::ArgumentDecoder& decoder, WebHitTestResultData& hitTestResultData)\n{\n if (!", "decoder.decode(hitTestResultData.absoluteImageURL)\n || !", "decoder.decode(hitTestResultData.absolutePDFURL)\n || !", "decoder.decode(hitTestResultData.absoluteLinkURL)\n || !", "decoder.decode(hitTestResultData.absoluteMediaURL)\n || !", "decoder.decode(hitTestResultData.linkLabel)\n || !", "decoder.decode(hitTestResultData.linkTitle)\n || !", "decoder.decode(hitTestResultData.isContentEditable)\n || !", "decoder.decode(hitTestResultData.elementBoundingBox)\n || !", "decoder.decode(hitTestResultData.isScrollbar)\n || !", "decoder.decode(hitTestResultData.isSelected)\n || !", "decoder.decode(hitTestResultData.isTextNode)\n || !", "decoder.decode(hitTestResultData.isOverTextInsideFormControlElement)\n || !", "decoder.decode(hitTestResultData.allowsCopy)\n || !", "decoder.decode(hitTestResultData.isDownloadableMedia)\n || !", "decoder.decode(hitTestResultData.lookupText)\n || !", "decoder.decode(hitTestResultData.dictionaryPopupInfo))\n return false;\n\n WebKit::SharedMemory::Handle imageHandle;\n if (!", "decoder.decode(imageHandle))\n return false;\n\n if (!", "imageHandle.isNull())\n hitTestResultData.imageSharedMemory = WebKit::SharedMemory::map(imageHandle, WebKit::SharedMemory::Protection::ReadOnly);\n\n if (!", "decoder.decode(hitTestResultData.imageSize))\n return false;\n\n bool hasLinkTextIndicator;\n if (!", "decoder.decode(hasLinkTextIndicator))\n return false;\n\n if (hasLinkTextIndicator) {\n WebCore::TextIndicatorData indicatorData;\n if (!", "decoder.decode(indicatorData))\n return false;\n\n hitTestResultData.linkTextIndicator = WebCore::TextIndicator::create(indicatorData);\n }\n\n return platformDecode(decoder, hitTestResultData);\n}\n\n#if !", "PLATFORM(MAC)\nvoid WebHitTestResultData::platformEncode(IPC::ArgumentEncoder& encoder) const\n{\n}\n\nbool WebHitTestResultData::platformDecode(IPC::ArgumentDecoder& decoder, WebHitTestResultData& hitTestResultData)\n{\n return true;\n}\n#endif // !", "PLATFORM(MAC)\n\nIntRect WebHitTestResultData::elementBoundingBoxInWindowCoordinates(const WebCore::HitTestResult& hitTestResult)\n{\n Node* node = hitTestResult.innerNonSharedNode();\n if (!", "node)\n return IntRect();\n\n Frame* frame = node->document().frame();\n if (!", "frame)\n return IntRect();\n\n FrameView* view = frame->view();\n if (!", "view)\n return IntRect();\n\n RenderObject* renderer = node->renderer();\n if (!", "renderer)\n return IntRect();\n\n return view->contentsToWindow(renderer->absoluteBoundingBoxRect());\n}\n\n} // WebKit\n" ]

{ "pile_set_name": "Github" }

[ 0.009009009009009009, 0, 0, 0.007407407407407408, 0.016666666666666666, 0.005363036303630363, 0.005714285714285714, 0, 0, 0, 0, 0, 0, 0, 0.015384615384615385, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.006172839506172839, 0, 0, 0.00904977375565611, 0.01639344262295082, 0.005208333333333333, 0.011494252873563218, 0.0125, 0.02247191011235955, 0.008064516129032258 ]

[ "Introduction {#sec1-1}\n============\n\nProtein misfolding and aberrant self-assembly in an infinitely propagating fashion-forming large molecular weight aggregates is a key pathognomonic feature of various seemingly unrelated neurodegenerative diseases, including Alzheimer\\'s disease (AD), Parkinson\\'s disease (PD), Huntington disease (HD) and Prion disease.\\[[@ref1]\\] The accumulation of these insoluble fibrous protein aggregates termed as amyloid in various organs/tissues causes their degeneration and loss of function. ", "The fibrillar deposits are commonly known as amyloid plaques or neurofibrillary tangles in AD, Lewy bodies in PD, and nuclear inclusions in HD. ", "The aggregating proteins are different in different neurodegenerative disorders. ", "Further, the occurrence of these diseases can be sporadic or hereditary and the aggregates can localize intracellularly or extracellularly \\[[Table 1](#T1){ref-type=\"table\"}\\]. ", "For past many decades, the field of amyloid disorders has been a major concern in medicine as these are progressive disorders and result in irreversible neurodegeneration. ", "The amyloids demonstrated as anomalous tissue deposits, were initially misidentified as starch but later determined to be composed of protein.\\[[@ref2]\\] Histologically, these deposits exhibit affinity for Congo red dye with concomitant apple-green birefringence under plane polarized light,\\[[@ref3]\\] which has long been considered as a gold standard in diagnosis of amyloid disorders. ", "In the mid-20th century, with the advancement in high resolution structural techniques such as electron microscopy, X-ray diffraction, NMR (nuclear magnetic resonance) it became evident that the amyloid deposits consists of bundles of unbranched fibrils\\[[@ref4]\\] and the constituent proteins were rich in highly ordered cross-β structures with β-strands running perpendicular to fibrillar axis.\\[[@ref5][@ref6][@ref7]\\] These fibrils have an indefinite length and diameters of ∼2-20 nm. ", "Moreover, these β-sheets have different arrangements in different amyloid proteins. ", "For instance, parallel in-register β-sheets as in α-synuclein, β-microglobulin, β-amyloid (Aβ~1-40~ and Aβ~1-42~), amylin, human prion protein\\[[@ref8][@ref9][@ref10][@ref11][@ref12]\\] and antiparallel β-sheets as in Aβ~16-22~ and Aβ~11-25~ fragments, C terminus of Aβ~34-42~, amylin and huntingtin.\\[[@ref13][@ref14][@ref15]\\] X-ray studies have revealed that mature fibrils have 'steric zipper structure' formed by interlocking of emanating side chains, like the teeth of a zipper.\\[[@ref16]\\]\n\n###### \n\nProteins causing amyloid diseases and location of filamentous lesions\n\n![](", "AIAN-18-138-g001)\n\nProtein Aggregation Model {#sec1-2}\n=========================\n\nAlthough various triggering factors (including increased temperature, mechanical stress, extreme pH, oxidative stress, glycation or mutation) have been implicated for protein misfolding, the basic mechanism of fibril formation is same in all amyloid diseases as shown in [Figure 1](#F1){ref-type=\"fig\"}. ", "According to the fibril formation model, native proteins misfold and undergo conformational change. ", "When these misfolded proteins reach a critical concentration, oligomers are formed that result in protofibrils and finally culminate in mature fibrils \\[[Figure 1a](#F1){ref-type=\"fig\"}\\]. ", "The oligomeric intermediates are referred to as prefibrillar or on-pathway oligomers. ", "In addition to these prefibrillar oligomers, off-pathway oligomers are formed which do not end up in fibrils \\[[Figure 1b\\]](#F1){ref-type=\"fig\"}. ", "Further, the mature fibrils can shear or fragment to form fibrillar oligomers (originated from fibrils) that can again aggregate to form mature fibrils \\[[Figure 1c](#F1){ref-type=\"fig\"}\\]. ", "The molecular structure of these transient species is not clearly defined but they exhibit spherical or annular morphology where annular oligomers are doughnut or ring shaped assemblies.\\[[@ref17]\\] In addition, FTIR (fourier-transform infrared spectroscopy) and NMR studies showed that oligomers lack the parallel in-register β-sheet structure present in fibrillar form but they contain the β-loop-β secondary and tertiary folds as in fibrils.\\[[@ref18][@ref19][@ref20]\\] However, these oligomers formed by various amyloid proteins (associated with different neurodegenerative disorders) have common immunological epitopes that are recognized by conformational antibodies, A11 (spherical) and αAPF (annular). ", "Interestingly, A11 and αAPF antibodies do not bind to mature fibrils or oligomers formed by fragmentation of fibrils/protofibrils (fibrillar oligomers).\\[[@ref17][@ref21]\\] The mature fibrils and the fibrillar oligomers share generic epitopes that are recognized by another conformational antibody OC. ", "OC antibody does not bind prefibrillar oligomers suggesting that prefibrillar and fibrillar oligomers are structurally different.\\[[@ref22]\\]\n\n{#F1}\n\nRecently, a new dimension has been added to this already existing fibrillization model. ", "Cohen *et al.*, ", "2013 have *in vitro* demonstrated secondary nucleation model for fibril formation using Aβ~42~ peptide. ", "This model suggests that mature fibrils catalyze the secondary nucleation reaction by forming diffusible oligomers from the monomers \\[[Figure 1d](#F1){ref-type=\"fig\"}\\]. ", "This reaction depends on the concentration of both native protein (monomers) and fibrils unlike the primary nucleation reaction which depends only on protein monomer concentration.\\[[@ref23]\\]\n\nToxicity of Protein Aggregation {#sec1-3}\n===============================\n\nAccording to the original amyloid cascade hypothesis (ACH) proposed in 1991, mutations in three genes: APP, Presenilin 1 and 2 were thought to be the initiating events in the abnormal accumulation of Aβ peptides (mature fibrils) leading to cell death and dementia in AD. ", "Earlier researchers used to believe that large fibril deposits were the primary cause of neuronal damage and disease pathogenesis. ", "Since protein aggregation is a dynamic process and involves formation of transient intermediate structures as well as dissociation of mature fibrils, it has been challenging to define the real culprit leading to toxicity. ", "However over the years, no linear correlation between dementia/cognitive decline and Aβ deposition in the brain was observed. ", "Further, increasing number of recent studies suggests intermediate oligomers formed during fibrillization to be more toxic species than mature fibrils. ", "Instead in some instances, formation of mature fibrils has been proposed to be protective by acting as harmless reservoir of toxic oligomers. ", "Hence, the original ACH has been modified to include the intermediary species formed during aggregation/fibrillization process and is commonly known as oligomer hypothesis. ", "In the following sections, we intent to review the ongoing debate of toxicity demonstrated by fibrils and oligomers and define the current status of amyloid pathology.", "\n\nToxicity of Oligomers versus Mature Fibrils {#sec1-4}\n===========================================\n\nPresence of amyloid plaque rich in Aβ fibrils in brains of AD patients supported the close link of protein aggregation with neurodegeneration. ", "Initial studies postulated amyloid fibrils to be the primary cause of cell death and disease pathogenesis. ", "It was demonstrated by independent approaches that application of extracellular Aβ fibrils induced AD-like changes. ", "In cultured neurons, Aβ fibrils increased frequency of action potentials and membrane depolarization and reduced the cell viability.\\[[@ref24]\\] Injection of Aβ fibrils in primary hippocampal neurons in rat-impaired synaptic transmission, induced cognitive/memory decline and caused neuronal cell death.\\[[@ref25]\\] However, there have been conflicting reports to define the correlation between amyloid plaque burden and severity of neuronal loss and other AD-symptoms. ", "Hey *et al.*, ", "2012 demonstrated that oligomeric Aβ infused into left ventricles of brain showed significant impairment of learning and memory functions while there was no significant effect observed due to fibrils.\\[[@ref26]\\] Moreover, in a cell culture study Aβ oligomers were found to suppress synaptic plasticity by specifically inhibiting presynaptic P/Q calcium currents unlike protein monomers and fibrils.\\[[@ref27]\\] Most of the earlier transgenic AD animal models exhibited plaque formation as the dominant pathology feature of AD with little neurotoxicity but a very recent AD transgenic mice supports that neurotoxicity is due to Aβ oligomers (oligomer hypothesis). ", "In this transgenic mouse with mutation in presenilin gene (PS1~V97L~-Tg), Aβ oligomers accumulated in the neurons and exhibited memory and synaptic dysfunction and tau hyperphosphorylation. ", "No extracellular amyloid plaques were observed in this transgenic mice.\\[[@ref28]\\] Interestingly, Aβ oligomers extracted from the cerebral cortex of AD patients when administered causes memory deficits and disrupts synaptic plasticity in normal rat.\\[[@ref29]\\] Later on, during the analysis of fractionated brain homogenate from patients with AD it was observed that levels of Aβ oligomers is associated with the loss of synaptic markers such as the synaptic vesicle protein VAMP2, and the post-synaptic protein PSD95.\\[[@ref30][@ref31]\\]\n\nIn case of PD, where α-synuclein (α-syn, the protein associated with PD) is considered as a major component of protein inclusions (Lewy bodies or Lewy neurites), a detailed attempt has been made to study the process of aggregate formation for identifying the toxic species responsible for neuronal dysfunction and cell death. ", "Specific α-synucleinopathies were achieved in transgenic mouse models by expressing human wild type and mutant α-syn in brain regions. ", "The mutant α-syn form inclusions containing fibrils showed behavioral impairment in these transgenic models such as locomotor dysfunction and cognitive decline leading to death.\\[[@ref32][@ref33]\\] Later, it has been shown that α-syn fibrils decrease cell survival in cell culture system by inducing the activation of caspase-3 pathway leading to apoptotic cell death.\\[[@ref34]\\] In yet other studies, soluble oligomeric α-syn have been proposed as the main toxic species instead of α-syn fibrils. ", "In an *in vivo* study using murine rat model, it was found that unlike fibrils, lentiviral injected α-syn oligomers increases loss of nigral tyrosine hydroxylase (TH)-positive neurons and showed strong reactivity with membrane, thus exhibiting its neuronal toxicity through membrane disruption.\\[[@ref35]\\] Furthermore, overexpression of α-syn mutant that has high propensity to form oligomers but not fibrils, in three established PD models: Mammalian neurons, *Caenorhabditis elegans* and *Drosophila melanogaster* resulted in lower mitochondrial dehydrogenase activity and degeneration of dopaminergic and other neurons.\\[[@ref36]\\] In this study, it was evident that α-syn variants with impaired ability to form fibril correlates better with increased toxicity and neurodegeneration in PD.", "\n\nThe reasons for the differential toxicity of amyloid oligomers and fibrils can be explained by their structural arrangements:\n\nIn oligomers hydrophobic surfaces are exposed in β-sheets while they are hidden inside the interacting stacks in fibrils\\[[@ref10]\\]Oligomers are smaller in size so can easily diffuse in tissues as compared to longer fibrils\\[[@ref37]\\]The more number of open active ends present in case of oligomers than fibrils allow improved interaction of oligomers with cellular targets.", "Oligomers are highly unstable disordered structures whereas fibrils are stable organized molecules.", "\n\nAlthough the toxicity studies reviewed here portrays only part of the advancing knowledge on the ongoing debate, it may be sufficient to clearly conclude that there is a link between protein aggregation and disease pathogenesis. ", "However, there is a need of development of *in vivo* imaging techniques to dissect the process of protein aggregation and provide definite clinical correlation of disease stage or severity with functional proteins, oligomeric intermediates and mature fibrils.", "\n\nCellular Mechanism of Toxicity {#sec1-5}\n==============================\n\nThe similarities in amyloid structures (oligomer and mature fibrils) of various amyloid proteins irrespective of its sequence, suggest a common shared mechanism of toxicity. ", "Since amyloid proteins are cytosolic and/or extracellular, interaction of higher molecular weight amyloid structures with molecules on the plasma membrane can be the primary target for toxicity. ", "There are two primary mechanisms proposed, which are proximal to the bifurcated downstream events in the causal cascade:\n\nMembrane interaction {#sec2-1}\n--------------------\n\nMature fibrils interact with monosialotetrahexosylganglioside (GM1)-rich membrane domains whereas oligomeric species are proposed to interact with glutamatergic receptors, voltage-gated channels and GM1-rich domains on the membrane to initiate the neurotoxic cascade.\\[[@ref38][@ref39][@ref40]\\] Aβ~1-40~ and Aβ~1-42~ peptides interact with GM1 and adopts an alternate conformation depending on the protein density.\\[[@ref41][@ref42]\\] Lower protein to ganglioside ratio prefers α-helix rich structure, whereas higher ratio facilitates β-sheet forms, which culminates into toxic fibrils. ", "Various studies showed that during the accumulation of fibrils on GM1 cluster-membrane, no oligomers were detected, accounting that cell death is only due to fibril-induced disruption of cell membranes.\\[[@ref43][@ref44]\\] Instead, recent studies demonstrating interaction of neuritic cell membrane with Aβ~1-40~ oligomers and fibrils, oligomers were observed to damage the lipid bilayer by stimulating phospholipid composition and negative net charge of the membranes.\\[[@ref45]\\]α-synuclein oligomers were shown to interact with large and small unilamellar negatively charged vesicles on membrane and transiently altering the membrane permeability.\\[[@ref46]\\]\n\nPerturbation of calcium homeostasis {#sec2-2}\n-----------------------------------\n\nCalcium homeostasis perturbation was found to be a ubiquitous toxicity mechanism for soluble oligomers whereas no detectable effect was observed for fibrils.\\[[@ref47][@ref48]\\] There are conflicting reports about the mechanism by which calcium homeostasis gets disturb. (", "i) Amyloid ion channel formation: Demuro and many other researches tried to check whether this Ca^2+^ dysfunction is due to the pre-existing ion channel or oligomers creates altogether different channels in the membrane. ", "Cobalt which is known to block the Ca^2+^ channel didn't affect the Ca^2+^ influx, supporting the new channel hypothesis.\\[[@ref48]\\] (ii) Lipid bilayer conductance: In contrary to the channel hypothesis, Sokolov, in his study found that the increase in membrane conductance is due to the thinning of the lipid bilayer and lowering the dielectric barrier for ion translocation rather than channel formation.\\[[@ref49]\\] Numerous studies have shown that amyloid oligomers increase the lipid bilayer conductance where as there was no observable effect on lipid bilayer in case of fibrils.\\[[@ref50][@ref51]\\]\n\nIrrespective of the precise mechanism of membrane conductance, elevated intracellular Ca^2+^ levels is proposed as the central mechanism of toxicity since it regulates key downstream pathological events. ", "Membrane permeabilization and calcium dyshomeostasis may initiate various downstream events including transmembrane signaling processes, mitochondrial dysfunction, reactive oxygen species production and apoptotic pathway leading to toxicity.\\[[@ref52][@ref53][@ref54][@ref55][@ref56][@ref57][@ref58]\\] Both amyloid fibrils and oligomers stimulate these downstream cascades of events but with unique signaling responses. ", "Oligomers stimulate inflammatory response through NF-κB and differentially activate microglia through increasing levels of phosphorylated Lyn and SyK kinase as well as p38 MAP kinase, whereas fibrils stimulate greater amount of Keratinocyte chemoattractant chemokine and active phosphorylated form of ERK (extracellular signal-regulated kinase).\\[[@ref26][@ref59]\\]\n\nIntermediary Oligomers as Potential Biomarker {#sec1-6}\n=============================================\n\nAlthough the debate between the correlation of oligomers or mature fibrils with disease severity in amyloid disorders is ongoing, Aβ and α-synuclein oligomer count in CSF (cerebrospinal fluid) or plasma has been shown to be potential diagnostic biomarker for AD and PD. ", "In all these studies the oligomers were found to be elevated in the body fluid samples of patients as compared to controls.\\[[@ref60][@ref61]\\] Thus, the thrust is now towards the development of sensitive and specific assays to detect oligomeric amyloid species. ", "Among the few assays developed, Aβ-PMCA (protein misfolding cyclic amplification assay) and monoclonal single antibody sandwich ELISA assay for Aβ and time-resolved Forster resonance energy transfer (TR-FRET)-based immunoassays for α-synuclein oligomer detection are most sensitive.\\[[@ref62][@ref63][@ref64][@ref65]\\] These oligomeric assays are specific and sensitive to Aβ oligomers and do not recognize APP (amyloid precursor protein), monomeric Aβ and other non-Aβ-peptide oligomers. ", "Though, these studies suggest Aβ oligomer as a promising biomarker but the number of patient sample is small. ", "Further research on larger cohorts using highly sensitivity and rigorous standard assays is needed before considering the analysis of oligomers as a routine diagnostic assay for the clinical evaluation. ", "Moreover, there is also a need to develop similar oligomeric-based assays for other amyloid proteins.", "\n\nTherapeutics Under Clinical Trials {#sec1-7}\n==================================\n\nAt present, there are limited approved drugs for amyloid diseases. ", "Moreover, effects and benefits of these drugs are mainly symptomatic and marginal rather than targeting the underlying cause of the amyloid formation. ", "However, in recent years, the understanding of different protein conformers (monomers, oligomers, protofibrils and fibrils) in amyloid cascade has lead to development of new strategies for therapeutics and diagnostics.\\[[@ref66]\\]\n\nBroadly two major strategies are being explored for development of drugs for amyloid disorders-1) Small molecules as therapeutic agent: Numerous small molecules are being studied that can target different conformers in amyloid cascade and prevent amyloid formation \\[[Table 2](#T2){ref-type=\"table\"}\\]. ", "Their interventions can be at different steps of amyloid formation. ", "a) Molecules that inhibit formation of oligomers or fibrils: EGCG, a flavonoid present in green tea is under phase 3 trials.\\[[@ref67]\\] Recently, Keampferol-3-O-rhamnoside has been found as a promising molecule in abrogating Aβ toxicity by modulating monomers and remodeling oligomers and fibrils to non-toxic oligomers (Sharoar *et al.*, ", "2012).\\[[@ref68]\\] b) Molecules that destabilize oligomers and fibrils: Two proteases plasmin and cathepsin B destabilize Aβ oligomers and fibrils.\\[[@ref69][@ref70]\\] Rifampicin disaggregates the alpha-synuclein fibrils and inhibits their fibrillization.\\[[@ref71]\\] Scyllo-inositol now in phase 2 trial, effectively impedes Aβ aggregation by accelerating aggregate dissociation. ", "We have also recently shown that curcumin, a component in turmeric has the potential to modulate fibril formation as well as dissociate the preformed huntingtin (htt) aggregates in a unicellular eukaryote model organism.\\[[@ref72]\\]2) Immunotherapy: Therapeutic interventions based on both active (stimulation of immune response by injecting amyloid peptides) and passive (direct administration of anti-Aβ monoclonal antibodies) immunization are under different phases of clinical trial for many amyloid diseases. ", "Here, we briefly summarize the outcomes of clinical trials for AD \\[[Table 3](#T3){ref-type=\"table\"}\\]. ", "Initial clinical trials for active immunization with full length Aβ vaccines were discontinued due to severe side effects. ", "Smaller fragments of Aβ under trial appear to be more promising. ", "CAD-106, Aβ 1-6 peptides has successfully completed the phase 2 trial for mild AD patients without any side effects.\\[[@ref73]\\] Passive immunotherapy using antibodies against Aβ monomers, oligomers and fibrils are under trial and show promising results \\[[Table 3](#T3){ref-type=\"table\"}\\]. ", "Among these, Solanezumab targeting Aβ oligomer and Gantenerumab against Aβ fibrils/plaques are now under phase 3 trial.\\[[@ref74][@ref75]\\] Since in AD pathophysiology both Aβ and tau proteins are involved, combined antibody (IVIG) therapy targeting these amyloid proteins is also under trial.\\[[@ref76]\\]\n\n###### \n\nSmall molecules in clinical trials\n\n![](", "AIAN-18-138-g003)\n\n###### \n\nActive and passive immunotherapies in different phases of Clinical trials\n\n![](", "AIAN-18-138-g004)\n\nConcluding Remarks {#sec1-8}\n==================\n\nAs the average life span has increased globally, the probability of getting age-related neurodegenerative diseases, including amyloid disorders has increased. ", "Amyloid disorders are a major concern as they cause irreversible degeneration and are usually fatal. ", "Despite complexity of the process of protein aggregation, the progress made in the field of amyloid biology in the past few years is really commendable. ", "Today, we understand the mechanistic and structural aspect of protein aggregation disorders. ", "This knowledge is being exploited to define the toxic species that can be clearly associated with disease pathogenesis. ", "It is incredible how the transient oligomeric species formed during protein aggregation can be used as a biomarker. ", "Highly specific and sensitive assays are being developed for the detection for amyloid oligomers from either CSF or plasma. ", "This can significantly improve early diagnosis, which can lead to better management and treatment of the disease. ", "Though further studies are warranted to establish the oligomeric species as the real culprit, tremendous efforts toward development of strategies to inhibit protein aggregation offers hope for cure for amyloid disorders in near future.", "\n\n**Source of Support:** VT acknowledges the funding and fellowship from Innovative Young Biotechnologist Award, Department of Biotechnology, India. ", "MV acknowledge the Senior Research Fellowship from Indian Council of Medical Research (ICMR), India. ", "AV acknowledge the Junior Research Fellowship from ICMR, India\n\n**Conflict of Interest:** None declared.", "\n" ]

{ "pile_set_name": "PubMed Central" }

[ 0.005714285714285714, 0, 0, 0, 0, 0.005154639175257732, 0.010224948875255624, 0, 0.01549053356282272, 0, 0, 0, 0, 0, 0, 0.008450704225352112, 0.009933774834437087, 0.006896551724137931, 0, 0, 0, 0.0625, 0, 0, 0.005555555555555556, 0, 0, 0, 0, 0, 0.005780346820809248, 0, 0, 0, 0, 0.00425531914893617, 0, 0.0030120481927710845, 0.005263157894736842, 0.004608294930875576, 0, 0.004008016032064128, 0.0025220680958385876, 0.0039603960396039604, 0, 0, 0, 0, 0, 0.007863695937090432, 0.005888125613346418, 0.004524886877828055, 0.008620689655172414, 0.002380952380952381, 0.014864864864864866, 0.0076045627376425855, 0.016359918200409, 0, 0, 0, 0, 0, 0.001869158878504673, 0, 0.0058823529411764705, 0.010498687664041995, 0.0019455252918287938, 0, 0, 0, 0.003424657534246575, 0.014044943820224719, 0, 0, 0, 0, 0, 0, 0, 0.008064516129032258, 0, 0, 0.013422818791946308, 0.019801980198019802, 0.019230769230769232, 0 ]

[ "// Copyright 2013 The Flutter Authors. ", "All rights reserved.", "\n// Use of this source code is governed by a BSD-style license that can be\n// found in the LICENSE file.", "\n\n#ifndef FLUTTER_FLUTTERENGINE_H_\n#define FLUTTER_FLUTTERENGINE_H_\n\n#import <Foundation/Foundation.h>\n#import <UIKit/UIKit.h>\n\n#include \"FlutterBinaryMessenger.h\"\n#include \"FlutterDartProject.h\"\n#include \"FlutterMacros.h\"\n#include \"FlutterPlugin.h\"\n#include \"FlutterTexture.h\"\n\n@class FlutterViewController;\n\n/**\n * The FlutterEngine class coordinates a single instance of execution for a\n * `FlutterDartProject`. ", " It may have zero or one `FlutterViewController` at a\n * time, which can be specified via `-setViewController:`.", "\n * `FlutterViewController`'s `initWithEngine` initializer will automatically call\n * `-setViewController:` for itself.", "\n *\n * A FlutterEngine can be created independently of a `FlutterViewController` for\n * headless execution. ", " It can also persist across the lifespan of multiple\n * `FlutterViewController` instances to maintain state and/or asynchronous tasks\n * (such as downloading a large file).", "\n *\n * Alternatively, you can simply create a new `FlutterViewController` with only a\n * `FlutterDartProject`. ", "That `FlutterViewController` will internally manage its\n * own instance of a FlutterEngine, but will not guarantee survival of the engine\n * beyond the life of the ViewController.", "\n *\n * A newly initialized FlutterEngine will not actually run a Dart Isolate until\n * either `-runWithEntrypoint:` or `-runWithEntrypoint:libraryURI` is invoked.", "\n * One of these methods must be invoked before calling `-setViewController:`.", "\n */\nFLUTTER_EXPORT\n@interface FlutterEngine\n : NSObject <FlutterBinaryMessenger, FlutterTextureRegistry, FlutterPluginRegistry>\n/**\n * Initialize this FlutterEngine with a `FlutterDartProject`.", "\n *\n * If the FlutterDartProject is not specified, the FlutterEngine will attempt to locate\n * the project in a default location (the flutter_assets folder in the iOS application\n * bundle).", "\n *\n * A newly initialized engine will not run the `FlutterDartProject` until either\n * `-runWithEntrypoint:` or `-runWithEntrypoint:libraryURI:` is called.", "\n *\n * @param labelPrefix The label prefix used to identify threads for this instance. ", "Should\n * be unique across FlutterEngine instances, and is used in instrumentation to label\n * the threads used by this FlutterEngine.", "\n * @param projectOrNil The `FlutterDartProject` to run.", "\n */\n- (instancetype)initWithName:(NSString*)labelPrefix\n project:(FlutterDartProject*)projectOrNil NS_DESIGNATED_INITIALIZER;\n\n/**\n * The default initializer is not available for this object.", "\n * Callers must use `-[FlutterEngine initWithName:project:]`.", "\n */\n- (instancetype)init NS_UNAVAILABLE;\n\n/**\n * Runs a Dart program on an Isolate from the main Dart library (i.e. the library that\n * contains `main()`).", "\n *\n * The first call to this method will create a new Isolate. ", "Subsequent calls will return\n * immediately.", "\n *\n * @param entrypoint The name of a top-level function from the same Dart\n * library that contains the app's main() function. ", " If this is nil, it will\n * default to `main()`. ", " If it is not the app's main() function, that function\n * must be decorated with `@pragma(vm:entry-point)` to ensure the method is not\n * tree-shaken by the Dart compiler.", "\n * @return YES if the call succeeds in creating and running a Flutter Engine instance; NO otherwise.", "\n */\n- (BOOL)runWithEntrypoint:(NSString*)entrypoint;\n\n/**\n * Runs a Dart program on an Isolate using the specified entrypoint and Dart library,\n * which may not be the same as the library containing the Dart program's `main()` function.", "\n *\n * The first call to this method will create a new Isolate. ", "Subsequent calls will return\n * immediately.", "\n *\n * @param entrypoint The name of a top-level function from a Dart library. ", " If nil, this will\n * default to `main()`. ", " If it is not the app's main() function, that function\n * must be decorated with `@pragma(vm:entry-point)` to ensure the method is not\n * tree-shaken by the Dart compiler.", "\n * @param uri The URI of the Dart library which contains the entrypoint method. ", " IF nil,\n * this will default to the same library as the `main()` function in the Dart program.", "\n * @return YES if the call succeeds in creating and running a Flutter Engine instance; NO otherwise.", "\n */\n- (BOOL)runWithEntrypoint:(NSString*)entrypoint libraryURI:(NSString*)uri;\n\n/**\n * Sets the `FlutterViewController` for this instance. ", " The FlutterEngine must be\n * running (e.g. a successful call to `-runWithEntrypoint:` or `-runWithEntrypoint:libraryURI`)\n * before calling this method. ", "Callers may pass nil to remove the viewController\n * and have the engine run headless in the current process.", "\n *\n * A FlutterEngine can only have one `FlutterViewController` at a time. ", "If there is\n * already a `FlutterViewController` associated with this instance, this method will replace\n * the engine's current viewController with the newly specified one.", "\n *\n * Setting the viewController will signal the engine to start animations and drawing, and unsetting\n * it will signal the engine to stop animations and drawing. ", " However, neither will impact the state\n * of the Dart program's execution.", "\n */\n@property(nonatomic, weak) FlutterViewController* viewController;\n\n/**\n * The `FlutterMethodChannel` used for localization related platform messages, such as\n * setting the locale.", "\n */\n@property(nonatomic, readonly) FlutterMethodChannel* localizationChannel;\n/**\n * The `FlutterMethodChannel` used for navigation related platform messages.", "\n *\n * @see [Navigation\n * Channel](https://docs.flutter.io/flutter/services/SystemChannels/navigation-constant.html)\n * @see [Navigator Widget](https://docs.flutter.io/flutter/widgets/Navigator-class.html)\n */\n@property(nonatomic, readonly) FlutterMethodChannel* navigationChannel;\n\n/**\n * The `FlutterMethodChannel` used for core platform messages, such as\n * information about the screen orientation.", "\n */\n@property(nonatomic, readonly) FlutterMethodChannel* platformChannel;\n\n/**\n * The `FlutterMethodChannel` used to communicate text input events to the\n * Dart Isolate.", "\n *\n * @see [Text Input\n * Channel](https://docs.flutter.io/flutter/services/SystemChannels/textInput-constant.html)\n */\n@property(nonatomic, readonly) FlutterMethodChannel* textInputChannel;\n\n/**\n * The `FlutterBasicMessageChannel` used to communicate app lifecycle events\n * to the Dart Isolate.", "\n *\n * @see [Lifecycle\n * Channel](https://docs.flutter.io/flutter/services/SystemChannels/lifecycle-constant.html)\n */\n@property(nonatomic, readonly) FlutterBasicMessageChannel* lifecycleChannel;\n\n/**\n * The `FlutterBasicMessageChannel` used for communicating system events, such as\n * memory pressure events.", "\n *\n * @see [System\n * Channel](https://docs.flutter.io/flutter/services/SystemChannels/system-constant.html)\n */\n@property(nonatomic, readonly) FlutterBasicMessageChannel* systemChannel;\n\n/**\n * The `FlutterBasicMessageChannel` used for communicating user settings such as\n * clock format and text scale.", "\n */\n@property(nonatomic, readonly) FlutterBasicMessageChannel* settingsChannel;\n\n@end\n\n#endif // FLUTTER_FLUTTERENGINE_H_\n" ]

{ "pile_set_name": "Github" }

[ 0, 0, 0.019230769230769232, 0.0024096385542168677, 0, 0, 0, 0, 0, 0.00558659217877095, 0.006172839506172839, 0, 0.01015228426395939, 0.005263157894736842, 0, 0.011494252873563218, 0.014492753623188406, 0.017857142857142856, 0, 0, 0, 0.015625, 0, 0.007633587786259542, 0, 0.011428571428571429, 0.009900990099009901, 0.008438818565400843, 0.015625, 0, 0.012658227848101266, 0, 0.011428571428571429, 0.012345679012345678, 0.010309278350515464, 0.009900990099009901, 0.007142857142857143, 0, 0, 0, 0, 0, 0.013333333333333334, 0.010810810810810811, 0.018867924528301886, 0.019851116625310174, 0.023391812865497075, 0.016835016835016835, 0.016129032258064516, 0.013114754098360656, 0.024193548387096774 ]

[ "he remainder when 12 is divided by a?", "\n2\nSuppose -3*i - 2*i + 235 = 0. ", "Let b = i + -34. ", "What is the remainder when 62 is divided by b?", "\n10\nSuppose 2*g + 2 = 0, 4*h - 355 - 18 = -3*g. ", "What is the remainder when h is divided by 49?", "\n45\nLet h(u) = u**2 - 21*u - 86. ", "What is the remainder when 218 is divided by h(26)?", "\n42\nLet i(q) = -q**2 + 14*q - 12. ", "What is the remainder when i(9) is divided by 12?", "\n9\nSuppose 3*o + 100 - 277 = 0. ", "What is the remainder when o is divided by (-8)/(13/((-520)/16))?", "\n19\nLet z(g) = -g**3 - 2*g**2 - g - 2. ", "Let f(c) = -13*c**2 - 3. ", "Let j be f(2). ", "What is the remainder when z(-4) is divided by j/(-3) + 12/(-36)?", "\n16\nSuppose -2*i + 6*i = 188. ", "Calculate the remainder when 138 is divided by i.\n44\nSuppose -53 = -3*p - 2*q - 3*q, 3*q = 3. ", "Calculate the remainder when 111 is divided by p.\n15\nLet w(d) = -d**2 - 3*d + 4. ", "Let j be w(-4). ", "What is the remainder when (22 - (j - 1)) + 1 is divided by 13?", "\n11\nSuppose 3*r + 2*s - 23 = -2*s, 0 = -4*r + 4*s + 12. ", "Suppose t = -r + 16. ", "What is the remainder when 52 is divided by t?", "\n8\nLet i(o) = -o**2 + 14*o + 14. ", "What is the remainder when 23 is divided by i(14)?", "\n9\nLet u = 35 - -24. ", "Calculate the remainder when u is divided by 15.", "\n14\nLet o = 82 + -81. ", "Suppose 5*b + 1 = 6. ", "What is the remainder when o is divided by b?", "\n0\nCalculate the remainder when (1/(-2) + 1)*126 is divided by 16.", "\n15\nSuppose -x + 1 + 2 = 0. ", "Suppose 5*n = 15, 5*n - 12 = -x*y + 27. ", "Suppose 2*k - 12 = 34. ", "What is the remainder when k is divided by y?", "\n7\nSuppose 10*s - 7*s - 72 = 0. ", "Let f = 2 + s. What is the remainder when f is divided by 7?", "\n5\nLet w be -8 - -2 - 1/(-1). ", "Let j = w - -13. ", "Calculate the remainder when 22 is divided by j.\n6\nLet w(f) = -2*f - 2. ", "Let a(t) = -6*t - 5. ", "Let i(u) = 4*a(u) - 11*w(u). ", "What is the remainder when i(-6) is divided by 5?", "\n4\nSuppose 5*k - 3*w + 0*w = 107, k - 5*w - 39 = 0. ", "Calculate the remainder when 54 is divided by k.\n16\nLet r(s) = -4*s - 13. ", "Calculate the remainder when r(-17) is divided by 14.", "\n13\nSuppose -53 = -2*w + 9. ", "Calculate the remainder when w is divided by 4.", "\n3\nLet j(y) = -y**3 - 7*y**2 - 5*y - 3. ", "What is the remainder when j(-7) is divided by 17?", "\n15\nLet a(f) = f + 13. ", "Suppose 4*b + 17 + 27 = 0. ", "Let o be a(b). ", "Suppose o*j - 11 = j. What is the remainder when 32 is divided by j?", "\n10\nSuppose 4*s + 3*a + 27 = 8, 0 = -3*a - 15. ", "Let o = 2 - 1. ", "Calculate the remainder when 8 is divided by o/((3/(-15))/s).", "\n3\nLet m(a) = 2*a**2 - 11*a - 16. ", "What is the remainder when 34 is divided by m(7)?", "\n4\nSuppose -5 - 7 = -4*s. ", "Calculate the remainder when s is divided by (-8)/(-20) - 26/(-10).", "\n0\nLet j = -52 + 27. ", "Let n be (-30)/j*(-10)/4. ", "What is the remainder when 41 is divided by 240/18 - 2/n?", "\n13\nLet v(n) = -n**2 + 5*n - 4. ", "Suppose s + 2*s = 2*d - 1, 0 = -2*s - 4*d + 26. ", "Calculate the remainder when v(s) is divided by 2.", "\n0\nLet b be (8/6)/((-3)/9). ", "Let r be (-6)/(-8) - 1/b. What is the remainder when 4 is divided by ((-2 + r)*-2)/1?", "\n0\nLet l(i) = 4*i + 4*i**2 + 6 - 2*i**2 + 4*i - 3*i**2. ", "What is the remainder when 22 is divided by l(8)?", "\n4\nCalculate the remainder when 5 is divided by (70/21)/(6/9).", "\n0\nLet q be (2 + -5 - 2)/(-1). ", "Let f = -1 + q. Suppose -3 = -d + 7. ", "Calculate the remainder when d is divided by f.\n2\nLet z = -102 + 115. ", "What is the remainder when z is divided by 4?", "\n1\nLet z(v) = -4 + v**2 + 4 + 1 - 5 - 10*v. ", "Suppose 2*r - 3*r = -12. ", "Calculate the remainder when z(r) is divided by 11.", "\n9\nWhat is the remainder when 73 is divided by 18/24*(-40)/(-2)?", "\n13\nLet n(o) = o**3 + 8*o**2 - 4*o + 12. ", "What is the remainder when n(-8) is divided by 12?", "\n8\nLet r(z) = 2*z - 1. ", "Let g(d) = -d**2 - 4*d + 2. ", "Let m(h) = -4*g(h) - 9*r(h). ", "Calculate the remainder when m(-3) is divided by 15.", "\n13\nSuppose -3*m + 5*p - 75 = -5*m, 3*m - p - 87 = 0. ", "Calculate the remainder when m is divided by 18.", "\n12\nSuppose 9*d = 3*d + 282. ", "What is the remainder when d is divided by 12?", "\n11\nSuppose -3*b + 3*k = -6*b + 435, -5*b + 4*k = -698. ", "Calculate the remainder when b is divided by 24.", "\n22\nLet h be 1 + (-2)/(-1) + -2. ", "Let j = 3 + h. Calculate the remainder when j is divided by (-1 + 2)/1*3.", "\n1\nLet k(y) = 25*y**3 - 3*y**2 - y + 2. ", "Calculate the remainder when 67 is divided by k(1).", "\n21\nLet r(o) = o**2 - 3*o - 1. ", "Let l be ((-8)/10)/((-1)/5). ", "Calculate the remainder when 8 is divided by r(l).", "\n2\nLet w(f) = f + 10. ", "Suppose -2*x - 29 = 5*b - 0*b, 0 = 2*b - 4*x + 2. ", "Let u(l) = 2*l**2 + 2*l + 6. ", "What is the remainder when u(b) is divided by w(6)?", "\n14\nLet u = 40 - 23. ", "What is the remainder when 32 is divided by u?", "\n15\nLet o(f) = -2*f + 0*f + 4 - f. Calculate the remainder when o(-10) is divided by 12.", "\n10\nLet y = -22 - -39. ", "Let i = 27 - y. What is the remainder when i is divided by 6?", "\n4\nLet s be 2 - 0/(-1) - -59. ", "Let y = 14 + s. What is the remainder when (-2)/((-231)/y + 3) is divided by 13?", "\n12\nWhat is the remainder when ((-5)/(-20))/((-1)/(-12)) is divided by 2?", "\n1\nLet i be (-1)/(-4 - (-1 + -2)). ", "What is the remainder when 28 is divided by 10*(-3 + i + 3)?", "\n8\nSuppose -3*b = -2*b - 14. ", "Calculate the remainder when 5/1*10/2 is divided by b.\n11\nSuppose -3*z + 16 = z. Suppose -4*v + 23 = -z*s - 21, -4*s = 4*v - 36. ", "What is the remainder when v is divided by 3?", "\n1\nSuppose 97 = 5*i + 82. ", "Let c = 5 + -3. ", "Suppose c*x + x - 12 = 3*y, 5*x - 22 = 3*y. ", "Calculate the remainder when x is divided by i.\n2\nSuppose -19 = u - 130. ", "Suppose -h + 4*h = u. What is the remainder when h is divided by 19?", "\n18\nSuppose 2*b + 0*b - 2 = 0. ", "Suppose -6*v - b = 4*y - v, 19 = 4*y + v. Let a(m) = m**2 - m - 2. ", "Calculate the remainder when y is divided by a(-2).", "\n2\nLet d(l) = -l + 2. ", "Let n be d(0). ", "Suppose -w - n*w = 5*p - 31, 3*w = -p + 23. ", "Calculate the remainder when (-2)/12 + (-121)/(-6) is divided by w.\n6\nSuppose -4*v + 43 = -5*a, -a + 37 = v + 3*a. ", "What is the remainder when 64 is divided by v?", "\n13\nLet h(m) = m**2 - 3*m - 7. ", "Calculate the remainder when h(8) is divided by 9.", "\n6\nSuppose 12 = 2*n + 2*d - 0*d, n - 12 = -3*d. ", "Suppose n*p = -c + 26, 4*c = 4*p - 1 - 55. ", "What is the remainder when p is divided by 3?", "\n1\nLet i be (-2)/12 + (-618)/(-36). ", "Let w = i - -1. ", "What is the remainder when w is divided by 5?", "\n3\nLet u(m) = -m**3 - 8*m**2 + 7*m + 5. ", "Suppose -13 = 2*j + 5. ", "Calculate the remainder when u(j) is divided by 6.", "\n5\nLet m = 51 + -38. ", "Calculate the remainder when 46 is divided by m.\n7\nLet r be 1/((-8)/(-10))*4. ", "Suppose r*x + 4 = 59. ", "Let y = -116 + 147. ", "Calculate the remainder when y is divided by x.\n9\nSuppose -c - 37 + 232 = 0. ", "Calculate the remainder when c is divided by 22.", "\n19\nSuppose 3*p + 2 = -u - 4*u, -4*p + 4*u + 40 = 0. ", "Let m = p - 4. ", "Calculate the remainder when 23 is divided by (24/(-9))/(m/(-9)).", "\n11\nLet u(x) = -5*x - 9. ", "Let f be -7 - 0 - (-6)/6. ", "Calculate the remainder when u(f) is divided by 11.", "\n10\nLet m = 5 - 3. ", "Let t = 0 + m. Suppose 3 = -d, 3*k - 5*k + 8 = -t*d. ", "Calculate the remainder when k is divided by 1.", "\n0\nSuppose -2*i + 4*r = -24, -8*r + 5*r - 38 = -4*i. ", "Let h = 13 + -8. ", "What is the remainder when i is divided by h?", "\n3\nLet d(m) = -m**3 + 3*m**2 + 19*m + 8. ", "Let r = 84 + -44. ", "What is the remainder when r is divided by d(6)?", "\n12\nLet p(w) = -7*w - 3. ", "Suppose -7*m = 2*f - 2*m - 197, 4*m = 4. ", "Suppose -3*s + f = s. Calculate the remainder when p(-7) is divided by s.\n22\nSuppose a + 5*u = 5, 6*a = 2*a - 5*u + 5. ", "Let j(y) = 2 - 3 + 2*y**3 + a. What is the remainder when j(1) is divided by 1?", "\n0\nLet v = -76 + 93. ", "Calculate the remainder when 100 is divided by v.\n15\nSuppose -2*s + 20 + 74 = 0. ", "Calculate the remainder when s is divided by 10.", "\n7\nSuppose 0*s - 4*s - 48 = 0. ", "Calculate the remainder when 55 is divided by 29 + 1 - s/(-6).", "\n27\nLet n = 52 + -29. ", "What is the remainder when n is divided by 8?", "\n7\nSuppose 0 = 2*c - 0*c - 32. ", "Let z = -1 - -3. ", "Suppose 3*s - z - 25 = 0. ", "What is the remainder when c is divided by s?", "\n7\nSuppose 1518 = 30*o + 198. ", "Let a(z) = -z**2 + 10*z - 1. ", "What is the remainder when o is divided by a(6)?", "\n21\nWhat is the remainder when 22 is divided by 10/(-40) - 66/(-8)?", "\n6\nLet q = 33 + -20. ", "What is the remainder when 36 is divided by q?", "\n10\nLet l(y) = -y**3 - 7*y**2 - 7*y - 10. ", "Let n be (16/14)/(2/7). ", "Suppose -5*j - 12 - 27 = r, 0 = n*r + 16. ", "Calculate the remainder when l(j) is divided by 10.", "\n9\nSuppose -7 + 241 = 3*w. ", "Calculate the remainder when w is divided by 16.", "\n14\nS" ]

{ "pile_set_name": "DM Mathematics" }

[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.02040816326530612, 0, 0, 0, 0, 0, 0, 0, 0.010638297872340425, 0, 0, 0, 0.017857142857142856, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.018867924528301886, 0.03571428571428571, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.02040816326530612, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.02040816326530612, 0, 0, 0.02702702702702703, 0, 0, 0, 0.04, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.0136986301369863, 0, 0.0196078431372549, 0, 0, 0, 0, 0.02, 0, 0.0196078431372549, 0, 0, 0, 0, 0, 0, 0, 0, 0.02857142857142857, 0, 0, 0, 0, 0, 0, 0.022727272727272728, 0.0136986301369863, 0, 0, 0, 0, 0, 0, 0, 0.008695652173913044, 0, 0, 0.02, 0, 0.023255813953488372, 0, 0, 0, 0, 0, 0, 0.02, 0, 0, 0.045454545454545456, 0, 0.012987012987012988, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.020833333333333332, 0, 0, 0, 0.012658227848101266, 0, 0, 0, 0, 0, 0, 0, 0, 0.058823529411764705, 0, 0, 0, 0, 0.020833333333333332, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]

[ "Computer scientist Zoran Popović is teaching algebra to children with an adaptive—\n\nand addictive—new videogame. ", "Mike Clinard. ", "GlueKit\n\nYou have an armadillo, a black-and-white trout, and a rainbow trout. ", "Can the animals fit inside the box? ", "No, this isn’t an intolerable ­riddle. ", "For Zoran Popović, it’s how you teach math to kids. ", "A computer scientist at the University of Washington, Popović first became known for his popular online game, Foldit, which challenges players to create intricate protein patterns by bending and rearranging amino acids—the constituent units of proteins—into new shapes. ", "Players win by building better molecules. ", "To Popović’s surprise, non-­scientist gamers developed more-­complex proteins than biochemists did, and these new proteins may lead to the discovery of new drugs or organic materials.", "\n\nOnce students learn the basic principals of algebra, the game introduces numbers and signs.", "\n\nCourtesy of UW Center for Game Science/WeWantToKnow\n\nDragonBox Adaptive draws on data from players' past games to adapt to their unique pace.", "\n\nCourtesy of UW Center for Game Science/WeWantToKnow\n\nNow Popović wants to put the power of games to an even better use. ", "Earlier this year he adapted DragonBox, a Norwegian game app that introduces algebraic concepts with ­animal-faced cards, then builds up to numbers and computational signs. ", "Kids are given a mix of cards and a box. ", "They follow simple rules to get rid of all the unnecessary cards. (", "Color cards cancel black-and-white ones, like positive and negative numbers.) ", "Eventually the player is left with a box on one side of the screen and an irreducible number of cards on the other—the equivalent of solving for x. As harder concepts are introduced, students who need more time on a level get additional problems; those who understand it move on. ", "In an experiment with DragonBox Adaptive in Washington state, an average of 93 percent of K–12 students successfully mastered concepts after only 90 minutes of gameplay, and they didn’t want to stop.", "\n\nPopović is using this method as the basis for an entire sixth-grade math curriculum in trial now in ­Seattle and next year in Brooklyn and Brazil. ", "He’s banking on digital tablets to help teachers adjust their lessons to individual performance. ", "Science, language arts, even coding are all on the table for the future. ", "Finally, an argument for more screen time for kids." ]

{ "pile_set_name": "OpenWebText2" }

[ 0.008849557522123894, 0.07142857142857142, 0, 0, 0, 0.019230769230769232, 0.007407407407407408, 0, 0.00546448087431694, 0, 0, 0, 0.005780346820809248, 0, 0, 0, 0, 0.005025125628140704, 0.006711409395973154, 0, 0, 0 ]

[ "Role of hypothalamic neuropeptide Y in feeding and obesity.", "\nThe 36-amino-acid peptide, neuropeptide Y (NPY), is the most abundant peptide in the rat brain. ", "When administered into the brain, NPY produces a variety of physiological actions including a pronounced stimulation of feeding in satiated rats. ", "Elevations in hypothalamic NPY have been reported after food deprivation and in genetically obese rodents. ", "NPY is believed to produce its actions through a portfolio of G-protein coupled receptors, Y1, Y2, Y4 and Y5. ", "Studies using peptide analogs, receptor knockout animals and specific receptor antagonists suggest the Y1 and Y5 receptors are important in mediating the effects of NPY on food intake in rats. ", "Development of specific receptor antagonists with improved pharmacokinetic properties will be required to determine the importance of NPY in human obesity and appetite disorders." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0.010309278350515464, 0.00684931506849315, 0.009345794392523364, 0.03636363636363636, 0.010362694300518135, 0.0056179775280898875 ]

[ "Q:\n\nHow can I create a layout with tabs completely in XML?", "\n\nI'm trying to make tabs in a layout. ", "I've found a lot of examples and tutorials using TabWidget, TabHost, but they all involve one of the following:\n\nJava code in the activity\nSeparate activities for each tabs\nSeparate fragments for each tabs\n\nThe content inside the tabs is static, so I should just be able to include everything in the layout, in pure XML.", "\nAnyway to do that?", "\n\nA:\n\nThe simple answer, no. ", " You have to setup your TabHost in Java code and create your tabs. ", " You can have static layouts for the tabs without using fragments but it still requires setup in Java.", "\nIf you don't do this setup in code, your TabWidget won't know which layout corresponds to which tab and wouldn't be able to function. ", " You're going to have to write a bit of code.", "\nThe code for doing this is really simple though.", "\nThe XML (placed inside your layout where you want it):\n<TabHost android:id=\"@+id/tab_host\"\n android:layout_width=\"fill_parent\"\n android:layout_height=\"wrap_content\" >\n <LinearLayout android:layout_width=\"fill_parent\"\n android:layout_height=\"wrap_content\"\n android:orientation=\"vertical\">\n <TabWidget android:id=\"@android:id/tabs\"\n android:layout_width=\"fill_parent\"\n android:layout_height=\"wrap_content\" />\n <FrameLayout android:id=\"@android:id/tabcontent\"\n android:layout_width=\"fill_parent\"\n android:layout_height=\"wrap_content\">\n <LinearLayout\n android:id=\"@+id/tab_one_container\"\n android:layout_width=\"fill_parent\"\n android:layout_height=\"wrap_content\" />\n <LinearLayout\n android:id=\"@+id/tab_two_container\"\n android:layout_width=\"fill_parent\"\n android:layout_height=\"wrap_content\" />\n </FrameLayout>\n </LinearLayout>\n</TabHost>\n\nThe Java code (placed wherever you setup your layout):\nTabHost host = (TabHost)findViewById(R.id.tab_host);\nhost.setup();\n\nTabSpec spec = host.newTabSpec(\"Tab One\");\nspec.setContent(R.id.tab_one_container);\nspec.setIndicator(\"Tab One\");\nhost.addTab(spec);\n\nspec = host.newTabSpec(\"Tab Two\");\nspec.setContent(R.id.tab_two_container);\nspec.setIndicator(\"Tab Two\");\nhost.addTab(spec);\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0, 0, 0.009375, 0, 0, 0.014925373134328358, 0.00980392156862745, 0.007407407407407408, 0, 0, 0.011251758087201125 ]

[ "Posts\n\nHow to get rid of small acne on the face\n\nEven the healthiest skin is not immune from minor defects: irregularities, wrinkles and, of course, small, but such harmful pimples. ", "Moreover, in an age of high technology and bad ecology, absolutely healthy skin is a particularly rare phenomenon. ", "But do not get upset because of such trifles, because there are many effective remedies for acne. ", "It remains only to find something that is right for you.", "\n\nYou will need\n\n– water;\n\n– soda;\n\n– sugar\n\nor\n\n– water;\n\n– mint leaves;\n\n– boric alcohol;\n\n– tincture of calendula;\n\n– lemon\n\nor\n\n– raspberry leaves\n\nor\n\n– plantain\n\nInstructions\n\n1\n\nTake one glass of boiling water and add one teaspoon of soda and sugar. ", "Stir when soda and sugar dissolve, dab a cotton pad or swab in it and carefully wipe the skin in places where there are pimples. ", "Then wash off with slightly cool water with household or tar мылом. ", "Теперь места, протертые ранее приготовленным раствором, смажьте маслом, лучше сливочным, чтобы снять покраснение и возможную сухость. ", "Через час снова умойтесь, но уже без мыла.", "\n\nRemedy for raspberry leaves Rinse raspberry leaves thoroughly. ", "Squeeze out the juice from them, scalding them with boiling water. ", "The resulting juice rub with butter until a thick creamy mass is obtained. ", "Lubricate the problem areas of the skin. ", "It is necessary to leave the ointment for the whole night.", "\n\n4\n\nMedicinal herbs are also a good helper in the fight against acne. ", "To prepare such a tool you need to take one tsp. ", "dry chamomile flowers, mint leaves and green tea. ", "All herbs pour a glass of boiling water. ", "Leave it to cool. ", "With this infusion, wipe the skin on Facial (especially problem areas) in the morning and evening. ", "Keep the infusion in the refrigerator." ]

{ "pile_set_name": "Pile-CC" }

[ 0, 0, 0, 0, 0, 0, 0, 0.022388059701492536, 0, 0.03076923076923077, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.010101010101010102, 0 ]

[ "Q:\n\nAS 3 Flash Countdown Timer For Game Over Screen\n\nI am currently in a Flash game programming class with action script 3 and can't seem to find a clock anywhere that counts down and then does an action. ", "I've been using Lynda tutorials and that hasn't been helpful and I have also Google searched quite a bit, but no luck. ", "I have a Countdown clock and have been try \"if\" statements, but no luck.", "\nCan someone guide me in the right direction on what I am doing wrong?", "\n //game timer\n\n var count:Number = 60; // amount of time \n\n var myTimer:Timer = new Timer(1000,count); // time in ms, count is calling from line above\n\n myTimer.addEventListener(TimerEvent.", "TIMER, countdown);\n\n myTimer.start();\n\n function countdown(event:TimerEvent):void\n {\n myText_txt.text = String((count)-myTimer.currentCount); //dynamic txt box shows current count \n }\n\n //if and else if statements\n\n if (((count)-myTimer.currentCount) == 58)\n {\n gotoAndStop(2);\n }\n\nA:\n\nThis should help ;) Without any magic numbers.", "\nvar myTimer:Timer = new Timer(1000,60); // every second for 60 seconds\nmyTimer.addEventListener(TimerEvent.", "TIMER, onTimer);\nmyTimer.addEventListener(TimerEvent.", "TIMER_COMPLETE, onComplete);\nmyTimer.start();\n\nfunction onTimer(e: TimerEvent):void {\n myText_txt.text = String(myTimer.repeatCount - myTimer.currentCount);\n}\n\nfunction onComplete(e: TimerEvent):void{\n gotoAndStop(2);\n}\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0.00975609756097561, 0.008403361344537815, 0, 0, 0, 0.0029411764705882353, 0.009259259259259259, 0.018867924528301886, 0 ]

[ "Toronto External Contracts Inquiry\n\nThe Toronto External Contracts Inquiry is a judicial inquiry into allegations of conflict of interest around contracts entered into by Toronto, Ontario's municipal government in 1998 and 1999.", "\n\nIt is a follow up to the city's earlier Computer Leasing Inquiry, and is investigating three other questionable contracts. ", "However, as with the earlier inquiry, the External Contracts Inquiry is commissioned by Madam Justice Denise Bellamy of the Ontario Superior Court of Justice.", "\n\nThe External Contracts Inquiry was established by Toronto City Council on October 1, 2002, as an expansion of the Computer Leasing Inquiry. ", "It began hearings on October 18, 2004, after the Computer Leasing Inquiry ended.", "\n\nTMACS\nThe first contract investigated by the External Contracts Inquiry involved the city's tax and billing management system.", "\n\nOn January 1, 1998, the city of Toronto and its five suburban cities (Scarborough, North York, Etobicoke, East York and York) were amalgamated into the single \"megacity\" of Toronto.", "\n\nDuring the amalgamation process, the six municipal governments had to merge their computer systems. ", "Five of the municipalities used a system called TXM, while North York used a system called TMACS. ", "North York's treasurer, Wanda Liczyk, was the sole voice in favour of adopting TMACS for the amalgamated city, and when she subsequently became treasurer of the \"megacity\", she implemented TMACS without consulting city council.", "\n\nTMACS was developed for North York's city government by two American software consultants, David Maxson and Michael Saunders. ", "Testimony before the inquiry has confirmed that from 1989 to 1991, Saunders and Liczyk had a personal romantic relationship.", "\n\nAlthough their relationship ended before Saunders was contracted by the city of Toronto, city staff have testified that Liczyk and Saunders still had a close friendship. ", "In her testimony to the inquiry, Liczyk has acknowledged that her relationship with Saunders may have created the appearance of favouritism in her decisions involving Saunders. ", "However, in response to suggestions that she should have declared a conflict of interest and recused herself from such decisions, Liczyk has testified that she is able to separate personal and business matters.", "\n\nIt has also been testified that Saunders was often rude and abrasive to city staff.", "\n\nOther Contracts\nTwo other city contracts will be investigated by the Inquiry. ", "Testimony around these contracts has not yet been presented.", "\n\nReferences\n\nExternal links\nToronto Computer Leasing Inquiry, and The Toronto External Contacts Inquiry\n\nExternal Contracts Inquiry\nCategory:Ontario political scandals" ]

{ "pile_set_name": "Wikipedia (en)" }

[ 0.0043859649122807015, 0, 0.02531645569620253, 0.014084507042253521, 0.0125, 0.0078125, 0.00546448087431694, 0, 0.02040816326530612, 0.013215859030837005, 0.015625, 0.008064516129032258, 0.011627906976744186, 0.011299435028248588, 0, 0.011764705882352941, 0, 0, 0.011904761904761904 ]

[ "(CBS News) In the first debate between the Massachusetts Senate candidates, challenger Elizabeth Warren attempted to paint incumbent Sen. Scott Brown as a corporate shill in lock-step with his Republican colleagues concerned about protecting the wealthy, while the senator framed Warren as liberal intent on raising taxes.", "\n\n\"She's obsessed with raising taxes,\" Brown said during Thursday night's debate in Boston, adding that \"there's a spending problem in Washington, D.C.\" and that Warren would continue to fuel the fire.", "\n\n\"She would have raised taxes on you,\" he said, pointing to the camera, in defense of voting against a bill in the Senate that would have reduced the interest rates of student loans by raising taxes on high-income earners. \"", "I am on the taxpayers' side.\"", "\n\nRepeating the position of national Democrats, Warren said Washington needs to take \"a balanced approach\" to reducing the deficit by cutting spending and raising revenue. ", "She also said that Brown has \"rigged the playing field\" by voting to extend subsidies for oil companies.", "\n\nBrown painted himself as a moderate Republican who has worked on issues important to the late Sen. Ted Kennedy, including preventing wind farms off the coast of Nantucket, and has voted with Democrats. ", "He said numerous times that he is the \"second most bipartisan senator in the Senate.\"", "\n\nHe used the line when Warren, a Harvard professor, said control of the Senate could hang on the Massachusetts Senate race. ", "She said Sen. James Inhofe, R-Okla., who believes global warming is a hoax, could lead the committee that overseas environmental policy if Republicans take control of the Senate.", "\n\nIn the first debate of one of the most closely watched Senate races in the country, which has been tight for many months, both Warren and Brown lobbed attacks while playing defense much of the one-hour face-off.", "\n\n\"Sen. Brown can say all he wants but he has voted,\" Warren said, continuously pointing to votes Brown has taken in the Senate.", "\n\n\"Women need someone they can depend on, not some of the time but all of the time,\" she added, charging Brown with voting against a bill that would require women make equal amounts of money as their male counterparts and co-sponsoring a measure, known as the Blunt amendment after the Missouri senator, that would have gutted President Obama's mandate that employers provide free contraception health coverage.", "\n\n\"You should stop scarring women,\" Brown shot back, adding that both he and Warren are both pro-choice and that he's been \"fighting for women's rights\" since he was a child as his mother was abused by his step-father.", "\n\nThe debate debate opened with both candidates on the attack. ", "Brown said Warren should release her personal files from Harvard, where she listed her ethnicity as Native American, to prove that she did not receive an advantage for being a minority.", "\n\n\"The only way to get this straight is for you to release your personnel records,\" he said, saying it \"speaks volumes\" that she has not done so.", "\n\nBrown also said the cost of higher education has skyrocketed, in part, because of administrative costs, pointing to Warren's several-hundred thousand dollar salary there \"in addition the housing and other perks,\" he said. \"", "No wonder the costs are high.\"", "\n\nMeantime, Warren said Brown would \"hold the other 98 percent of families hostage if there is not tax breaks for the top.\"", "\n\nBrown did not utter the words of the Republican presidential candidate Mitt Romney, the former governor of the state, who has no chance of winning the liberal state in November.", "\n\nWarren invoked President Obama's name, however, when the debate turned to foreign policy, an issue in which polls show voters approve of the president's handling.", "\n\n\"I support President Obama,\" she said, arguing that the president has \"done a first-rate job\" in using sanctions and diplomacy and leaving all options of the table to prevent Iran from obtaining a nuclear weapon. \"", "I'm still working to have President Obama be the commander-in-chief, not Mitt Romney.\"", "\n\nEmphasizing his independence from the Republican Party, Brown praised Secretary of State Hillary Clinton as \"a bright star in that administration,\" but said the president started strong but has gotten weak on Iran, adding that \"we cannot have a nuanced approach.\"" ]

{ "pile_set_name": "Pile-CC" }

[ 0.012422360248447204, 0.004975124378109453, 0.0044444444444444444, 0, 0, 0.009615384615384616, 0.00980392156862745, 0.011764705882352941, 0.032, 0.011235955056179775, 0.014084507042253521, 0.0234375, 0.0072992700729927005, 0.0045871559633027525, 0, 0.016216216216216217, 0, 0.008888888888888889, 0, 0.016260162601626018, 0.0111731843575419, 0.006097560975609756, 0.004629629629629629, 0.023255813953488372, 0.01509433962264151 ]

[ "Q:\n\nWrite an output file with a user defined name in C\n\nI am new to C and am trying to define an output filename before the program runs. ", "I am getting a Bus error\nHere is my code:\n#include <stdio.h>\n\nint main (int argc, const char * argv[]) {\n char fname[128];\n printf(\"Enter the file name\\n\");\n scanf(\"%123s\",fname);\n strcat(\"/Users/user/Desktop/learn/\", fname);\n strcat(fname, \".txt\");\n\n FILE *fp;\n fp=fopen(fname,\"a\");\n fprintf(fp, \"Testing... OK I think it worked\\n\");\n return 0;\n}\n\nA:\n\nYou didn't #include <string.h> for strcat.", "\nThe first argument to strcat must be a pointer, not a string literal.", "\nstrcat itself isn't safe, use strncat instead.", "\nDon't forget to check the result of scanf and fopen.", "\nAnd close fp when you're done with it.", "\nThe signature of main should be int main(int argc, char * argv[]).", "\n\nThe use of scanf is also generally discouraged, use fscanf & sscanf instead.", "\n\n" ]

{ "pile_set_name": "StackExchange" }

[ 0, 0, 0, 0, 0, 0, 0, 0, 0 ]

[ "Planning safe anesthesia: the role of collective resources management.", "\nAnticipation and planning are essential steps of risk management but the mechanisms of planning behavior are incompletely understood, especially the factors including collective work. ", "The aim of this research is to understand how anesthetists plan safe solution to perform anesthesia. ", "A study based on interviews was conducted in two French hospitals. ", "Data processing focused on the main decisions made by 20 anesthetists during two simulated pre-anesthetic consultation. ", "The main decisions made have been identified and the decision criteria have been analyzed. ", "To ensure patient safety, all anesthetists do not plan the same solution. ", "The rejection or the selection of solutions by each physician rests on two types of criteria: the assessment of risks for the patient and the assessment of resources available to handle the situation. ", "For the latter, the knowledge on the individual skills of each and the adoption of \"local benchmark practices\" play an essential role. ", "Ultra safe performance in highly variable systems cannot be achieved only through standardization but also through the possibility and ability of the subjects to adapt their practices to their own skills and to that of their colleagues. ", "The conditions for the development of this \"adaptative safety\" are discussed." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ]

[ "Even if they privately agree with him CCHQ strategists don’t appreciate the media interventions by Matthew Parris. ", "Lynton Crosby has decreed that UKIP is to be downplayed and all but ignored insomuch as that is possible. ", "The Clacton bullet is to be bitten and the pain taken without too much complaint or squealing. ", "Hence the lack of Tory MPs in the studios stoking up the row or giving biting anti-Carswell soundbites to be lapped up by the media and inserted in stories of Tory splits and infighting. ", "So in their place we have Matthew Parris, a former Tory MP and the arch voice of the Tory left, on the Daily Politics, Newsnight and in his columns he is stoking up the euro-row, amplifying it and providing soundbites, inventing motives and attributing malice to the eurosceptic wing of the Tories. ", "Matthew’s latest column in The Times paints Clacton as full of people so infra dig that they should just be ignored by the Tories. ", "The column is so patronising and sneering that it could be reprinted by UKIP and delivered as a recruiting leaflet to every former Tory voter in Clacton…" ]

{ "pile_set_name": "Pile-CC" }

[ 0.008695652173913044, 0.009433962264150943, 0.010526315789473684, 0, 0.006688963210702341, 0.007633587786259542, 0 ]

[ "Expression and significance of plasma 3-NT and ox-LDL in patients with Alzheimer's disease.", "\nTo examine the expression and clinical significance of plasma 3-nitrotyrosine (3-NT) and oxidized low-density lipoprotein (ox-LDL) levels in patients with Alzheimer's disease (AD), we examined 48 AD patients and 37 healthy control subjects. ", "The Mini-Mental State Examination, Activities of Daily Living Scale, and Hachinski Ischemic Scale were examined in all subjects. ", "AD patients were classified using the Global Deterioration Scale. ", "The concentrations of plasma 3-NT and ox-LDL were detected using an enzyme-linked immunosorbent assay. ", "We found that the plasma 3-NT concentration in the AD group (119.46 ± 21.82 nM) was significantly higher than that in the control group (55.09 ± 9.63 nM) (P < 0.05). ", "Spearman analysis showed that plasma 3-NT level was negatively associated with the Mini-Mental State Examination results of AD patients. ", "Plasma ox-LDL level in the AD group (112.25 ± 17.81 mg/L) was significantly higher than that in the control group (47.46 ± 10.04 mg/L) (P < 0.05). ", "Spearman analysis showed that plasma ox-LDL level was positively correlated with AD severity in AD patients. ", "However, plasma 3-NT level in the AD group was not associated with plasma ox-LDL level. ", "Therefore, plasma 3-NT and ox-LDL levels in AD patients were significantly increased, which may be related to the degree of AD severity in AD patients." ]

{ "pile_set_name": "PubMed Abstracts" }

[ 0, 0, 0.007751937984496124, 0.015151515151515152, 0, 0, 0.0072992700729927005, 0.006802721088435374, 0, 0, 0 ]