File size: 21,813 Bytes
6fa4bc9 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 |
{
"paper_id": "M92-1011",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T03:13:17.082042Z"
},
"title": "LANGUAGE SYSTEMS, INC . MUC-4 TEST RESULTS AND ANALYSIS 1",
"authors": [
{
"first": "Christine",
"middle": [
"A"
],
"last": "Montgomery",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Language Systems, Inc",
"location": {
"addrLine": "6269 Variel Avenue, Suite F Woodland Hills",
"postCode": "91367",
"region": "CA"
}
},
"email": ""
},
{
"first": "Bonnie",
"middle": [
"Glover"
],
"last": "Stalls",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Language Systems, Inc",
"location": {
"addrLine": "6269 Variel Avenue, Suite F Woodland Hills",
"postCode": "91367",
"region": "CA"
}
},
"email": ""
},
{
"first": "Robert",
"middle": [
"R"
],
"last": "Stumberger",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Language Systems, Inc",
"location": {
"addrLine": "6269 Variel Avenue, Suite F Woodland Hills",
"postCode": "91367",
"region": "CA"
}
},
"email": ""
},
{
"first": "Naicong",
"middle": [],
"last": "Li",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Language Systems, Inc",
"location": {
"addrLine": "6269 Variel Avenue, Suite F Woodland Hills",
"postCode": "91367",
"region": "CA"
}
},
"email": ""
},
{
"first": "Robert",
"middle": [
"S"
],
"last": "Belvin",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Language Systems, Inc",
"location": {
"addrLine": "6269 Variel Avenue, Suite F Woodland Hills",
"postCode": "91367",
"region": "CA"
}
},
"email": ""
},
{
"first": "Alfredo",
"middle": [],
"last": "Arnai",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Language Systems, Inc",
"location": {
"addrLine": "6269 Variel Avenue, Suite F Woodland Hills",
"postCode": "91367",
"region": "CA"
}
},
"email": ""
},
{
"first": "Susan",
"middle": [
"B"
],
"last": "Hirsh",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Language Systems, Inc",
"location": {
"addrLine": "6269 Variel Avenue, Suite F Woodland Hills",
"postCode": "91367",
"region": "CA"
}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "LSI's overall natural language processing (NLP) objective is the development of a broad coverage, reusable system which is readily transportable to additional domains, applications, and sublanguages in English, as well as providing a foundation for our multilingual work. Our system, called DBG, for Data Base Generator, is comprised of a set of NLP components which have been developed, extended, and rebuilt over a period of some years. The core of the system is an innovative Principle-based parser, using ideas from [1], which we began developing in the course of MUC-3 to replace our previous chart parser. Our approach thus relies on the concept of powerful, robust parsing as the most crucial component in an NLP system. In applying our NLP system to text extraction, our ultimate objective is to develop a high quality text extraction system, where \"high quality \" is defined as scoring above 80%-a number well beyond any current MUC scores. In line with these NLP objectives, our major focus for MUC-4 was a follow-up to our main \"lesson learned\" i n MUC-3, which was to acquire a machine-readable dictionary (MRD) and integrate its content into the DBG system. When attempts to acquire the computer-friendly Longmans or one of the Oxford Dictionaries were unsuccessful, we turned to ACL's CD-ROM containing the Collins English Dictionary. The most correct version of the CED on the ACL CD-ROM was apparently developed directly from a medium prepared for the typographer , and unfortunately lacks any documentation of features, fonts, language, etc. The effort of acquiring an d integrating the CED was clearly a worthwhile endeavor, since we were able to increase the number of entries i n our lexicon threefold in a relatively short time (see Table 1). The increase in lexicon size will benefit all th e applications LSI is currently working on .",
"pdf_parse": {
"paper_id": "M92-1011",
"_pdf_hash": "",
"abstract": [
{
"text": "LSI's overall natural language processing (NLP) objective is the development of a broad coverage, reusable system which is readily transportable to additional domains, applications, and sublanguages in English, as well as providing a foundation for our multilingual work. Our system, called DBG, for Data Base Generator, is comprised of a set of NLP components which have been developed, extended, and rebuilt over a period of some years. The core of the system is an innovative Principle-based parser, using ideas from [1], which we began developing in the course of MUC-3 to replace our previous chart parser. Our approach thus relies on the concept of powerful, robust parsing as the most crucial component in an NLP system. In applying our NLP system to text extraction, our ultimate objective is to develop a high quality text extraction system, where \"high quality \" is defined as scoring above 80%-a number well beyond any current MUC scores. In line with these NLP objectives, our major focus for MUC-4 was a follow-up to our main \"lesson learned\" i n MUC-3, which was to acquire a machine-readable dictionary (MRD) and integrate its content into the DBG system. When attempts to acquire the computer-friendly Longmans or one of the Oxford Dictionaries were unsuccessful, we turned to ACL's CD-ROM containing the Collins English Dictionary. The most correct version of the CED on the ACL CD-ROM was apparently developed directly from a medium prepared for the typographer , and unfortunately lacks any documentation of features, fonts, language, etc. The effort of acquiring an d integrating the CED was clearly a worthwhile endeavor, since we were able to increase the number of entries i n our lexicon threefold in a relatively short time (see Table 1). The increase in lexicon size will benefit all th e applications LSI is currently working on .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "The complete LSI TST3 and TST4 score reports are included in Appendix G, \"Final Test Score Summaries\" . As an indication of system development during MUC4, we can compare our TST3 results with our results on th e MUC-4 interim test (TST2) . The relevant portions of the TST3 and TST2 results are shown in Tables 2 and 3 . Figure 1 graphically presents the TST2 and TST3 recall and precision matrices . Although our overall TST3 and TST4 scores clearly fell short of our goals, there are important comparisons t o be made between TST2 and TST3 . Most importantly, our recall scores made a definite improvement, as can be seen in the TST3 REC column (vs . the TST2 REC column) and on the Recall axis in Figure 1 . This is due to improvements in the extraction of events and entities, from the text, at our knowledge representation level . (Some examples of this are given in the system summary paper in our discussion of Message 0048) . Unfortunately, our precision did not significantly improve . This is in large part due to template overgeneration, which is caused by deficiencies in our event template merging . We are not yet properly merging event references across multiple sentences .",
"cite_spans": [],
"ref_spans": [
{
"start": 305,
"end": 319,
"text": "Tables 2 and 3",
"ref_id": "TABREF2"
},
{
"start": 322,
"end": 330,
"text": "Figure 1",
"ref_id": null
},
{
"start": 701,
"end": 709,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "RESULTS",
"sec_num": null
},
{
"text": "Although improvements in both recall and precision are required, we anticipate that first solving the overgeneration problem will give us a more accurate picture of how the system is really performing in terms of recall an d precision, and where additional work will produce the most significant improvement in system performance . Figure 1 of LSI's system summary in this proceedings presents an overview of the DBG system as configure d for MUC-4 . A new module has been added at the front end to select sentences of potential interest for th e application . Work on our Principle-based parser has continued throughout the past year, extending the inventor y of syntactic structures that can currently be handled .",
"cite_spans": [],
"ref_spans": [
{
"start": 332,
"end": 340,
"text": "Figure 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "SLOT",
"sec_num": null
},
{
"text": "The major MUC-4 effort was devoted to the lexicon (approximately 35%) and to the parser (about 20%), wit h other modules getting substantially less of the total effort, as shown in Table 4 . ",
"cite_spans": [],
"ref_spans": [
{
"start": 181,
"end": 188,
"text": "Table 4",
"ref_id": null
}
],
"eq_spans": [],
"section": "ALLOCATION OF EFFOR T",
"sec_num": null
},
{
"text": "MUC is unfortunately a resource-limited undertaking for LSI; however, we did expend a significant effort on the lexicon and parser for MUC-4 . Although LSI is a small company, we were able to devote these resources t o MUC-4 in part due to the sponsorship of DARPA and BRL (see Footnote 1), and additionally, because the work was d irectly in line with our overall NLP objectives mentioned previously.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "LIMITING FACTORS",
"sec_num": null
},
{
"text": "Limiting factors included all those on the list --time, people, cpu cycles --as well as the budgetary limits mentioned above. Knowledge was also a limiting factor in the sense that portions of the knowledge embedded i n the system were not exploited, and other crucial knowledge was not added, due to resource limitations .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "LIMITING FACTORS",
"sec_num": null
},
{
"text": "On the other hand, the amount of knowledge represented in the expanded lexicon is significant, so significan t achievements are possible if limited resources are focused on particular problem areas .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "LIMITING FACTORS",
"sec_num": null
},
{
"text": "During our preparation for MUC-4 testing, we were able to use the entire development corpus this year, an d found it extremely valuable in our system development .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "TRAINING",
"sec_num": null
},
{
"text": "The code for our Lexical Unexpected Inputs/Word Acquisition Module (LUX/WAM), which deals with erroneous (e .g ., misspelled) or new words is still the one which has gone for the longest period of time without rewriting or optimization of any kind. However, with our new, much larger lexicon, LUX/WAM was invoked far les s frequently than during MUC-3 processing, and so was not really a significant factor in MUC-4 .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "MODULE MOST OVERDUE FOR REWRITIN G",
"sec_num": null
},
{
"text": "A second module mentioned last year as a candidate for rewriting was LXI, the lexical lookup component . Some modification of LXI code was carried out to provide more efficient processing for MUC-4 .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "MODULE MOST OVERDUE FOR REWRITIN G",
"sec_num": null
},
{
"text": "Throughout LSI's MUC participation, our goal has been to exploit this opportunity to achieve a generic, broa d coverage, text extraction capability . To this end, with the exception of specific MUC-oriented parameters suc h as the names of critical events, the DBG system as configured for MUC is completely reusable in another application (and is in fact being used for all other NLP projects currently in house, including the NLP component of our voice translation testbed for English-->Spanish-->English). For example, the new sentence selection modul e added this year can be used to search any text; only the tables containing MUC-oriented words that indicate critical event content are MUC-specific .",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "REUSABILITY",
"sec_num": null
}
],
"back_matter": [],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Principle-Based Parsing",
"authors": [
{
"first": "R",
"middle": [
"C"
],
"last": "Berwick",
"suffix": ""
}
],
"year": 1987,
"venue": "AI TR",
"volume": "972",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Berwick, R . C ., Principle-Based Parsing, AI TR No, . 972, June, 1987 .",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "MUC-3 Test Results an d Analysis",
"authors": [
{
"first": "C",
"middle": [
"A"
],
"last": "Montgomery",
"suffix": ""
},
{
"first": "B",
"middle": [
"G"
],
"last": "Stalls",
"suffix": ""
},
{
"first": "R",
"middle": [
"S"
],
"last": "Belvin",
"suffix": ""
},
{
"first": "R",
"middle": [
"E"
],
"last": "Stumberger",
"suffix": ""
}
],
"year": 1991,
"venue": "Proceedings of the Third Message Understanding Conference",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Montgomery, C . A., Stalls, B . G ., Belvin, R. S., and Stumberger, R . E. MUC-3 Test Results an d Analysis. Proceedings of the Third Message Understanding Conference. Morgan Kaufmann Publishers , San Mateo, May, 1991 .",
"links": null
}
},
"ref_entries": {
"TABREF0": {
"html": null,
"type_str": "table",
"num": null,
"content": "<table><tr><td/><td>MUC3</td><td>MUC4</td></tr><tr><td>STEMS</td><td/><td>15,285</td></tr><tr><td>INFLECTED FORMS</td><td/><td>14,56 1</td></tr><tr><td>TOTALS</td><td>est at 10,000</td><td>29,846</td></tr><tr><td colspan=\"2\">Table 1 . LSI Lexicon Statistics</td><td/></tr></table>",
"text": "1 . The work reported in this paper was supported in part by the Defense Advanced Research Projects Agency , Software and Intelligent Systems Technology Office, ruder Contract No . N66001-90-C-0192 (Subcontrac t 19-930042-31 to SAIC), and by the U . S . Army Ballistic Research Laboratory under Contract No . DAAA15-89-C-0004 (Subcontract No . 05-562-01 to Logicon, Inc . )"
},
"TABREF2": {
"html": null,
"type_str": "table",
"num": null,
"content": "<table><tr><td>inc-total</td><td>529</td><td>1189</td><td>160</td><td>63</td><td>24</td><td>0</td><td>23</td><td>942</td><td>282</td><td>718</td><td>36</td><td>16</td><td>79</td></tr><tr><td>perp-total</td><td>249</td><td>687</td><td>39</td><td>19</td><td>41</td><td>0</td><td>4</td><td>588</td><td>150</td><td>631</td><td>19</td><td>7</td><td>86</td></tr><tr><td>phys-tgt-total</td><td>255</td><td>280</td><td>26</td><td>12</td><td>28</td><td>1</td><td>10</td><td>214</td><td>189</td><td>1788</td><td>12</td><td>11</td><td>76</td></tr><tr><td>hum-tgt-total</td><td>594</td><td>236</td><td>82</td><td>42</td><td>28</td><td>1</td><td>38</td><td>84</td><td>442</td><td>2038</td><td>17</td><td>44</td><td>36</td></tr><tr><td>Matched/Missing</td><td>1627</td><td>614</td><td>307</td><td>136</td><td>121</td><td>2</td><td>75</td><td>50</td><td>1063</td><td>1203</td><td>23</td><td>61</td><td>8</td></tr><tr><td>Matched/Spurious</td><td>971</td><td>2392</td><td>307</td><td>136</td><td>121</td><td>2</td><td>75</td><td>1828</td><td>407</td><td>4601</td><td>39</td><td>16</td><td>7 6</td></tr><tr><td>Matched Only</td><td>971</td><td>614</td><td>307</td><td>136</td><td>121</td><td>2</td><td>75</td><td>50</td><td>407</td><td>629</td><td>39</td><td>61</td><td>8</td></tr><tr><td>All Templates</td><td colspan=\"2\">1627 2392</td><td>307</td><td>136</td><td>121</td><td>2</td><td>75</td><td>1828</td><td>1063</td><td>5175</td><td>23</td><td>16</td><td>7 6</td></tr><tr><td>Set Fills Only</td><td>778</td><td>333</td><td>177</td><td>35</td><td>74</td><td>0</td><td>7</td><td>47</td><td>492</td><td>538</td><td>25</td><td>58</td><td>1 4</td></tr><tr><td>String Fills Only</td><td>419</td><td>105</td><td>50</td><td>30</td><td>22</td><td>1</td><td>30</td><td>3</td><td>317</td><td>353</td><td>16</td><td>62</td><td>3</td></tr></table>",
"text": "TST2 (MUC4 Interim COR PAR INC ICR IPA SPU MIS NON REC PRE OVG"
},
"TABREF3": {
"html": null,
"type_str": "table",
"num": null,
"content": "<table/>",
"text": ""
}
}
}
} |