Datasets:

ds3lab

/

ac-sgd-arxiv21

like 0

Advances in synthetic biology and nanotechnology have contributed to the design of tools that can be used to control, reuse, modify, and re-engineer cells' structure, as well as enabling engineers to effectively use biological cells as programmable substrates to realize Bio-Nano Things (biological embedded computing devices). Bio-NanoThings are generally tiny, non-intrusive, and concealable devices that can be used for in-vivo applications such as intra-body sensing and actuation networks, where the use of artificial devices can be detrimental. Such (nano-scale) devices can be used in various healthcare settings such as continuous health monitoring, targeted drug delivery, and nano-surgeries. These services can also be grouped to form a collaborative network (i.e., nanonetwork), whose performance can potentially be improved when connected to higher bandwidth external networks such as the Internet, say via 5G. However, to realize the IoBNT paradigm, it is also important to seamlessly connect the biological environment with the technological landscape by having a dynamic interface design to convert biochemical signals from the human body into an equivalent electromagnetic signal (and vice versa). This, unfortunately, risks the exposure of internal biological mechanisms to cyber-based sensing and medical actuation, with potential security and privacy implications. This paper comprehensively reviews bio-cyber interface for IoBNT architecture, focusing on bio-cyber interfacing options for IoBNT like biologically inspired bio-electronic devices, RFID enabled implantable chips, and electronic tattoos. This study also identifies known and potential security and privacy vulnerabilities and mitigation strategies for consideration in future IoBNT designs and implementations.

In this paper, we introduce a novel deep neural network suitable for multi-scale analysis and propose efficient model-agnostic methods that help the network extract information from high-frequency domains to reconstruct clearer images. Our model can be applied to multi-scale image enhancement problems including denoising, deblurring and single image super-resolution. Experiments on SIDD, Flickr2K, DIV2K, and REDS datasets show that our method achieves state-of-the-art performance on each task. Furthermore, we show that our model can overcome the over-smoothing problem commonly observed in existing PSNR-oriented methods and generate more natural high-resolution images by applying adversarial training.

Topological phases exhibit unconventional order that cannot be detected by any local order parameter. In the framework of Projected Entangled Pair States(PEPS), topological order is characterized by an entanglement symmetry of the local tensor which describes the model. This symmetry can take the form of a tensor product of group representations, or in the more general case a correlated symmetry action in the form of a Matrix Product Operator(MPO), which encompasses all string-net models. Among other things, these entanglement symmetries allow for the description of ground states and anyon excitations. Recently, the idea has been put forward to use those symmetries and the anyonic objects they describe as order parameters for probing topological phase transitions, and the applicability of this idea has been demonstrated for Abelian groups. In this paper, we extend this construction to the domain of non-Abelian models with MPO symmetries, and use it to study the breakdown of topological order in the double Fibonacci (DFib) string-net and its Galois conjugate, the non-hermitian double Yang-Lee (DYL) string-net. We start by showing how to construct topological order parameters for condensation and deconfinement of anyons using the MPO symmetries. Subsequently, we set up interpolations from the DFib and the DYL model to the trivial phase, and show that these can be mapped to certain restricted solid on solid(RSOS) models, which are equivalent to the $((5\pm\sqrt{5})/2)$-state Potts model, respectively. The known exact solutions of the statistical models allow us to locate the critical points, and to predict the critical exponents for the order parameters. We complement this by numerical study of the phase transitions, which fully confirms our theoretical predictions; remarkably, we find that both models exhibit a duality between the order parameters for condensation and deconfinement.

In this paper we propose a novel class of methods for high order accurate integration of multirate systems of ordinary differential equation initial-value problems. The proposed methods construct multirate schemes by approximating the action of matrix $\varphi$-functions within explicit exponential Rosenbrock (ExpRB) methods, thereby called Multirate Exponential Rosenbrock (MERB) methods. They consist of the solution to a sequence of modified "fast" initial-value problems, that may themselves be approximated through subcycling any desired IVP solver. In addition to proving how to construct MERB methods from certain classes of ExpRB methods, we provide rigorous convergence analysis of these methods and derive efficient MERB schemes of orders two through six (the highest order ever constructed infinitesimal multirate methods). We then present numerical simulations to confirm these theoretical convergence rates, and to compare the efficiency of MERB methods against other recently-introduced high order multirate methods.

When the inflaton couples to photons and amplifies electric fields, charged particles produced via the Schwinger effect can dominate the universe after inflation, which is dubbed as the Schwinger preheating. Using the hydrodynamic approach for the Boltzmann equation, we numerically study two cases, the Starobinsky inflation model with the kinetic coupling and the anisotropic inflation model. The Schwinger preheating is not observed in the latter model but occurs for a sufficiently large inflaton-photon coupling in the first model. We analytically address its condition and derive a general attractor solution of the electric fields. The occurrence of the Schwinger preheating in the first model is determined by whether the electric fields enter the attractor solution during inflation or not.

In this paper we have presented the mechanism of the barrier crossing dynamics of a Brownian particle which is coupled to a thermal bath in the presence of both time independent and fluctuating magnetic fields. Here the following three aspects are important in addition to the role of the thermal bath on the barrier crossing dynamics. Magnetic field induced coupling may introduce a resonance like effect. Another role of the field is that enhancement of its strength reduces the frequency factor of the barrier crossing rate constant. Finally, the fluctuating magnetic field introduces an induced electric field which activates the Brownian particle to cross the energy barrier. As a result of interplay among these aspects versatile non-monotonic behavior may appear in the variation of the rate constant as a function of the strength of the time independent magnetic field.

The space radiation environment is a complex combination of fast-moving ions derived from all atomic species found in the periodic table. The energy spectrum of each ion species varies widely but is prominently in the range of 400 - 600 MeV/n. The large dynamic range in ion energy is difficult to simulate in ground-based radiobiology experiments. Most ground-based irradiations with mono-energetic beams of a single one ion species are delivered at comparatively high dose rates. In some cases, sequences of such beams are delivered with various ion species and energies to crudely approximate the complex space radiation environment. This approximation may cause profound experimental bias in processes such as biologic repair of radiation damage, which are known to have strong temporal dependancies. It is possible that this experimental bias leads to an overprediction of risks of radiation effects that have not been observed in the astronaut cohort. None of the primary health risks presumely attributed to space radiation exposure, such as radiation carciogenesis, cardiovascular disease, cognitive deficits, etc., have been observed in astronaut or cosmonaut crews. This fundamentally and profoundly limits our understanding of the effects of GCR on humans and limits the development of effective radiation countermeasures.

Time series forecasting methods play critical role in estimating the spread of an epidemic. The coronavirus outbreak of December 2019 has already infected millions all over the world and continues to spread on. Just when the curve of the outbreak had started to flatten, many countries have again started to witness a rise in cases which is now being referred as the 2nd wave of the pandemic. A thorough analysis of time-series forecasting models is therefore required to equip state authorities and health officials with immediate strategies for future times. This aims of the study are three-fold: (a) To model the overall trend of the spread; (b) To generate a short-term forecast of 10 days in countries with the highest incidence of confirmed cases (USA, India and Brazil); (c) To quantitatively determine the algorithm that is best suited for precise modelling of the linear and non-linear features of the time series. The comparison of forecasting models for the total cumulative cases of each country is carried out by comparing the reported data and the predicted value, and then ranking the algorithms (Prophet, Holt-Winters, LSTM, ARIMA, and ARIMA-NARNN) based on their RMSE, MAE and MAPE values. The hybrid combination of ARIMA and NARNN (Nonlinear Auto-Regression Neural Network) gave the best result among the selected models with a reduced RMSE, which proved to be almost 35.3% better than one of the most prevalent method of time-series prediction (ARIMA). The results demonstrated the efficacy of the hybrid implementation of the ARIMA-NARNN model over other forecasting methods such as Prophet, Holt Winters, LSTM, and the ARIMA model in encapsulating the linear as well as non-linear patterns of the epidemical datasets.

We present mid infrared imaging of two young clusters, the Coronet in the CrA cloud core and B59 in the Pipe Nebula, using the FORCAST camera on the Stratospheric Observatory for Infrared Astronomy. We also analyze Herschel Space Observatory PACS and SPIRE images of the associated clouds. The two clusters are at similar, and very close, distances. Star formation is ongoing in the Coronet, which hosts at least one Class 0 source and several pre-stellar cores, which may collapse and form stars. The B59 cluster is older, although it still has a few Class I sources, and is less compact. The CrA cloud has a diameter of about 0.16 pc, and we determine a dust temperature of 15.7 K and a star formation efficiency of about 27 %, while the B59 core is approximately twice as large, has a dust temperature of about 11.4 K and a star formation efficiency of about 14 %. We infer that the gas densities are much higher in the Coronet, which has also formed intermediate mass stars, while B59 has only formed low-mass stars.

This paper presents approximation methods for time-dependent thermal radiative transfer problems in high energy density physics. It is based on the multilevel quasidiffusion method defined by the high-order radiative transfer equation (RTE) and the low-order quasidiffusion (aka VEF) equations for the moments of the specific intensity. A large part of data storage in TRT problems between time steps is determined by the dimensionality of grid functions of the radiation intensity. The approximate implicit methods with reduced memory for the time-dependent Boltzmann equation are applied to the high-order RTE, discretized in time with the backward Euler (BE) scheme. The high-dimensional intensity from the previous time level in the BE scheme is approximated by means of the low-rank proper orthogonal decomposition (POD). Another version of the presented method applies the POD to the remainder term of P2 expansion of the intensity. The accuracy of the solution of the approximate implicit methods depends of the rank of the POD. The proposed methods enable one to reduce storage requirements in time dependent problems. Numerical results of a Fleck-Cummings TRT test problem are presented.

In this paper we study nonlinear interpolation problems for interpolation and peak-interpolation sets of function algebras. The subject goes back to the classical Rudin-Carleson interpolation theorem. In particular, we prove the following nonlinear version of this theorem: Let $\bar{\mathbb D}\subset \mathbb C$ be the closed unit disk, $\mathbb T\subset\bar{\mathbb D}$ the unit circle, $S\subset\mathbb T$ a closed subset of Lebesgue measure zero and $M$ a connected complex manifold. Then for every continuous $M$-valued map $f$ on $S$ there exists a continuous $M$-valued map $g$ on $\bar{\mathbb D}$ holomorphic on its interior such that $g|_S=f$. We also consider similar interpolation problems for continuous maps $f: S\rightarrow\bar M$, where $\bar M$ is a complex manifold with boundary $\partial M$ and interior $M$. Assuming that $f(S)\cap\partial M\ne\emptyset$ we are looking for holomorphic extensions $g$ of $f$ such that $g(\bar{\mathbb D}\setminus S)\subset M$.

In this paper, we investigate Riesz energy problems on unbounded conductors in $\R^d$ in the presence of general external fields $Q$, not necessarily satisfying the growth condition $Q(x)\to\infty$ as $x\to\infty$ assumed in several previous studies. We provide sufficient conditions on $Q$ for the existence of an equilibrium measure and the compactness of its support. Particular attention is paid to the case of the hyperplanar conductor $\R^{d}$, embedded in $\R^{d+1}$, when the external field is created by the potential of a signed measure $\nu$ outside of $\R^{d}$. Simple cases where $\nu$ is a discrete measure are analyzed in detail. New theoretic results for Riesz potentials, in particular an extension of a classical theorem by de La Vall\'ee-Poussin, are established. These results are of independent interest.

DNA sequencing is becoming increasingly commonplace, both in medical and direct-to-consumer settings. To promote discovery, collected genomic data is often de-identified and shared, either in public repositories, such as OpenSNP, or with researchers through access-controlled repositories. However, recent studies have suggested that genomic data can be effectively matched to high-resolution three-dimensional face images, which raises a concern that the increasingly ubiquitous public face images can be linked to shared genomic data, thereby re-identifying individuals in the genomic data. While these investigations illustrate the possibility of such an attack, they assume that those performing the linkage have access to extremely well-curated data. Given that this is unlikely to be the case in practice, it calls into question the pragmatic nature of the attack. As such, we systematically study this re-identification risk from two perspectives: first, we investigate how successful such linkage attacks can be when real face images are used, and second, we consider how we can empower individuals to have better control over the associated re-identification risk. We observe that the true risk of re-identification is likely substantially smaller for most individuals than prior literature suggests. In addition, we demonstrate that the addition of a small amount of carefully crafted noise to images can enable a controlled trade-off between re-identification success and the quality of shared images, with risk typically significantly lowered even with noise that is imperceptible to humans.

Recent works on Binary Neural Networks (BNNs) have made promising progress in narrowing the accuracy gap of BNNs to their 32-bit counterparts. However, the accuracy gains are often based on specialized model designs using additional 32-bit components. Furthermore, almost all previous BNNs use 32-bit for feature maps and the shortcuts enclosing the corresponding binary convolution blocks, which helps to effectively maintain the accuracy, but is not friendly to hardware accelerators with limited memory, energy, and computing resources. Thus, we raise the following question: How can accuracy and energy consumption be balanced in a BNN network design? We extensively study this fundamental problem in this work and propose a novel BNN architecture without most commonly used 32-bit components: \textit{BoolNet}. Experimental results on ImageNet demonstrate that BoolNet can achieve 4.6x energy reduction coupled with 1.2\% higher accuracy than the commonly used BNN architecture Bi-RealNet. Code and trained models are available at: https://github.com/hpi-xnor/BoolNet.

Many task-oriented dialogue systems use deep reinforcement learning (DRL) to learn policies that respond to the user appropriately and complete the tasks successfully. Training DRL agents with diverse dialogue trajectories prepare them well for rare user requests and unseen situations. One effective diversification method is to let the agent interact with a diverse set of learned user models. However, trajectories created by these artificial user models may contain generation errors, which can quickly propagate into the agent's policy. It is thus important to control the quality of the diversification and resist the noise. In this paper, we propose a novel dialogue diversification method for task-oriented dialogue systems trained in simulators. Our method, Intermittent Short Extension Ensemble (I-SEE), constrains the intensity to interact with an ensemble of diverse user models and effectively controls the quality of the diversification. Evaluations on the Multiwoz dataset show that I-SEE successfully boosts the performance of several state-of-the-art DRL dialogue agents.

When manipulating three-dimensional data, it is possible to ensure that rotational and translational symmetries are respected by applying so-called SE(3)-equivariant models. Protein structure prediction is a prominent example of a task which displays these symmetries. Recent work in this area has successfully made use of an SE(3)-equivariant model, applying an iterative SE(3)-equivariant attention mechanism. Motivated by this application, we implement an iterative version of the SE(3)-Transformer, an SE(3)-equivariant attention-based model for graph data. We address the additional complications which arise when applying the SE(3)-Transformer in an iterative fashion, compare the iterative and single-pass versions on a toy problem, and consider why an iterative model may be beneficial in some problem settings. We make the code for our implementation available to the community.

In the framework of the Standard Model (SM) a theoretical description of the neutron beta decay is given at the level of 10^{-5}. The neutron lifetime and correlation coefficients of the neutron beta decay for a polarized neutron, a polarized electron and an unpolarized proton are calculated at the account for i) the radiative corrections of order O(\alpha E_e/m_N) ~ 10^{-5} to Sirlin's outer and inner radiative corrections of order O(\alpha/\pi), ii) the corrections of order O(E^2_e/m^2_N) ~ 10^{-5}, caused by weak magnetism and proton recoil, and iii) Wilkinson's corrections of order 10^{-5} (Wilkinson, Nucl. Phys. A377, 474 (1982)). These corrections define the SM background of the theoretical description of the neutron beta decay at the level of 10^{-5}, which is required by experimental searches of interactions beyond the SM with experimental uncertainties of a few parts of 10^{-5}.

Face swapping has both positive applications such as entertainment, human-computer interaction, etc., and negative applications such as DeepFake threats to politics, economics, etc. Nevertheless, it is necessary to understand the scheme of advanced methods for high-quality face swapping and generate enough and representative face swapping images to train DeepFake detection algorithms. This paper proposes the first Megapixel level method for one shot Face Swapping (or MegaFS for short). Firstly, MegaFS organizes face representation hierarchically by the proposed Hierarchical Representation Face Encoder (HieRFE) in an extended latent space to maintain more facial details, rather than compressed representation in previous face swapping methods. Secondly, a carefully designed Face Transfer Module (FTM) is proposed to transfer the identity from a source image to the target by a non-linear trajectory without explicit feature disentanglement. Finally, the swapped faces can be synthesized by StyleGAN2 with the benefits of its training stability and powerful generative capability. Each part of MegaFS can be trained separately so the requirement of our model for GPU memory can be satisfied for megapixel face swapping. In summary, complete face representation, stable training, and limited memory usage are the three novel contributions to the success of our method. Extensive experiments demonstrate the superiority of MegaFS and the first megapixel level face swapping database is released for research on DeepFake detection and face image editing in the public domain. The dataset is at this link.

In this paper, we study the problem of mobile user profiling, which is a critical component for quantifying users' characteristics in the human mobility modeling pipeline. Human mobility is a sequential decision-making process dependent on the users' dynamic interests. With accurate user profiles, the predictive model can perfectly reproduce users' mobility trajectories. In the reverse direction, once the predictive model can imitate users' mobility patterns, the learned user profiles are also optimal. Such intuition motivates us to propose an imitation-based mobile user profiling framework by exploiting reinforcement learning, in which the agent is trained to precisely imitate users' mobility patterns for optimal user profiles. Specifically, the proposed framework includes two modules: (1) representation module, which produces state combining user profiles and spatio-temporal context in real-time; (2) imitation module, where Deep Q-network (DQN) imitates the user behavior (action) based on the state that is produced by the representation module. However, there are two challenges in running the framework effectively. First, epsilon-greedy strategy in DQN makes use of the exploration-exploitation trade-off by randomly pick actions with the epsilon probability. Such randomness feeds back to the representation module, causing the learned user profiles unstable. To solve the problem, we propose an adversarial training strategy to guarantee the robustness of the representation module. Second, the representation module updates users' profiles in an incremental manner, requiring integrating the temporal effects of user profiles. Inspired by Long-short Term Memory (LSTM), we introduce a gated mechanism to incorporate new and old user characteristics into the user profile.

We study the polarization dynamics of ultrafast solitons in mode-locked fiber lasers. We find that when a stable soliton is generated, it's state-of-polarization shifts toward a stable state and when the soliton is generated with excess power levels it experiences relaxation oscillations in its intensity and timing. On the other hand, when a soliton is generated in an unstable state-of-polarization, it either decays in intensity until it disappears, or its temporal width decreases until it explodes into several solitons and then it disappears. We also found that when two solitons are simultaneously generated close to each other, they attract each other until they collide and merge into a single soliton. Although, these two solitons are generated with different states-of-polarization, they shift their state-of-polarization closer to each other until the polarization coincides when they collide. We support our findings by numerical calculations of a non-Lagrangian approach by simulating the Ginzburg-Landau equation governing the dynamics of solitons in a laser cavity. Our model also predicts the relaxation oscillations of stable solitons and the two types of unstable solitons observed in the experimental measurements.

This paper presents Favalon, a functional programming language built on the premise of a lambda calculus for use as an interactive shell replacement. Favalon seamlessly integrates with typed versions of existing libraries and commands using type inference, flexible runtime type metadata, and the same techniques employed by shells to link commands together. Much of Favalon's syntax is customizable via user-defined functions, allowing it to be extended by anyone who is familiar with a command-line shell. Furthermore, Favalon's type inference engine can be separated from its runtime library and easily repurposed for other applications.

Recently, asymmetric plasmonic nanojunctions [Karnetzky et. al., Nature Comm. 2471, 9 (2018)] have shown promise as on-chip electronic devices to convert femtosecond optical pulses to current bursts, with a bandwidth of multi-terahertz scale, although yet at low temperatures and pressures. Such nanoscale devices are of great interest for novel ultrafast electronics and opto-electronic applications. Here, we operate the device in air and at room temperature, revealing the mechanisms of photoemission from plasmonic nanojunctions, and the fundamental limitations on the speed of optical-to-electronic conversion. Inter-cycle interference of coherent electronic wavepackets results in a complex energy electron distribution and birth of multiphoton effects. This energy structure, as well as reshaping of the wavepackets during their propagation from one tip to the other, determine the ultrafast dynamics of the current. We show that, up to some level of approximation, the electron flight time is well-determined by the mean ponderomotive velocity in the driving field.

The necessary and sufficient conditions are given for a sequence of complex numbers to be the periodic (or antiperiodic) spectrum of non-self-adjoint Dirac operator.

In this paper, we consider a type of image quality assessment as a task-specific measurement, which can be used to select images that are more amenable to a given target task, such as image classification or segmentation. We propose to train simultaneously two neural networks for image selection and a target task using reinforcement learning. A controller network learns an image selection policy by maximising an accumulated reward based on the target task performance on the controller-selected validation set, whilst the target task predictor is optimised using the training set. The trained controller is therefore able to reject those images that lead to poor accuracy in the target task. In this work, we show that the controller-predicted image quality can be significantly different from the task-specific image quality labels that are manually defined by humans. Furthermore, we demonstrate that it is possible to learn effective image quality assessment without using a ``clean'' validation set, thereby avoiding the requirement for human labelling of images with respect to their amenability for the task. Using $6712$, labelled and segmented, clinical ultrasound images from $259$ patients, experimental results on holdout data show that the proposed image quality assessment achieved a mean classification accuracy of $0.94\pm0.01$ and a mean segmentation Dice of $0.89\pm0.02$, by discarding $5\%$ and $15\%$ of the acquired images, respectively. The significantly improved performance was observed for both tested tasks, compared with the respective $0.90\pm0.01$ and $0.82\pm0.02$ from networks without considering task amenability. This enables image quality feedback during real-time ultrasound acquisition among many other medical imaging applications.

We show that a novel, general phase space mapping Hamiltonian for nonadiabatic systems, which is reminiscent of the renowned Meyer-Miller mapping Hamiltonian, involves a commutator variable matrix rather than the conventional zero-point-energy parameter. In the exact mapping formulation on constraint space for phase space approaches for nonadiabatic dynamics, the general mapping Hamiltonian with commutator variables can be employed to generate approximate trajectory-based dynamics. Various benchmark model tests, which range from gas phase to condensed phase systems, suggest that the overall performance of the general mapping Hamiltonian is better than that of the conventional Meyer-Miller Hamiltonian.

Many man-made objects are characterised by a shape that is symmetric along one or more planar directions. Estimating the location and orientation of such symmetry planes can aid many tasks such as estimating the overall orientation of an object of interest or performing shape completion, where a partial scan of an object is reflected across the estimated symmetry plane in order to obtain a more detailed shape. Many methods processing 3D data rely on expensive 3D convolutions. In this paper we present an alternative novel encoding that instead slices the data along the height dimension and passes it sequentially to a 2D convolutional recurrent regression scheme. The method also comprises a differentiable least squares step, allowing for end-to-end accurate and fast processing of both full and partial scans of symmetric objects. We use this approach to efficiently handle 3D inputs to design a method to estimate planar reflective symmetries. We show that our approach has an accuracy comparable to state-of-the-art techniques on the task of planar reflective symmetry estimation on full synthetic objects. Additionally, we show that it can be deployed on partial scans of objects in a real-world pipeline to improve the outputs of a 3D object detector.

In this paper we present a novel mechanism for producing the observed Dark Matter(DM) relic abundance during the First Order Phase Transition (FOPT) in the early universe. We show that the bubble expansion with ultra-relativistic velocities can lead to the abundance of DM particles with masses much larger than the scale of the transition. We study this non-thermal production mechanism in the context of a generic phase transition and the electroweak phase transition. The application of the mechanism to the Higgs portal DM as well as the signal in the Stochastic Gravitational Background are discussed.

Introduction. Can the infection due to the human immunodeficiency virus type 1 induce a change in the differentiation status or process in T cells?. Methods. We will consider two stochastic Markov chain models, one which will describe the T-helper cell differentiation process, and another one describing that process of infection of the T-helper cell by the virus; in these Markov chains, we will consider a set of states $\{X_t \}$ comprised of those proteins involved in each of the processes and their interactions (either differentiation or infection of the cell), such that we will obtain two stochastic transition matrices ($A,B$), one for each process; afterwards, the computation of their eigenvalues shall be performed, in which, should the eigenvalue $\lambda_i=1$ exist, the computation for the equilibrium distribution $\pi^n$ will be obtained for each of the matrices, which will inform us on the trends of interactions amongst the proteins in the long-term. Results. The stochastic processes considered possess an equilibrium distribution, when reaching their equilibrium distribution, there exists an increase in their informational entropy, and their log-rank distributions can be modeled as discrete beta generalized distributions (DGBD). Discussion. The equilibrium distributions of both process can be regarded as states in which the cell is well-differentiated, ergo there exists an induction of a novel HIV-dependent differentiated state in the T-cell; these processes due to their DGBD distribution can be considered complex processes; due to the increasing entropy, the equilibrium states are stable ones. Conclusion. The HIV virus can promote a novel differentiated state in the T-cell, which can give account for clinical features seen in patients; this model, notwithstanding does not give account of YES/NO logical switches involved in the regulatory networks.

We propose an optimal MMSE precoding technique using quantized signals with constant envelope. Unlike the existing MMSE design that relies on 1-bit resolution, the proposed approach employs uniform phase quantization and the bounding step in the branch-and-bound method is different in terms of considering the most restrictive relaxation of the nonconvex problem, which is then utilized for a suboptimal design also. Moreover, unlike prior studies, we propose three different soft detection methods and an iterative detection and decoding scheme that allow the utilization of channel coding in conjunction with low-resolution precoding. Besides an exact approach for computing the extrinsic information, we propose two approximations with reduced computational complexity. Numerical simulations show that utilizing the MMSE criterion instead of the established maximum-minimum distance to the decision threshold yields a lower bit-error-rate in many scenarios. Furthermore, when using the MMSE criterion, a smaller number of bound evaluations in the branch-and-bound method is required for low and medium SNR. Finally, results based on an LDPC block code indicate that the receive processing schemes yield a lower bit-error-rate compared to the conventional design.

We consider the geodesic of the directed last passage percolation with iid exponential weights. We find the explicit one point distribution of the geodesic location joint with the last passage times, and its limit when the size goes to infinity.

We consider the problem of minimizing age of information in general single-hop and multihop wireless networks. First, we formulate a way to convert AoI optimization problems into equivalent network stability problems. Then, we propose a heuristic low complexity approach for achieving stability that can handle general network topologies; unicast, multicast and broadcast flows; interference constraints; link reliabilities; and AoI cost functions. We provide numerical results to show that our proposed algorithms behave as well as the best known scheduling and routing schemes available in the literature for a wide variety of network settings.

We develop a theory for the non-equilibrium screening of a charged impurity in a two-dimensional electron system under a strong time-periodic drive. Our analysis of the time-averaged polarization function and dielectric function reveals that Floquet driving modifies the screened impurity potential in two main regimes. In the weak drive regime, the time-averaged screened potential exhibits unconventional Friedel oscillations with multiple spatial periods contributed by a principal period modulated by higher-order periods, which are due to the emergence of additional Kohn anomalies in the polarization function. In the strong drive regime, the time-averaged impurity potential becomes almost unscreened and does not exhibit Friedel oscillations. This tunable Friedel oscillations is a result of the dynamic gating effect of the time-dependent driving field on the two-dimensional electron system.

In this paper, based on the idea of self-adjusting steepness based schemes[5], a two-dimensional calculation method of steepness parameter is proposed, and thus a two-dimensional self-adjusting steepness based limiter is constructed. With the application of such limiter to the over-intersection based remapping framework, a low dissipation remapping method has been proposed that can be applied to the existing ALE method.

We derive the Laws of Cosines and Sines in the super hyperbolic plane using Minkowski supergeometry and find the identical formulae to the classical case, but remarkably involving different expressions for cosines and sines of angles which include substantial fermionic corrections. In further analogy to the classical case, we apply these results to show that two parallel supergeodesics which are not ultraparallel admit a unique common orthogonal supergeodesic, and we briefly describe aspects of elementary supernumber theory, leading to a prospective analogue of the Gauss product of quadratic forms.

We present an analysis of the galaxy environment and physical properties of a partial Lyman limit system at z = 0.83718 with HI and metal line components closely separated in redshift space ($|\Delta v| \approx 400$ km/s) towards the background quasar HE1003+0149. The HST/COS far-ultraviolet spectrum provides coverage of lines of oxygen ions from OI to OV. Comparison of observed spectral lines with synthetic profiles generated from Bayesian ionization modeling reveals the presence of two distinct gas phases in the absorbing medium. The low-ionization phase of the absorber has sub-solar metallicities (1/10-th solar) with indications of [C/O] < 0 in each of the components. The OIV and OV trace a more diffuse higher-ionization medium with predicted HI column densities that are $\approx 2$ dex lower. The quasar field observed with VLT/MUSE reveals three dwarf galaxies with stellar masses of $M^* \sim 10^{8} - 10^{9}$ M$_\odot$, and with star formation rates of $\approx 0.5 - 1$ M$_\odot$ yr$^{-1}$, at projected separations of $\rho/R_{\mathrm{vir}} \approx 1.8 - 3.0$ from the absorber. Over a wider field with projected proper separation of $\leq 5$ Mpc and radial velocity offset of $|\Delta v| \leq 1000$ km/s from the absorber, 21 more galaxies are identified in the $VLT$/VIMOS and Magellan deep galaxy redshift surveys, with 8 of them within $1$ Mpc and $500$ km/s, consistent with the line of sight penetrating a group of galaxies. The absorber presumably traces multiple phases of cool ($T \sim 10^4$ K) photoionized intragroup medium. The inferred [C/O] < 0 hints at preferential enrichment from core-collapse supernovae, with such gas displaced from one or more of the nearby galaxies, and confined to the group medium.

Transition metal dichalcogenides (TMDs) combine interesting optical and spintronic properties in an atomically-thin material, where the light polarization can be used to control the spin and valley degrees-of-freedom for the development of novel opto-spintronic devices. These promising properties emerge due to their large spin-orbit coupling in combination with their crystal symmetries. Here, we provide simple symmetry arguments in a group-theory approach to unveil the symmetry-allowed spin scattering mechanisms, and indicate how one can use these concepts towards an external control of the spin lifetime. We perform this analysis for both monolayer (inversion asymmetric) and bilayer (inversion symmetric) crystals, indicating the different mechanisms that play a role in these systems. We show that, in monolayer TMDs, electrons and holes transform fundamentally differently -- leading to distinct spin-scattering processes. We find that one of the electronic states in the conduction band is partially protected by time-reversal symmetry, indicating a longer spin lifetime for that state. In bilayer and bulk TMDs, a hidden spin-polarization can exist within each layer despite the presence of global inversion symmetry. We show that this feature enables control of the interlayer spin-flipping scattering processes via an out-of-plane electric field, providing a mechanism for electrical control of the spin lifetime.

We study the dynamics of a one-dimensional Rydberg lattice gas under facilitation (anti-blockade) conditions which implements a so-called kinetically constrained spin system. Here an atom can only be excited to a Rydberg state when one of its neighbors is already excited. Once two or more atoms are simultaneously excited mechanical forces emerge, which couple the internal electronic dynamics of this many-body system to external vibrational degrees of freedom in the lattice. This electron-phonon coupling results in a so-called phonon dressing of many-body states which in turn impacts on the facilitation dynamics. In our theoretical study we focus on a scenario in which all energy scales are sufficiently separated such that a perturbative treatment of the coupling between electronic and vibrational states is possible. This allows to analytically derive an effective Hamiltonian for the evolution of consecutive clusters of Rydberg excitations in the presence of phonon dressing. We analyze the spectrum of this Hamiltonian and show -- by employing Fano resonance theory -- that the interaction between Rydberg excitations and lattice vibrations leads to the emergence of slowly decaying bound states that inhibit fast relaxation of certain initial states.

Drafting as a process to reduce drag and to benefit from the presence of other competitors is applied in various sports with several recent examples of competitive running in formations. In this study, the aerodynamics of a realistic model of a female runner is calculated by computational fluid dynamics (CFD) simulations at four running speeds of 15 km/h, 18 km/h, 21 km/h, and 36 km/h. Aerodynamic power fractions of the total energy expenditure are found to be in the range of 2.6-8.5%. Additionally, four exemplary formations are analysed with respect to their drafting potential and resulting drag values are compared for the main runner and her pacers. The best of the formations achieves a total drag reduction on the main runner of 75.6%. Moreover, there are large variations in the drag reduction between the considered formations of up to 42% with respect to the baseline single-runner case. We conclude that major drag reduction of more than 70% can already be achieved with fairly simple formations, while certain factors, such as runners on the sides, can have a detrimental effect on drag reduction due to local acceleration of the passing flow. Using an empirical model for mechanical power output during running, gains of metabolic power and performance predictions are evaluated for all considered formations. Improvements in running economy are up to 3.5% for the best formation, leading to velocity gains of 2.3%. This translates to 154 s (~2.6 min) saved over a marathon distance. Consequently, direct conclusions are drawn from the obtained data for ideal drafting of long-distance running in highly packed formations.

Turbulence in the upper ocean in the submesoscale range (scales smaller than the deformation radius) plays an important role for the heat exchange with the atmosphere and for oceanic biogeochemistry. Its dynamics should strongly depend on the seasonal cycle and the associated mixed-layer instabilities. The latter are particularly relevant in winter and are responsible for the formation of energetic small scales that extend over the whole depth of the mixed layer. The knowledge of the transport properties of oceanic flows at depth, which is essential to understand the coupling between surface and interior dynamics, however, is still limited. By means of numerical simulations, we explore the Lagrangian dispersion properties of turbulent flows in a quasi-geostrophic model system allowing for both thermocline and mixed-layer instabilities. The results indicate that, when mixed-layer instabilities are present, the dispersion regime is local from the surface down to depths comparable with that of the interface with the thermocline, while in their absence dispersion quickly becomes nonlocal versus depth. We then identify the origin of such behavior in the existence of fine-scale energetic structures due to mixed-layer instabilities. We further discuss the effect of vertical shear on the Lagrangian particle spreading and address the correlation between the dispersion properties at the surface and at depth, which is relevant to assess the possibility of inferring the dynamical features of deeper flows from the more accessible surface ones.

Electrochemically mediated selective adsorption is an emerging electrosorption technique that utilizes Faradaically enhanced redox active electrodes, which can adsorb ions not only electrostatically, but also electrochemically. The superb selectivity (>100) of this technique enables selective removal of toxic or high-value target ions under low energy consumption. Here, we develop a general theoretical framework to describe the competitive electrosorption phenomena involving multiple ions and surface-bound redox species. The model couples diffusion, convection and electromigration with competitive surface adsorption reaction kinetics, consistently derived from non-equilibrium thermodynamics. To optimize the selective removal of the target ions, design criteria were derived analytically from physically relevant dimensionless groups and time scales, where the propagation of the target anions concentration front is the limiting step. Detailed computational studies are reported for three case studies that cover a wide range of inlet concentration ratios between the competing ions. And in all three cases, target anions in the electrosorption cell forms a self-sharpening reaction-diffusion wave front. Based on the model, a three-step stop-flow operation scheme with a pure stripping solution of target anions is proposed that optimizes the ion adsorption performance and increases the purity of the regeneration stream to almost 100%, which is beneficial for downstream processing.

Quality of website design is one of the influential factors of website success. How the design helps the users using effectively and efficiently website and satisfied at the end of the use. However, it is a common tendency that websites are designed based on the developer's perspectives and lack considering user importance. Thus, the degree of website usability tends to be low according to user perceptions. This study purposed to understand the user experiences using an institutional repository (IR) website in a public university in Indonesia. The research was performed based on usability testing framework as the usability testing method. About 12 participants were purposely involved concerning their key informant characteristics. Following three empirical data collection techniques (i.e., query technique, formal experiment, and thinking aloud), both descriptive analysis using usability scale matric and content analysis using qualitative data analysis (QDA) Miner Lite software were used in the data analysis stage. Lastly, several visual design recommendations were then proposed at the end of the study. In terms of a case study, besides the practical recommendations which may contextually useful for the next website development; the clarity of the research design may also help scholars how to combine more than one usability testing technique within a multi-technique study design.

The SIMT execution model is commonly used for general GPU development. CUDA and OpenCL developers write scalar code that is implicitly parallelized by compiler and hardware. On Intel GPUs, however, this abstraction has profound performance implications as the underlying ISA is SIMD and important hardware capabilities cannot be fully utilized. To close this performance gap we introduce C-For-Metal (CM), an explicit SIMD programming framework designed to deliver close-to-the-metal performance on Intel GPUs. The CM programming language and its vector/matrix types provide an intuitive interface to exploit the underlying hardware features, allowing fine-grained register management, SIMD size control and cross-lane data sharing. Experimental results show that CM applications from different domains outperform the best-known SIMT-based OpenCL implementations, achieving up to 2.7x speedup on the latest Intel GPU.

We systematically investigate axisymmetric extremal isolated horizons (EIHs) defined by vanishing surface gravity, corresponding to zero temperature. In the first part, using the Newman-Penrose and GHP formalism we derive the most general metric function for such EIHs in the Einstein-Maxwell theory, which complements the previous result of Lewandowski and Pawlowski. We prove that it depends on 5 independent parameters, namely deficit angles on the north and south poles of a spherical-like section of the horizon, its radius (area), and total electric and magnetic charges of the black hole. The deficit angles and both charges can be separately set to zero. In the second part of our paper, we identify this general axially symmetric solution for EIH with extremal horizons in exact electrovacuum Plebanski-Demianski spacetimes, using the convenient parametrization of this family by Griffiths and Podolsky. They represent all (double aligned) black holes of algebraic type D without a cosmological constant. Apart from a conicity, they depend on 6 physical parameters (mass, Kerr-like rotation, NUT parameter, acceleration, electric and magnetic charges) constrained by the extremality condition. We were able to determine their relation to the EIH geometrical parameters. This explicit identification of type D extremal black holes with a unique form of EIH includes several interesting subclasses, such as accelerating extremely charged Reissner-Nordstrom black hole (C-metric), extremal accelerating Kerr-Newman, accelerating Kerr-NUT, or non-accelerating Kerr-Newman-NUT black holes.

Multimode nonlinear optics offers to overcome a long-standing limitation of fiber optics, tightly phase locking several spatial modes and enabling the coherent transport of a wavepacket through a multimode fiber. A similar problem is encountered in the temporal compression of multi-mJ pulses to few-cycle duration in hollow gas-filled fibers. Scaling the fiber length to up to six meters, hollow fibers have recently reached 1 TW of peak power. Despite the remarkable utility of the hollow fiber compressor and its widespread application, however, no analytical model exists to enable insight into the scaling behavior of maximum compressibility and peak power. Here we extend a recently introduced formalism for describing mode-locking to the spatially analogue scenario of locking spatial fiber modes together. Our formalism unveils the coexistence of two soliton branches for anomalous modal dispersion and indicates the formation of stable spatio-temporal light bullets that would be unstable in free space, similar to the temporal cage solitons in mode-locking theory. Our model enables deeper understanding of the physical processes behind the formation of such light bullets and predict the existence of multimode solitons in a much wider range of fiber types than previously considered possible.

Let $f : X \to S$ be a family of smooth projective algebraic varieties over a smooth connected base $S$, with everything defined over $\overline{\mathbb{Q}}$. Denote by $\mathbb{V} = R^{2i} f_{*} \mathbb{Z}(i)$ the associated integral variation of Hodge structure on the degree $2i$ cohomology. We consider the following question: when can a fibre $\mathbb{V}_{s}$ above an algebraic point $s \in S(\overline{\mathbb{Q}})$ be isomorphic to a transcendental fibre $\mathbb{V}_{s'}$ with $s' \in S(\mathbb{C}) \setminus S(\overline{\mathbb{Q}})$? When $\mathbb{V}$ induces a quasi-finite period map $\varphi : S \to \Gamma \backslash D$, conjectures in Hodge theory predict that such isomorphisms cannot exist. We introduce new differential-algebraic techniques to show this is true for all points $s \in S(\overline{\mathbb{Q}})$ outside of an explicit proper closed algebraic subset of $S$. As a corollary we establish the existence of a canonical $\overline{\mathbb{Q}}$-algebraic model for normalizations of period images.

We study co-dimension two monodromy defects in theories of conformally coupled scalars and free Dirac fermions in arbitrary $d$ dimensions. We characterise this family of conformal defects by computing the one-point functions of the stress-tensor and conserved current for Abelian flavour symmetries as well as two-point functions of the displacement operator. In the case of $d=4$, the normalisation of these correlation functions are related to defect Weyl anomaly coefficients, and thus provide crucial information about the defect conformal field theory. We provide explicit checks on the values of the defect central charges by calculating the universal part of the defect contribution to entanglement entropy, and further, we use our results to extract the universal part of the vacuum R\'enyi entropy. Moreover, we leverage the non-supersymmetric free field results to compute a novel defect Weyl anomaly coefficient in a $d=4$ theory of free $\mathcal{N}=2$ hypermultiplets. Including singular modes in the defect operator product expansion of fundamental fields, we identify notable relevant deformations in the singular defect theories and show that they trigger a renormalisation group flow towards an IR fixed point with the most regular defect OPE. We also study Gukov-Witten defects in free $d=4$ Maxwell theory and show that their central charges vanish.

Loneliness (i.e., the distressing feeling that often accompanies the subjective sense of social disconnection) is detrimental to mental and physical health, and deficits in self-reported feelings of being understood by others is a risk factor for loneliness. What contributes to these deficits in lonely people? We used functional magnetic resonance imaging (fMRI) to unobtrusively measure the relative alignment of various aspects of people's mental processing of naturalistic stimuli (specifically, videos) as they unfold over time. We thereby tested whether lonely people actually process the world in idiosyncratic ways, rather than only exaggerating or misperceiving how dissimilar others' views are to their own (which could lead them to feel misunderstood, even if they actually see the world similarly to those around them). We found evidence for such idiosyncrasy: lonely individuals' neural responses during free viewing of the videos were dissimilar to peers in their communities, particularly in brain regions (e.g., regions of the default-mode network) in which similar responses have been associated with shared psychological perspectives and subjective understanding. Our findings were robust even after controlling for demographic similarities, participants' overall levels of objective social isolation, and their friendships with each other. These results suggest that being surrounded predominantly by people who see the world differently from oneself may be a risk factor for loneliness, even if one is friends with them.

The Traditional Approximation of Rotation (TAR) is a treatment of the hydrodynamic equations of rotating and stably stratified fluids in which the action of the Coriolis acceleration along the direction of the entropy and chemical stratifications is neglected because it is weak in comparison with the buoyancy force. The dependent variables in the equations for the dynamics of gravito-inertial waves (GIWs) then become separable into radial and horizontal parts as in the non-rotating case. The TAR is built on the assumptions that the star is spherical (i.e. its centrifugal deformation is neglected) and uniformly rotating. We study the feasibility of carrying out a generalisation of the TAR to account for the centrifugal acceleration in the case of strongly deformed uniformly and rapidly rotating stars (and planets), and to identify the validity domain of this approximation. We built analytically a complete formalism that allows the study of the dynamics of GIWs in spheroidal coordinates which take into account the flattening of rapidly rotating stars by assuming the hierarchies of frequencies adopted within the TAR in the spherical case and by deriving a generalised Laplace tidal equation for the horizontal eigenfunctions of the GIWs and their asymptotic wave periods, which can be used to probe the structure and dynamics of rotating deformed stars with asteroseismology. Using 2D ESTER stellar models, we determine the validity domain of the generalised TAR as a function of the rotation rate of the star normalised by its critical angular velocity and its pseudo-radius. This generalisation allows us to study the signature of the centrifugal effects on GIWs in rapidly rotating deformed stars. We found that the effects of the centrifugal acceleration in rapidly rotating early-type stars on GIWs are theoretically detectable in modern space photometry using observations from Kepler.

Artificial intelligence is applied in a range of sectors, and is relied upon for decisions requiring a high level of trust. For regression methods, trust is increased if they approximate the true input-output relationships and perform accurately outside the bounds of the training data. But often performance off-test-set is poor, especially when data is sparse. This is because the conditional average, which in many scenarios is a good approximation of the `ground truth', is only modelled with conventional Minkowski-r error measures when the data set adheres to restrictive assumptions, with many real data sets violating these. To combat this there are several methods that use prior knowledge to approximate the `ground truth'. However, prior knowledge is not always available, and this paper investigates how error measures affect the ability for a regression method to model the `ground truth' in these scenarios. Current error measures are shown to create an unhelpful bias and a new error measure is derived which does not exhibit this behaviour. This is tested on 36 representative data sets with different characteristics, showing that it is more consistent in determining the `ground truth' and in giving improved predictions in regions beyond the range of the training data.

This paper addresses the task of (complex) conversational question answering over a knowledge graph. For this task, we propose LASAGNE (muLti-task semAntic parSing with trAnsformer and Graph atteNtion nEtworks). It is the first approach, which employs a transformer architecture extended with Graph Attention Networks for multi-task neural semantic parsing. LASAGNE uses a transformer model for generating the base logical forms, while the Graph Attention model is used to exploit correlations between (entity) types and predicates to produce node representations. LASAGNE also includes a novel entity recognition module which detects, links, and ranks all relevant entities in the question context. We evaluate LASAGNE on a standard dataset for complex sequential question answering, on which it outperforms existing baseline averages on all question types. Specifically, we show that LASAGNE improves the F1-score on eight out of ten question types; in some cases, the increase in F1-score is more than 20% compared to the state of the art.

Calculation of conductivity in the Hubbard model is a challenging task. Recent years have seen much progress in this respect and numerically exact solutions are now possible in certain regimes. In this paper we discuss the calculation of conductivity for the square lattice Hubbard model in the presence of a perpendicular magnetic field, focusing on orbital effects. We present the relevant formalism in all detail and in full generality, and then discuss the simplifications that arise at the level of the dynamical mean field theory (DMFT). We prove that the Kubo bubble preserves gauge and translational invariance, and that in the DMFT the vertex corrections cancel regardless of the magnetic field. We present the DMFT results for the spectral function and both the longitudinal and Hall conductivity in several regimes of parameters. We analyze thoroughly the quantum oscillations of the longitudinal conductivity and identify a high-frequency oscillation component, arising as a combined effect of scattering and temperature, in line with recent experimental observations in moir\'e systems.

Millions of people use platforms such as YouTube, Facebook, Twitter, and other mass media. Due to the accessibility of these platforms, they are often used to establish a narrative, conduct propaganda, and disseminate misinformation. This work proposes an approach that uses state-of-the-art NLP techniques to extract features from video captions (subtitles). To evaluate our approach, we utilize a publicly accessible and labeled dataset for classifying videos as misinformation or not. The motivation behind exploring video captions stems from our analysis of videos metadata. Attributes such as the number of views, likes, dislikes, and comments are ineffective as videos are hard to differentiate using this information. Using caption dataset, the proposed models can classify videos among three classes (Misinformation, Debunking Misinformation, and Neutral) with 0.85 to 0.90 F1-score. To emphasize the relevance of the misinformation class, we re-formulate our classification problem as a two-class classification - Misinformation vs. others (Debunking Misinformation and Neutral). In our experiments, the proposed models can classify videos with 0.92 to 0.95 F1-score and 0.78 to 0.90 AUC ROC.

The 2D TI edge states are considered within the Volkov-Pankratov (VP) Hamiltonian. A smooth transition between TI and OI is assumed. The edge states are formed in the total gap of homogeneous 2D material. A pair of these states are of linear dispersion, others have gapped Dirac spectra. The optical selection rules are found. The optical transitions between the neighboring edge states appear in the global 2D gap for the in-plane light electric field directed across the edge. The electrons in linear edge states have no backscattering, that is indicative of the fact of topological protection. However, when linear edge states get to the energy domain of Dirac edge states, the backscattering becomes permitted. The elastic backscattering rate is found. The Drude-like conductivity is found when the Fermi level gets into the energy domain of the coexistence of linear and Dirac edge states. The localization edge conductance of a finite sample at zero temperature is determined.

This paper compares the advantages, limitations, and computational considerations of using Finite-Time Lyapunov Exponents (FTLEs) and Lagrangian Descriptors (LDs) as tools for identifying barriers and mechanisms of fluid transport in two-dimensional time-periodic flows. These barriers and mechanisms of transport are often referred to as "Lagrangian Coherent Structures," though this term often changes meaning depending on the author or context. This paper will specifically focus on using FTLEs and LDs to identify stable and unstable manifolds of hyperbolic stagnation points, and the Kolmogorov-Arnold-Moser (KAM) tori associated with elliptic stagnation points. The background and theory behind both methods and their associated phase space structures will be presented, and then examples of FTLEs and LDs will be shown based on a simple, periodic, time-dependent double-gyre toy model with varying parameters.

Railway systems provide pivotal support to modern societies, making their efficiency and robustness important to ensure. However, these systems are susceptible to disruptions and delays, leading to accumulating economic damage. The large spatial scale of delay spreading typically make it difficult to distinguish which regions will ultimately affected from an initial disruption, creating uncertainty for risk assessment. In this paper, we identify geographical structures that reflect how delay spreads through railway networks. We do so by proposing a graph-based, hybrid schedule and empirical-based model for delay propagation and apply spectral clustering. We apply the model to four European railway systems: the Netherlands, Germany, Switzerland and Italy. We characterize geographical structures in the railway systems of these countries and interpret these regions in terms of delay severity and how dynamically disconnected they are from the rest. The method also allows us to point out important differences between these countries' railway systems. For practitioners, this geographical characterization of railways provide natural boundaries for local decision-making structures and a first-order prioritization on which regions are at risk, given an initial disruption.

Brain parcellations play a ubiquitous role in the analysis of magnetic resonance imaging (MRI) datasets. Over 100 years of research has been conducted in pursuit of an ideal brain parcellation. Different methods have been developed and studied for constructing brain parcellations using different imaging modalities. More recently, several data-driven parcellation methods have been adopted from data mining, machine learning, and statistics communities. With contributions from different scientific fields, there is a rich body of literature that needs to be examined to appreciate the breadth of existing research and the gaps that need to be investigated. In this work, we review the large body of in vivo brain parcellation research spanning different neuroimaging modalities and methods. A key contribution of this work is a semantic organization of this large body of work into different taxonomies, making it easy to understand the breadth and depth of the brain parcellation literature. Specifically, we categorized the existing parcellations into three groups: Anatomical parcellations, functional parcellations, and structural parcellations which are constructed using T1-weighted MRI, functional MRI (fMRI), and diffusion-weighted imaging (DWI) datasets, respectively. We provide a multi-level taxonomy of different methods studied in each of these categories, compare their relative strengths and weaknesses, and highlight the challenges currently faced for the development of brain parcellations.

Effective molecular representation learning is of great importance to facilitate molecular property prediction, which is a fundamental task for the drug and material industry. Recent advances in graph neural networks (GNNs) have shown great promise in applying GNNs for molecular representation learning. Moreover, a few recent studies have also demonstrated successful applications of self-supervised learning methods to pre-train the GNNs to overcome the problem of insufficient labeled molecules. However, existing GNNs and pre-training strategies usually treat molecules as topological graph data without fully utilizing the molecular geometry information. Whereas, the three-dimensional (3D) spatial structure of a molecule, a.k.a molecular geometry, is one of the most critical factors for determining molecular physical, chemical, and biological properties. To this end, we propose a novel Geometry Enhanced Molecular representation learning method (GEM) for Chemical Representation Learning (ChemRL). At first, we design a geometry-based GNN architecture that simultaneously models atoms, bonds, and bond angles in a molecule. To be specific, we devised double graphs for a molecule: The first one encodes the atom-bond relations; The second one encodes bond-angle relations. Moreover, on top of the devised GNN architecture, we propose several novel geometry-level self-supervised learning strategies to learn spatial knowledge by utilizing the local and global molecular 3D structures. We compare ChemRL-GEM with various state-of-the-art (SOTA) baselines on different molecular benchmarks and exhibit that ChemRL-GEM can significantly outperform all baselines in both regression and classification tasks. For example, the experimental results show an overall improvement of 8.8% on average compared to SOTA baselines on the regression tasks, demonstrating the superiority of the proposed method.

We investigate program equivalence for linear higher-order(sequential) languages endowed with primitives for computational effects. More specifically, we study operationally-based notions of program equivalence for a linear $\lambda$-calculus with explicit copying and algebraic effects \emph{\`a la} Plotkin and Power. Such a calculus makes explicit the interaction between copying and linearity, which are intensional aspects of computation, with effects, which are, instead, \emph{extensional}. We review some of the notions of equivalences for linear calculi proposed in the literature and show their limitations when applied to effectful calculi where copying is a first-class citizen. We then introduce resource transition systems, namely transition systems whose states are built over tuples of programs representing the available resources, as an operational semantics accounting for both intensional and extensional interactive behaviors of programs. Our main result is a sound and complete characterization of contextual equivalence as trace equivalence defined on top of resource transition systems.

One of the exciting recent developments in decentralized finance (DeFi) has been the development of decentralized cryptocurrency exchanges that can autonomously handle conversion between different cryptocurrencies. Decentralized exchange protocols such as Uniswap, Curve and other types of Automated Market Makers (AMMs) maintain a liquidity pool (LP) of two or more assets constrained to maintain at all times a mathematical relation to each other, defined by a given function or curve. Examples of such functions are the constant-sum and constant-product AMMs. Existing systems however suffer from several challenges. They require external arbitrageurs to restore the price of tokens in the pool to match the market price. Such activities can potentially drain resources from the liquidity pool. In particular, dramatic market price changes can result in low liquidity with respect to one or more of the assets and reduce the total value of the LP. We propose in this work a new approach to constructing the AMM by proposing the idea of dynamic curves. It utilizes input from a market price oracle to modify the mathematical relationship between the assets so that the pool price continuously and automatically adjusts to be identical to the market price. This approach eliminates arbitrage opportunities and, as we show through simulations, maintains liquidity in the LP for all assets and the total value of the LP over a wide range of market prices.

We develop a model of interacting zwitterionic membranes with rotating surface dipoles immersed in a monovalent salt, and implement it in a field theoretic formalism. In the mean-field regime of monovalent salt, the electrostatic forces between the membranes are characterized by a non-uniform trend: at large membrane separations, the interfacial dipoles on the opposing sides behave as like-charge cations and give rise to repulsive membrane interactions; at short membrane separations, the anionic field induced by the dipolar phosphate groups sets the behavior in the intermembrane region. The attraction of the cationic nitrogens in the dipolar lipid headgroups leads to the adhesion of the membrane surfaces via dipolar bridging. The underlying competition between the opposing field components of the individual dipolar charges leads to the non-uniform salt ion affinity of the zwitterionic membrane with respect to the separation distance; large inter-membrane separations imply anionic excess while small, nanometer size separations, favor cationic excess. This complex ionic selectivity of zwitterionic membranes may have relevant repercussions on nanofiltration and nanofluidic transport techniques.

Recently the leading order of the correlation energy of a Fermi gas in a coupled mean-field and semiclassical scaling regime has been derived, under the assumption of an interaction potential with a small norm and with compact support in Fourier space. We generalize this result to large interaction potentials, requiring only $|\cdot| \hat{V} \in \ell^1 (\mathbb{Z}^3)$. Our proof is based on approximate, collective bosonization in three dimensions. Significant improvements compared to recent work include stronger bounds on non-bosonizable terms and more efficient control on the bosonization of the kinetic energy.

The localization spread gives a criterion to decide between metallic versus insulating behaviour of a material. It is defined as the second moment cumulant of the many-body position operator, divided by the number of electrons. Different operators are used for systems treated with Open or Periodic Boundary Conditions. In particular, in the case of periodic systems, we use the complex-position definition, that was already used in similar contexts for the treatment of both classical and quantum situations. In this study, we show that the localization spread evaluated on a finite ring system of radius $R$ with Open Boundary Conditions leads, in the large $R$ limit, to the same formula derived by Resta et al. for 1D systems with periodic Born-von K\'arm\'an boundary conditions. A second formula, alternative to the Resta's one, is also given, based on the sum-over-state formalism, allowing for an interesting generalization to polarizability and other similar quantities.

In the recent years, there has been a shift in facial behavior analysis from the laboratory-controlled conditions to the challenging in-the-wild conditions due to the superior performance of deep learning based approaches for many real world applications.However, the performance of deep learning approaches relies on the amount of training data. One of the major problems with data acquisition is the requirement of annotations for large amount of training data. Labeling process of huge training data demands lot of human support with strong domain expertise for facial expressions or action units, which is difficult to obtain in real-time environments.Moreover, labeling process is highly vulnerable to ambiguity of expressions or action units, especially for intensities due to the bias induced by the domain experts. Therefore, there is an imperative need to address the problem of facial behavior analysis with weak annotations. In this paper, we provide a comprehensive review of weakly supervised learning (WSL) approaches for facial behavior analysis with both categorical as well as dimensional labels along with the challenges and potential research directions associated with it. First, we introduce various types of weak annotations in the context of facial behavior analysis and the corresponding challenges associated with it. We then systematically review the existing state-of-the-art approaches and provide a taxonomy of these approaches along with their insights and limitations. In addition, widely used data-sets in the reviewed literature and the performance of these approaches along with evaluation principles are summarized. Finally, we discuss the remaining challenges and opportunities along with the potential research directions in order to apply facial behavior analysis with weak labels in real life situations.

We show that the action of the mapping class group on the space of closed curves of a closed surface effectively tracks the corresponding action on Teichm\"uller space in the following sense: for all but quantitatively few mapping classes, the information of how a mapping class moves a given point of Teichm\"uller space determines, up to a power saving error term, how it changes the geometric intersection numbers of a given closed curve with respect to arbitrary geodesic currents. Applications include an effective estimate describing the speed of convergence of Teichm\"uller geodesic rays to the boundary at infinity of Teichm\"uller space, an effective estimate comparing the Teichm\"uller and Thurston metrics along mapping class group orbits of Teichm\"uller space, and, in the sequel, effective estimates for countings of filling closed geodesics on closed, negatively curved surfaces.

We discuss a model based on dark sector described by non-Abelian $SU(2)_D$ gauge symmetry where we introduce $SU(2)_L \times SU(2)_D$ bi-doublet vector-like leptons to generate active neutrino masses and kinetic mixing between $SU(2)_D$ and $U(1)_Y$ gauge fields at one-loop level. After spontaneous symmetry breaking of $SU(2)_D$, we have remnant $Z_4$ symmetry guaranteeing stability of dark matter candidates. We formulate neutrino mass matrix and related lepton flavor violating processes and discus dark matter physics estimating relic density. It is found that our model realize multicomponent dark matter scenario due to the $Z_4$ symmetry and relic density can be explained by gauge interactions with kinetic mixing effect.

In this article we continue with the research initiated in our previous work on singular Liouville equations with quantized singularity. The main goal of this article is to prove that as long as the bubbling solutions violate the spherical Harnack inequality near a singular source, the first derivatives of coefficient functions must tend to zero.

We compute explicit solutions $\Lambda^\pm_m$ of the Painleve VI (PVI) differential equation from equivariant instanton bundles $E_m$ corresponding to Yang-Mills instantons with "quadrupole symmetry." This is based on a generalization of Hitchin's logarithmic connection to vector bundles with an $SL_2({\mathbb C})$ action. We then identify explicit Okamoto transformation which play the role of "creation operators" for construction $\Lambda^\pm_m$ from the "ground state" $\Lambda^\pm_0$, suggesting that the equivariant instanton bundles $E_m$ might similarly be related to the trivial "ground state" $E_0$.

Machine unlearning has great significance in guaranteeing model security and protecting user privacy. Additionally, many legal provisions clearly stipulate that users have the right to demand model providers to delete their own data from training set, that is, the right to be forgotten. The naive way of unlearning data is to retrain the model without it from scratch, which becomes extremely time and resource consuming at the modern scale of deep neural networks. Other unlearning approaches by refactoring model or training data struggle to gain a balance between overhead and model usability. In this paper, we propose an approach, dubbed as DeepObliviate, to implement machine unlearning efficiently, without modifying the normal training mode. Our approach improves the original training process by storing intermediate models on the hard disk. Given a data point to unlearn, we first quantify its temporal residual memory left in stored models. The influenced models will be retrained and we decide when to terminate the retraining based on the trend of residual memory on-the-fly. Last, we stitch an unlearned model by combining the retrained models and uninfluenced models. We extensively evaluate our approach on five datasets and deep learning models. Compared to the method of retraining from scratch, our approach can achieve 99.0%, 95.0%, 91.9%, 96.7%, 74.1% accuracy rates and 66.7$\times$, 75.0$\times$, 33.3$\times$, 29.4$\times$, 13.7$\times$ speedups on the MNIST, SVHN, CIFAR-10, Purchase, and ImageNet datasets, respectively. Compared to the state-of-the-art unlearning approach, we improve 5.8% accuracy, 32.5$\times$ prediction speedup, and reach a comparable retrain speedup under identical settings on average on these datasets. Additionally, DeepObliviate can also pass the backdoor-based unlearning verification.

Theoretical studies of superradiant lasing on optical clock transitions predict a superb frequency accuracy and precision closely tied to the bare atomic linewidth. Such a superradiant laser is also robust against cavity fluctuations when the spectral width of the lasing mode is much larger than that of the atomic medium. Recent predictions suggest that this unique feature persists even for a hot and thus strongly broadened ensemble, provided the effective atom number is large enough. Here we use a second-order cumulant expansion approach to study the power, linewidth and lineshifts of such a superradiant laser as a function of the inhomogeneous width of the ensemble including variations of the spatial atom-field coupling within the resonator. We present conditions on the atom numbers, the pump and coupling strengths required to reach the buildup of collective atomic coherence as well as scaling and limitations for the achievable laser linewidth.

Many sequence-to-sequence tasks in natural language processing are roughly monotonic in the alignment between source and target sequence, and previous work has facilitated or enforced learning of monotonic attention behavior via specialized attention functions or pretraining. In this work, we introduce a monotonicity loss function that is compatible with standard attention mechanisms and test it on several sequence-to-sequence tasks: grapheme-to-phoneme conversion, morphological inflection, transliteration, and dialect normalization. Experiments show that we can achieve largely monotonic behavior. Performance is mixed, with larger gains on top of RNN baselines. General monotonicity does not benefit transformer multihead attention, however, we see isolated improvements when only a subset of heads is biased towards monotonic behavior.

For hidden Markov models one of the most popular estimates of the hidden chain is the Viterbi path -- the path maximising the posterior probability. We consider a more general setting, called the pairwise Markov model (PMM), where the joint process consisting of finite-state hidden process and observation process is assumed to be a Markov chain. It has been recently proven that under some conditions the Viterbi path of the PMM can almost surely be extended to infinity, thereby defining the infinite Viterbi decoding of the observation sequence, called the Viterbi process. This was done by constructing a block of observations, called a barrier, which ensures that the Viterbi path goes trough a given state whenever this block occurs in the observation sequence. In this paper we prove that the joint process consisting of Viterbi process and PMM is regenerative. The proof involves a delicate construction of regeneration times which coincide with the occurrences of barriers. As one possible application of our theory, some results on the asymptotics of the Viterbi training algorithm are derived.

A generalized Kummer surface $X=Km_{3}(A,G_{A})$ is the minimal resolution of the quotient of a $2$-dimensional complex torus by an order 3 symplectic automorphism group $G_{A}$. A Kummer structure on $X$ is an isomorphism class of pairs $(B,G_{B})$ such that $X\simeq Km_{3}(B,G_{B})$. When the surface is algebraic, we obtain that the number of Kummer structures is linked with the number of order $3$ elliptic points on some Shimura curve naturally related to $A$. For each $n\in\mathbb{N}$, we obtain generalized Kummer surfaces $X_{n}$ for which the number of Kummer structures is $2^{n}$. We then give a classification of the moduli spaces of generalized Kummer surfaces. When the surface is non algebraic, there is only one Kummer structure, but the number of irreducible components of the moduli spaces of such surfaces is large compared to the algebraic case. The endomorphism rings of the complex $2$-tori we study are mainly quaternion orders, these order contain the ring of Eisenstein integers. One can also see this paper as a study of quaternion orders $\mathcal{O}$ over $\mathbb{Q}$ that contain the ring of Eisenstein integers. We obtain that such order is determined up to isomorphism by its discriminant, and when the quaternion algebra is indefinite, the order $\mathcal{O}$ is principal.

The neural network and quantum computing are both significant and appealing fields, with their interactive disciplines promising for large-scale computing tasks that are untackled by conventional computers. However, both developments are restricted by the scope of the hardware development. Nevertheless, many neural network algorithms had been proposed before GPUs become powerful enough for running very deep models. Similarly, quantum algorithms can also be proposed as knowledge reserves before real quantum computers are easily accessible. Specifically, taking advantage of both the neural networks and quantum computation and designing quantum deep neural networks (QDNNs) for acceleration on Noisy Intermediate-Scale Quantum (NISQ) processors is also an important research problem. As one of the most widely used neural network architectures, convolutional neural network (CNN) remains to be accelerated by quantum mechanisms, with only a few attempts have been demonstrated. In this paper, we propose a new hybrid quantum-classical circuit, namely Quantum Fourier Convolutional Network (QFCN). Our model achieves exponential speed-up compared with classical CNN theoretically and improves over the existing best result of quantum CNN. We demonstrate the potential of this architecture by applying it to different deep learning tasks, including traffic prediction and image classification.

Our goal, in the context of open-domain textual question-answering (QA), is to explain answers by showing the line of reasoning from what is known to the answer, rather than simply showing a fragment of textual evidence (a "rationale'"). If this could be done, new opportunities for understanding and debugging the system's reasoning become possible. Our approach is to generate explanations in the form of entailment trees, namely a tree of multipremise entailment steps from facts that are known, through intermediate conclusions, to the hypothesis of interest (namely the question + answer). To train a model with this skill, we created ENTAILMENTBANK, the first dataset to contain multistep entailment trees. Given a hypothesis (question + answer), we define three increasingly difficult explanation tasks: generate a valid entailment tree given (a) all relevant sentences (b) all relevant and some irrelevant sentences, or (c) a corpus. We show that a strong language model can partially solve these tasks, in particular when the relevant sentences are included in the input (e.g., 35% of trees for (a) are perfect), and with indications of generalization to other domains. This work is significant as it provides a new type of dataset (multistep entailments) and baselines, offering a new avenue for the community to generate richer, more systematic explanations.

We study the problem of list-decodable linear regression, where an adversary can corrupt a majority of the examples. Specifically, we are given a set $T$ of labeled examples $(x, y) \in \mathbb{R}^d \times \mathbb{R}$ and a parameter $0< \alpha <1/2$ such that an $\alpha$-fraction of the points in $T$ are i.i.d. samples from a linear regression model with Gaussian covariates, and the remaining $(1-\alpha)$-fraction of the points are drawn from an arbitrary noise distribution. The goal is to output a small list of hypothesis vectors such that at least one of them is close to the target regression vector. Our main result is a Statistical Query (SQ) lower bound of $d^{\mathrm{poly}(1/\alpha)}$ for this problem. Our SQ lower bound qualitatively matches the performance of previously developed algorithms, providing evidence that current upper bounds for this task are nearly best possible.

In this work, we aim to address the 3D scene stylization problem - generating stylized images of the scene at arbitrary novel view angles. A straightforward solution is to combine existing novel view synthesis and image/video style transfer approaches, which often leads to blurry results or inconsistent appearance. Inspired by the high quality results of the neural radiance fields (NeRF) method, we propose a joint framework to directly render novel views with the desired style. Our framework consists of two components: an implicit representation of the 3D scene with the neural radiance field model, and a hypernetwork to transfer the style information into the scene representation. In particular, our implicit representation model disentangles the scene into the geometry and appearance branches, and the hypernetwork learns to predict the parameters of the appearance branch from the reference style image. To alleviate the training difficulties and memory burden, we propose a two-stage training procedure and a patch sub-sampling approach to optimize the style and content losses with the neural radiance field model. After optimization, our model is able to render consistent novel views at arbitrary view angles with arbitrary style. Both quantitative evaluation and human subject study have demonstrated that the proposed method generates faithful stylization results with consistent appearance across different views.

In this paper, we propose a simple yet effective method to deal with the violation of the Closed-World Assumption for a classifier. Previous works tend to apply a threshold either on the classification scores or the loss function to reject the inputs that violate the assumption. However, these methods cannot achieve the low False Positive Ratio (FPR) required in safety applications. The proposed method is a rejection option based on hypothesis testing with probabilistic networks. With probabilistic networks, it is possible to estimate the distribution of outcomes instead of a single output. By utilizing Z-test over the mean and standard deviation for each class, the proposed method can estimate the statistical significance of the network certainty and reject uncertain outputs. The proposed method was experimented on with different configurations of the COCO and CIFAR datasets. The performance of the proposed method is compared with the Softmax Response, which is a known top-performing method. It is shown that the proposed method can achieve a broader range of operation and cover a lower FPR than the alternative.

I describe the rationale for, and design of, an agent-based simulation model of a contemporary online sports-betting exchange: such exchanges, closely related to the exchange mechanisms at the heart of major financial markets, have revolutionized the gambling industry in the past 20 years, but gathering sufficiently large quantities of rich and temporally high-resolution data from real exchanges - i.e., the sort of data that is needed in large quantities for Deep Learning - is often very expensive, and sometimes simply impossible; this creates a need for a plausibly realistic synthetic data generator, which is what this simulation now provides. The simulator, named the "Bristol Betting Exchange" (BBE), is intended as a common platform, a data-source and experimental test-bed, for researchers studying the application of AI and machine learning (ML) techniques to issues arising in betting exchanges; and, as far as I have been able to determine, BBE is the first of its kind: a free open-source agent-based simulation model consisting not only of a sports-betting exchange, but also a minimal simulation model of racetrack sporting events (e.g., horse-races or car-races) about which bets may be made, and a population of simulated bettors who each form their own private evaluation of odds and place bets on the exchange before and - crucially - during the race itself (i.e., so-called "in-play" betting) and whose betting opinions change second-by-second as each race event unfolds. BBE is offered as a proof-of-concept system that enables the generation of large high-resolution data-sets for automated discovery or improvement of profitable strategies for betting on sporting events via the application of AI/ML and advanced data analytics techniques. This paper offers an extensive survey of relevant literature and explains the motivation and design of BBE, and presents brief illustrative results.

Pinch-off and satellite droplets formation during breakup of near-inviscid liquid bridge sandwiched between two given equal and coaxial circular plates have been investigated. The breakup always results in the formation of a spindle shape which is the precursor of the satellite droplet at the moment of pinch-off. Interestingly, the slenderness of this spindle is always bigger than 2{\pi} and always results in the formation of only one satellite droplet regardless of the surface tension and the slenderness of the liquid bridge. We predict the cone angle of this spindle formed during the pinch-off of inviscid fluids should be 18.086122158...{\deg}. After pinch-off, the satellite droplets will drift out of the pinch-off regions in the case of symmetrical short bridge, and merge again with the sessile drop in the case of unsymmetrical long bridge. We demonstrate that the velocity of the satellite droplet is consistent with a scaling model based on a balance between capillary forces and the inertia at the pinch-off region.

In this paper, we generalize fractional $q$-integrals by the method of $q$-difference equation. In addition, we deduce fractional Askey--Wilson integral, reversal type fractional Askey--Wilson integral and Ramanujan type fractional Askey--Wilson integral.

Person Re-Identification (Re-ID) is of great importance to the many video surveillance systems. Learning discriminative features for Re-ID remains a challenge due to the large variations in the image space, e.g., continuously changing human poses, illuminations and point of views. In this paper, we propose HAVANA, a novel extensible, light-weight HierArchical and VAriation-Normalized Autoencoder that learns features robust to intra-class variations. In contrast to existing generative approaches that prune the variations with heavy extra supervised signals, HAVANA suppresses the intra-class variations with a Variation-Normalized Autoencoder trained with no additional supervision. We also introduce a novel Jensen-Shannon triplet loss for contrastive distribution learning in Re-ID. In addition, we present Hierarchical Variation Distiller, a hierarchical VAE to factorize the latent representation and explicitly model the variations. To the best of our knowledge, HAVANA is the first VAE-based framework for person ReID.

We classify Frobenius forms, a special class of homogeneous polynomials in characteristic $p>0$, in up to five variables over an algebraically closed field. We also point out some of the similarities with quadratic forms.

We present optical spectroscopy for 18 halo white dwarfs identified using photometry from the Canada-France Imaging Survey and Pan-STARRS1 DR1 3$\pi$ survey combined with astrometry from Gaia DR2. The sample contains 13 DA, 1 DZ, 2 DC, and two potentially exotic types of white dwarf. We fit both the spectrum and the spectral energy distribution in order to obtain the temperature and surface gravity, which we then convert into a mass, and then an age, using stellar isochrones and the initial-to-final mass relation. We find a large spread in ages that is not consistent with expected formation scenarios for the Galactic halo. We find a mean age of 9.03$^{+2.13}_{-2.03}$ Gyr and a dispersion of 4.21$^{+2.33}_{-1.58}$ Gyr for the inner halo using a maximum likelihood method. This result suggests an extended star formation history within the local halo population.

According to their strength, the tracing properties of a code can be categorized as frameproof, separating, IPP and TA. It is known that if the minimum distance of the code is larger than a certain threshold then the TA property implies the rest. Silverberg et al. ask if there is some kind of tracing capability left when the minimum distance falls below the threshold. Under different assumptions, several papers have given a negative answer to the question. In this paper further progress is made. We establish values of the minimum distance for which Reed-Solomon codes do not posses the separating property.

By using ab-initio-accurate force fields and molecular dynamics simulations we demonstrate that the layer stiffness has profound effects on the superlubricant state of two-dimensional van der Waals heterostructures. These are engineered to have identical inter-layer sliding energy surfaces, but layers of different rigidity, so that the effects of the stiffness on the microscopic friction in the superlubricant state can be isolated. A twofold increase in the intra-layer stiffness reduces the friction by approximately a factor six. Most importantly, we find two sliding regimes as a function of the sliding velocity. At low velocity the heat generated by the motion is efficiently exchanged between the layers and the friction is independent on whether the sliding layer is softer or harder than the substrate. In contrast, at high velocity the friction heat flux cannot be exchanged fast enough, and the build up of significant temperature gradients between the layers is observed. In this situation the temperature profile depends on whether the slider is softer than the substrate.

Previous studies have predicted the failure of Fourier's law of thermal conduction due to the existence of wave like propagation of heat with finite propagation speed. This non-Fourier thermal transport phenomenon can appear in both the hydrodynamic and (quasi) ballistic regimes. Hence, it is not easy to clearly distinguish these two non-Fourier regimes only by this phenomenon. In this work, the transient heat propagation in homogeneous thermal system is studied based on the phonon Boltzmann transport equation (BTE) under the Callaway model. Given a quasi-one or quasi-two (three) dimensional simulation with homogeneous environment temperature, at initial moment, a heat source is added suddenly at the center with high temperature, then the heat propagates from the center to the outer. Numerical results show that in quasi-two (three) dimensional simulations, the transient temperature will be lower than the lowest value of initial temperature in the hydrodynamic regime within a certain range of time and space. This phenomenon appears only when the normal scattering dominates heat conduction. Besides, it disappears in quasi-one dimensional simulations. Similar phenomenon is also observed in thermal systems with time varying heat source. This novel transient heat propagation phenomenon of hydrodynamic phonon transport distinguishes it well from (quasi) ballistic phonon transport.

Despite the advances in the autonomous driving domain, autonomous vehicles (AVs) are still inefficient and limited in terms of cooperating with each other or coordinating with vehicles operated by humans. A group of autonomous and human-driven vehicles (HVs) which work together to optimize an altruistic social utility -- as opposed to the egoistic individual utility -- can co-exist seamlessly and assure safety and efficiency on the road. Achieving this mission without explicit coordination among agents is challenging, mainly due to the difficulty of predicting the behavior of humans with heterogeneous preferences in mixed-autonomy environments. Formally, we model an AV's maneuver planning in mixed-autonomy traffic as a partially-observable stochastic game and attempt to derive optimal policies that lead to socially-desirable outcomes using a multi-agent reinforcement learning framework. We introduce a quantitative representation of the AVs' social preferences and design a distributed reward structure that induces altruism into their decision making process. Our altruistic AVs are able to form alliances, guide the traffic, and affect the behavior of the HVs to handle competitive driving scenarios. As a case study, we compare egoistic AVs to our altruistic autonomous agents in a highway merging setting and demonstrate the emerging behaviors that lead to a noticeable improvement in the number of successful merges as well as the overall traffic flow and safety.

Improving existing widely-adopted prediction models is often a more efficient and robust way towards progress than training new models from scratch. Existing models may (a) incorporate complex mechanistic knowledge, (b) leverage proprietary information and, (c) have surmounted barriers to adoption. Compared to model training, model improvement and modification receive little attention. In this paper we propose a general approach to model improvement: we combine gradient boosting with any previously developed model to improve model performance while retaining important existing characteristics. To exemplify, we consider the context of Mendelian models, which estimate the probability of carrying genetic mutations that confer susceptibility to disease by using family pedigrees and health histories of family members. Via simulations we show that integration of gradient boosting with an existing Mendelian model can produce an improved model that outperforms both that model and the model built using gradient boosting alone. We illustrate the approach on genetic testing data from the USC-Stanford Cancer Genetics Hereditary Cancer Panel (HCP) study.

Traditional on-die, three-level cache hierarchy design is very commonly used but is also prone to latency, especially at the Level 2 (L2) cache. We discuss three distinct ways of improving this design in order to have better performance. Performance is especially important for systems with high workloads. The first method proposes to eliminate L2 altogether while proposing a new prefetching technique, the second method suggests increasing the size of L2, while the last method advocates the implementation of optical caches. After carefully contemplating the results in performance gains and the advantages and disadvantages of each method, we found the last method to be the best of the three.

It is known that general relativity (GR) theory is not consistent with the latest observations. The modified gravity of GR known as $\mathrm{f(R)}$ where $\mathrm{R}$ is the Ricci scalar, is considered to be a good candidate for dealing with the anomalies present in classical GR. In this context, we study static rotating uncharged anti-de Sitter and de Sitter (AdS and dS) black holes (BHs) using $\mathrm{f(R)}$ theory without assuming any constraints on the Ricci scalar or on $\mathrm{f(R)}$. We derive BH solutions depend on the convolution function and deviate from the AdS/dS Schwarzschild BH solution of GR. Although the field equations have no dependence on the cosmological constant, the BHs are characterized by an effective cosmological constant that depends on the convolution function. The asymptotic form of this BH solution depends on the gravitational mass of the system and on extra terms that lead to BHs being different from GR BHs but to correspond to GR BHs under certain conditions. We also investigate how these extra terms are responsible for making the singularities of the invariants milder than those of the GR BHs. We study some physical properties of the BHs from the point of view of thermodynamics and show that there is an outer event horizon in addition to the inner Cauchy horizons. Among other things, we show that our BH solutions satisfy the first law of thermodynamics. To check the stability of these BHs we use the geodesic deviations and derive the stability conditions. Finally, using the odd-type mode it is shown that all the derived BHs are stable and have a radial speed equal to one.

In physical experiments, reference frames are standardly modelled through a specific choice of coordinates used to describe the physical systems, but they themselves are not considered as such. However, any reference frame is a physical system that ultimately behaves according to quantum mechanics. We develop a framework for rotational (i.e. spin) quantum reference frames, with respect to which quantum systems with spin degrees of freedom are described. We give an explicit model for such frames as systems composed of three spin coherent states of angular momentum $j$ and introduce the transformations between them by upgrading the Euler angles occurring in classical $\textrm{SO}(3)$ spin transformations to quantum mechanical operators acting on the states of the reference frames. To ensure that an arbitrary rotation can be applied on the spin we take the limit of infinitely large $j$, in which case the angle operator possesses a continuous spectrum. We prove that rotationally invariant Hamiltonians (such as that of the Heisenberg model) are invariant under a larger group of quantum reference frame transformations. Our result is the first development of the quantum reference frame formalism for a non-Abelian group.

We consider a simple scalar dark matter model within the frame of gauged $L_{\mu}-L_{\tau}$ symmetry. A gauge boson $Z'$ as well as two scalar fields $S$ and $\Phi$ are introduced to the Standard Model (SM). $S$ and $\Phi$ are SM singlet but both with $U(1)_{L_{\mu}-L_{\tau}}$ charge. The real component and imaginary component of $S$ can acquire different masses after spontaneously symmetry breaking, and the lighter one can play the role of dark matter which is stabilized by the residual $Z_2$ symmetry. A viable parameter space is considered to discuss the possibility of light dark matter as well as co-annihilation case, and we present current $(g-2)_{\mu}$ anomaly, Higgs invisible decay, dark matter relic density as well as direct detection constriants on the parameter space.

Decentralized financial (DeFi) applications on the Ethereum blockchain are highly interoperable because they share a single state in a deterministic computational environment. Stakeholders can deposit claims on assets, referred to as 'liquidity shares', across applications producing effects equivalent to rehypothecation in traditional financial systems. We seek to understand the degree to which this practice may contribute to financial integration on Ethereum by examining transactions in 'composed' derivatives for the assets DAI, USDC, USDT, ETH and tokenized BTC for the full set of 344.8 million Ethereum transactions computed in 2020. We identify a salient trend for 'composing' assets in multiple sequential generations of derivatives and comment on potential systemic implications for the Ethereum network.

The paper presents the submission of the team indicnlp@kgp to the EACL 2021 shared task "Offensive Language Identification in Dravidian Languages." The task aimed to classify different offensive content types in 3 code-mixed Dravidian language datasets. The work leverages existing state of the art approaches in text classification by incorporating additional data and transfer learning on pre-trained models. Our final submission is an ensemble of an AWD-LSTM based model along with 2 different transformer model architectures based on BERT and RoBERTa. We achieved weighted-average F1 scores of 0.97, 0.77, and 0.72 in the Malayalam-English, Tamil-English, and Kannada-English datasets ranking 1st, 2nd, and 3rd on the respective tasks.

The goal of this study was to improve the post-processing of precipitation forecasts using convolutional neural networks (CNNs). Instead of post-processing forecasts on a per-pixel basis, as is usually done when employing machine learning in meteorological post-processing, input forecast images were combined and transformed into probabilistic output forecast images using fully convolutional neural networks. CNNs did not outperform regularized logistic regression. Additionally, an ablation analysis was performed. Combining input forecasts from a global low-resolution weather model and a regional high-resolution weather model improved performance over either one.

Deep learning can promote the mammography-based computer-aided diagnosis (CAD) for breast cancers, but it generally suffers from the small sample size problem. Self-supervised learning (SSL) has shown its effectiveness in medical image analysis with limited training samples. However, the network model sometimes cannot be well pre-trained in the conventional SSL framework due to the limitation of the pretext task and fine-tuning mechanism. In this work, a Task-driven Self-supervised Bi-channel Networks (TSBN) framework is proposed to improve the performance of classification model the mammography-based CAD. In particular, a new gray-scale image mapping (GSIM) is designed as the pretext task, which embeds the class label information of mammograms into the image restoration task to improve discriminative feature representation. The proposed TSBN then innovatively integrates different network architecture, including the image restoration network and the classification network, into a unified SSL framework. It jointly trains the bi-channel network models and collaboratively transfers the knowledge from the pretext task network to the downstream task network with improved diagnostic accuracy. The proposed TSBN is evaluated on a public INbreast mammogram dataset. The experimental results indicate that it outperforms the conventional SSL and multi-task learning algorithms for diagnosis of breast cancers with limited samples.

The true topological nature of the Kondo insulator SmB$_6$ remains to be unveiled. Our previous tunneling study not only found evidence for the existence of surface Dirac fermions, but it also uncovered that they inherently interact with the spin excitons, collective excitations in the bulk. We have extended such a spectroscopic investigation into crystals containing a Sm deficiency. The bulk hybridization gap is found to be insensitive to the deficiency up to 1% studied here, but the surface states in Sm-deficient crystals exhibit quite different temperature evolutions from those in stoichiometric ones. We attribute this to the topological surface states remaining incoherent down to the lowest measurement temperature due to their continued interaction with the spin excitons that remain uncondensed. This result shows that the detailed topological nature of SmB$_6$ could vary drastically in the presence of disorder in the lattice. This sensitiveness to disorder is seemingly contradictory to the celebrated topological protection, but it can be understood as being due to the intimate interplay between strong correlations and topological effects.

Effective environmental planning and management to address climate change could be achieved through extensive environmental modeling with machine learning and conventional physical models. In order to develop and improve these models, practitioners and researchers need comprehensive benchmark datasets that are prepared and processed with environmental expertise that they can rely on. This study presents an extensive dataset of rainfall events for the state of Iowa (2016-2019) acquired from the National Weather Service Next Generation Weather Radar (NEXRAD) system and processed by a quantitative precipitation estimation system. The dataset presented in this study could be used for better disaster monitoring, response and recovery by paving the way for both predictive and prescriptive modeling.

This paper is concerned with polynomially generated multiplier invariant subspaces of the weighted Bergman space $A_{\boldsymbol{\beta}}^2$ in infinitely many variables. We completely classify these invariant subspaces under the unitary equivalence. Our results not only cover cases of both the Hardy space $H^{2}(\mathbb{D}_{2}^{\infty})$ and the Bergman space $A^{2}(\mathbb{D}_{2}^{\infty})$ in infinitely many variables, but also apply in finite-variable setting.

The capability of generalization to unseen domains is crucial for deep learning models when considering real-world scenarios. However, current available medical image datasets, such as those for COVID-19 CT images, have large variations of infections and domain shift problems. To address this issue, we propose a prior knowledge driven domain adaptation and a dual-domain enhanced self-correction learning scheme. Based on the novel learning schemes, a domain adaptation based self-correction model (DASC-Net) is proposed for COVID-19 infection segmentation on CT images. DASC-Net consists of a novel attention and feature domain enhanced domain adaptation model (AFD-DA) to solve the domain shifts and a self-correction learning process to refine segmentation results. The innovations in AFD-DA include an image-level activation feature extractor with attention to lung abnormalities and a multi-level discrimination module for hierarchical feature domain alignment. The proposed self-correction learning process adaptively aggregates the learned model and corresponding pseudo labels for the propagation of aligned source and target domain information to alleviate the overfitting to noises caused by pseudo labels. Extensive experiments over three publicly available COVID-19 CT datasets demonstrate that DASC-Net consistently outperforms state-of-the-art segmentation, domain shift, and coronavirus infection segmentation methods. Ablation analysis further shows the effectiveness of the major components in our model. The DASC-Net enriches the theory of domain adaptation and self-correction learning in medical imaging and can be generalized to multi-site COVID-19 infection segmentation on CT images for clinical deployment.