text
null | inputs
dict | prediction
null | prediction_agent
null | annotation
list | annotation_agent
null | multi_label
bool 1
class | explanation
null | id
stringlengths 1
5
| metadata
null | status
stringclasses 2
values | event_timestamp
null | metrics
null |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
null | {
"abstract": " Tensor factorization with hard and/or soft constraints has played an\nimportant role in signal processing and data analysis. However, existing\nalgorithms for constrained tensor factorization have two drawbacks: (i) they\nrequire matrix-inversion; and (ii) they cannot (or at least is very difficult\nto) handle structured regularizations. We propose a new tensor factorization\nalgorithm that circumvents these drawbacks. The proposed method is built upon\nalternating optimization, and each subproblem is solved by a primal-dual\nsplitting algorithm, yielding an efficient and flexible algorithmic framework\nto constrained tensor factorization. The advantages of the proposed method over\na state-of-the-art constrained tensor factorization algorithm, called AO-ADMM,\nare demonstrated on regularized nonnegative tensor factorization.\n",
"title": "Efficient Constrained Tensor Factorization by Alternating Optimization with Primal-Dual Splitting"
} | null | null | null | null | true | null | 19701 | null | Default | null | null |
null | {
"abstract": " In this paper, we propose a replay attack spoofing detection system for\nautomatic speaker verification using multitask learning of noise classes. We\ndefine the noise that is caused by the replay attack as replay noise. We\nexplore the effectiveness of training a deep neural network simultaneously for\nreplay attack spoofing detection and replay noise classification. The\nmulti-task learning includes classifying the noise of playback devices,\nrecording environments, and recording devices as well as the spoofing\ndetection. Each of the three types of the noise classes also includes a genuine\nclass. The experiment results on the ASVspoof2017 datasets demonstrate that the\nperformance of our proposed system is improved by 30% relatively on the\nevaluation set.\n",
"title": "Replay spoofing detection system for automatic speaker verification using multi-task learning of noise classes"
} | null | null | null | null | true | null | 19702 | null | Default | null | null |
null | {
"abstract": " For linear inverse problem with Gaussian random noise we show that Tikhonov\nregularization algorithm is minimax in the class of linear estimators and is\nasymptotically minimax in the sense of sharp asymptotic in the class of all\nestimators. The results are valid if some a priori information on a Fourier\ncoefficients of solution is provided. For trigonometric basis this a priori\ninformation implies that the solution belongs to a ball in Besov space\n$B^r_{2\\infty}$.\n",
"title": "On One Property of Tikhonov Regularization Algorithm"
} | null | null | null | null | true | null | 19703 | null | Default | null | null |
null | {
"abstract": " Light carries momentum which induces on atoms a recoil for each photon\nabsorbed. In vacuum, for a monochromatic beam of frequency $\\nu$, the global\nmomentum per photon is bounded by general principles and is smaller than $h\n\\nu/c$ leading to a limit on the recoil. However, locally this limit can be\nbroken. In this paper, we give a general formula to calculate the recoil in\nvacuum. We show that in a laser beam with a distorted optical field, there are\nregions where the recoil can be higher than this limit. Using atoms placed in\nthose regions we are able to measure directly the extra recoil.\n",
"title": "Observation of Extra Photon Recoil in a Distorted Optical Field"
} | null | null | null | null | true | null | 19704 | null | Default | null | null |
null | {
"abstract": " For a contraction $C_0$-semigroup on a separable Hilbert space, the decay\nrate is estimated by using the weak Poincaré inequalities for the symmetric\nand anti-symmetric part of the generator. As applications, non-exponential\nconvergence rate is characterized for a class of degenerate diffusion\nprocesses, so that the study of hypocoercivity is extended. Concrete examples\nare presented.\n",
"title": "Weak Poincaré Inequalities for Convergence Rate of Degenerate Diffusion Processes"
} | null | null | null | null | true | null | 19705 | null | Default | null | null |
null | {
"abstract": " Traceroute is the main tools to explore Internet path. It provides limited\ninformation about each node along the path. However, Traceroute cannot go\nfurther in statistics analysis, or \\emph{Man-Machine Interface (MMI)}.\nIndeed, there are no graphical tool that is able to draw all paths used by IP\nroutes. We present a new tool that can handle more than 1,000 Traceroute\nresults, map them, identify graphically MPLS links, get information of usage of\nall routes (in percent) to improve the knowledge between countries' links.\nrTraceroute want to go deeper in usage of atomic traces. In this paper, we will\ndiscuss the concept of rTraceroute and present some example of usage.\n",
"title": "rTraceroute: Réunion Traceroute Visualisation"
} | null | null | null | null | true | null | 19706 | null | Default | null | null |
null | {
"abstract": " We studied the thermodynamic behaviors of non-interacting bosons and fermions\ntrapped by a scale-invariant branching structure of adjustable degree of\nheterogeneity. The full energy spectrum in tight-binding approximation was\nanalytically solved . We found that the log-periodic oscillation of the\nspecific heat for Fermi gas depended on the heterogeneity of hopping. Also, low\ndimensional Bose-Einstein condensation occurred only for non-homogeneous setup.\n",
"title": "Effects of heterogeneity in site-site couplings for tight-binding models on scale-invariant structures"
} | null | null | null | null | true | null | 19707 | null | Default | null | null |
null | {
"abstract": " Many deep models have been recently proposed for anomaly detection. This\npaper presents comparison of selected generative deep models and classical\nanomaly detection methods on an extensive number of non--image benchmark\ndatasets. We provide statistical comparison of the selected models, in many\nconfigurations, architectures and hyperparamaters. We arrive to conclusion that\nperformance of the generative models is determined by the process of selection\nof their hyperparameters. Specifically, performance of the deep generative\nmodels deteriorates with decreasing amount of anomalous samples used in\nhyperparameter selection. In practical scenarios of anomaly detection, none of\nthe deep generative models systematically outperforms the kNN.\n",
"title": "Are generative deep models for novelty detection truly better?"
} | null | null | null | null | true | null | 19708 | null | Default | null | null |
null | {
"abstract": " Consider a terminal in which users arrive continuously over a finite period\nof time at a variable rate known in advance. A fleet of shuttles has to carry\nthe users over a fixed trip. What is the shuttle schedule that minimizes their\nwaiting time? This is the question addressed in the present paper. We propose\nefficient algorithms for several variations of this question with proven\nperformance guarantees. The techniques used are of various types (convex\noptimization, shortest paths,...). The paper ends with numerical experiments\nshowing that most of our algorithms behave also well in practice.\n",
"title": "Minimizing the waiting time for a one-way shuttle service"
} | null | null | null | null | true | null | 19709 | null | Default | null | null |
null | {
"abstract": " In this paper we extend the theory of two weight, $A_p$ bump conditions to\nthe setting of matrix weights. We prove two matrix weight inequalities for\nfractional maximal operators, fractional and singular integrals, sparse\noperators and averaging operators. As applications we prove quantitative, one\nweight estimates, in terms of the matrix $A_p$ constant, for singular\nintegrals, and prove a Poincaré inequality related to those that appear in\nthe study of degenerate elliptic PDEs.\n",
"title": "Two weight bump conditions for matrix weights"
} | null | null | null | null | true | null | 19710 | null | Default | null | null |
null | {
"abstract": " We show that a link in an open book can be realized as a strongly\nquasipositive braid if and only if it bounds a Legendrian ribbon with respect\nto the associated contact structure. This generalizes a result due to Baader\nand Ishikawa for links in the three-sphere. We highlight some related\ntechniques for determining whether or not a link is strongly quasipositive,\nemphasizing applications to fibered links and satellites.\n",
"title": "Legendrian ribbons and strongly quasipositive links in an open book"
} | null | null | null | null | true | null | 19711 | null | Default | null | null |
null | {
"abstract": " The \\Delta-convolution of real probability measures, introduced by Bożejko,\ngeneralizes both free and boolean convolutions. It is linearized by the\n\\Delta-cumulants, and Yoshida gave a combinatorial formula for moments in terms\nof \\Delta-cumulants, that implicitly defines the latter. It relies on the\ndefinition of an appropriate weight on noncrossing partitions. We give here two\ndifferent expressions for the \\Delta-cumulants: the first one is a simple\nvariant of Lagrange inversion formula, and the second one is a combinatorial\ninversion of Yoshida's formula involving Schröder trees.\n",
"title": "Δ-cumulants in terms of moments"
} | null | null | [
"Mathematics"
]
| null | true | null | 19712 | null | Validated | null | null |
null | {
"abstract": " When classifying point clouds, a large amount of time is devoted to the\nprocess of engineering a reliable set of features which are then passed to a\nclassifier of choice. Generally, such features - usually derived from the\n3D-covariance matrix - are computed using the surrounding neighborhood of\npoints. While these features capture local information, the process is usually\ntime-consuming, and requires the application at multiple scales combined with\ncontextual methods in order to adequately describe the diversity of objects\nwithin a scene. In this paper we present a 1D-fully convolutional network that\nconsumes terrain-normalized points directly with the corresponding spectral\ndata,if available, to generate point-wise labeling while implicitly learning\ncontextual features in an end-to-end fashion. Our method uses only the\n3D-coordinates and three corresponding spectral features for each point.\nSpectral features may either be extracted from 2D-georeferenced images, as\nshown here for Light Detection and Ranging (LiDAR) point clouds, or extracted\ndirectly for passive-derived point clouds,i.e. from muliple-view imagery. We\ntrain our network by splitting the data into square regions, and use a pooling\nlayer that respects the permutation-invariance of the input points. Evaluated\nusing the ISPRS 3D Semantic Labeling Contest, our method scored second place\nwith an overall accuracy of 81.6%. We ranked third place with a mean F1-score\nof 63.32%, surpassing the F1-score of the method with highest accuracy by\n1.69%. In addition to labeling 3D-point clouds, we also show that our method\ncan be easily extended to 2D-semantic segmentation tasks, with promising\ninitial results.\n",
"title": "A Fully Convolutional Network for Semantic Labeling of 3D Point Clouds"
} | null | null | null | null | true | null | 19713 | null | Default | null | null |
null | {
"abstract": " Some boundedness properties of function spaces (considered as topological\ngroups) are studied.\n",
"title": "Boundedness properties in function spaces"
} | null | null | null | null | true | null | 19714 | null | Default | null | null |
null | {
"abstract": " In this paper, we propose a novel approach to combine \\emph{compact directed\nacyclic word graphs} (CDAWGs) and grammar-based compression. This leads us to\nan efficient self-index, called Linear-size CDAWGs (L-CDAWGs), which can be\nrepresented with $O(\\tilde e_T \\log n)$ bits of space allowing for $O(\\log\nn)$-time random and $O(1)$-time sequential accesses to edge labels, and $O(m\n\\log \\sigma + occ)$-time pattern matching. Here, $\\tilde e_T$ is the number of\nall extensions of maximal repeats in $T$, $n$ and $m$ are respectively the\nlengths of the text $T$ and a given pattern, $\\sigma$ is the alphabet size, and\n$occ$ is the number of occurrences of the pattern in $T$. The repetitiveness\nmeasure $\\tilde e_T$ is known to be much smaller than the text length $n$ for\nhighly repetitive text. For constant alphabets, our L-CDAWGs achieve $O(m +\nocc)$ pattern matching time with $O(e_T^r \\log n)$ bits of space, which\nimproves the pattern matching time of Belazzougui et al.'s run-length\nBWT-CDAWGs by a factor of $\\log \\log n$, with the same space complexity. Here,\n$e_T^r$ is the number of right extensions of maximal repeats in $T$. As a\nbyproduct, our result gives a way of constructing an SLP of size $O(\\tilde\ne_T)$ for a given text $T$ in $O(n + \\tilde e_T \\log \\sigma)$ time.\n",
"title": "Linear-size CDAWG: new repetition-aware indexing and grammar compression"
} | null | null | null | null | true | null | 19715 | null | Default | null | null |
null | {
"abstract": " Topic lifecycle analysis on Twitter, a branch of study that investigates\nTwitter topics from their birth through lifecycle to death, has gained immense\nmainstream research popularity. In the literature, topics are often treated as\none of (a) hashtags (independent from other hashtags), (b) a burst of keywords\nin a short time span or (c) a latent concept space captured by advanced text\nanalysis methodologies, such as Latent Dirichlet Allocation (LDA). The first\ntwo approaches are not capable of recognizing topics where different users use\ndifferent hashtags to express the same concept (semantically related), while\nthe third approach misses out the user's explicit intent expressed via\nhashtags. In our work, we use a word embedding based approach to cluster\ndifferent hashtags together, and the temporal concurrency of the hashtag\nusages, thus forming topics (a semantically and temporally related group of\nhashtags).We present a novel analysis of topic lifecycles with respect to\ncommunities. We characterize the participation of social communities in the\ntopic clusters, and analyze the lifecycle of topic clusters with respect to\nsuch participation. We derive first-of-its-kind novel insights with respect to\nthe complex evolution of topics over communities and time: temporal morphing of\ntopics over hashtags within communities, how the hashtags die in some\ncommunities but morph into some other hashtags in some other communities (that,\nit is a community-level phenomenon), and how specific communities adopt to\nspecific hashtags. Our work is fundamental in the space of topic lifecycle\nmodeling and understanding in communities: it redefines our understanding of\ntopic lifecycles and shows that the social boundaries of topic lifecycles are\ndeeply ingrained with community behavior.\n",
"title": "Topic Lifecycle on Social Networks: Analyzing the Effects of Semantic Continuity and Social Communities"
} | null | null | null | null | true | null | 19716 | null | Default | null | null |
null | {
"abstract": " Previous research on unstable footwear has suggested that it may induce\nplantar mechanical noise during walking. The purpose of this study was to\nexplore whether unstable footwear could be considered as a noise-based training\ngear to exercise body center of mass (CoM) motion during walking or not. Ground\nreaction forces were collected among 24 healthy young women walking at speeds\nbetween 3 and 6 km h-1 with control running shoes and unstable rocker-bottom\nshoes. The external mechanical work, the recovery of mechanical energy of the\nCoM during and within the step cycles, and the phase shift between potential\nand kinetic energy curves of the CoM were computed. Our findings support the\nidea that unstable rocker-bottom footwear could serve as a speed-dependent\nnoise- based training gear to exercise CoM motion during walking. At slow\nspeed, it acts as a stochastic resonance or facilitator, whereas at brisk speed\nit acts as a constraint.\n",
"title": "Unstable Footwear as a Speed-Dependent Noise-Based Training Gear to Exercise Inverted Pendulum Motion During Walking"
} | null | null | null | null | true | null | 19717 | null | Default | null | null |
null | {
"abstract": " In this paper we introduce a modified version of a gaussian standard\nfirst-order autoregressive process where we allow for a dependence structure\nbetween the state variable $Y_{t-1}$ and the next innovation $\\xi_t$. We call\nthis model dependent innovations gaussian AR(1) process (DIG-AR(1)). We analyze\nthe moment and temporal dependence properties of the new model. After proving\nthat the OLS estimator does not consistently estimate the autoregressive\nparameter, we introduce an infeasible estimator and we provide its\n$\\sqrt{T}$-asymptotic normality.\n",
"title": "Gaussian autoregressive process with dependent innovations. Some asymptotic results"
} | null | null | null | null | true | null | 19718 | null | Default | null | null |
null | {
"abstract": " Structured Peer Learning (SPL) is a form of peer-based supplemental\ninstruction that focuses on mentoring, guidance, and development of technical,\ncommunication, and social skills in both the students receiving assistance and\nthe students in teaching roles. This paper explores the methodology, efficacy,\nand reasoning behind the practical realization of a SPL program designed to\nincrease student knowledge and success in undergraduate Computer Science\ncourses. Students expressed an increased level of comfort when asking for help\nfrom student teachers versus traditional educational resources, historically\nshowed an increased average grade in lower-level courses, and felt that the\nprogram positively impacted their desire to continue in or switch to a Computer\nmajor. Additionally, results indicated that advances in programming, analytical\nthinking, and abstract analysis skills were evident in not only the students\nbut also the student teachers, suggesting a strong bidirectional flow of\nknowledge.\n",
"title": "Structured Peer Learning Program - An Innovative Approach to Computer Science Education"
} | null | null | null | null | true | null | 19719 | null | Default | null | null |
null | {
"abstract": " The monograph represents analysis of the possibility for time-optimal control\nof a prototype of an aircraft-mounted scanning-imaging system of a Light\nDetection and Ranging based wake vortex detection system. The study is a part\nof the research project \"Demonstration of Light Detection and Ranging based\nwake vortex detection system incorporating an atmospheric hazard map\" or\nGREEN-WAKE (Project ID 213254) of the European Union Seventh Framework Program\nfor Research and Technological Development. The scanning system comprises two\nlight mirror actuators. The study is decomposed into several group of problems.\nThe first and second groups consider the mathematical models of the scanning\nsystem and the mirror actuators. The third group of problems deals with the\ndesign of closed loop tracking control systems of both the large and small\nmirror actuators. The control of each one system is synthesized as a near\ntime-optimal control of the precise linear model of the mirror actuator. The\ncontrol algorithms realize the state of the art method for synthesis of\ntime-optimal control of any order for a class of linear time-optimal control\nproblems developed in the author's dissertation. The last discovers and\ntheoretically proves new properties of a class of linear time-optimal control\nproblems. From the point of view of the control synthesis algorithms the main\nadvantage is that the time-optimal control is produced by a multistage\nprocedure within the class of problems but without the need of describing the\nhyper-surfaces of switching over. The fourth group of problems considers\nmodeling the real scan picture but with inclusion also of the Coulomb's\nfriction model. The fifth group of problems investigates the ways of\nimprovement of the real scan picture. As a result an excellent repetition and\nclearness of the scanning alongside with symmetry and matching the scan pattern\nare seen.\n",
"title": "Analysis of the possibility for time-optimal control of the scanning system of the GREEN-WAKE's project lidar"
} | null | null | null | null | true | null | 19720 | null | Default | null | null |
null | {
"abstract": " Recently, researchers have discovered that the state-of-the-art object\nclassifiers can be fooled easily by small perturbations in the input\nunnoticeable to human eyes. It is also known that an attacker can generate\nstrong adversarial examples if she knows the classifier parameters. Conversely,\na defender can robustify the classifier by retraining if she has access to the\nadversarial examples. We explain and formulate this adversarial example problem\nas a two-player continuous zero-sum game, and demonstrate the fallacy of\nevaluating a defense or an attack as a static problem. To find the best\nworst-case defense against whitebox attacks, we propose a continuous minimax\noptimization algorithm. We demonstrate the minimax defense with two types of\nattack classes -- gradient-based and neural network-based attacks. Experiments\nwith the MNIST and the CIFAR-10 datasets demonstrate that the defense found by\nnumerical minimax optimization is indeed more robust than non-minimax defenses.\nWe discuss directions for improving the result toward achieving robustness\nagainst multiple types of attack classes.\n",
"title": "Machine vs Machine: Minimax-Optimal Defense Against Adversarial Examples"
} | null | null | null | null | true | null | 19721 | null | Default | null | null |
null | {
"abstract": " Over the last decades, the Internet and mobile technology have consolidated\nthe digital as a public sphere of life. Designers are asked to create engaging\ndigital experiences. However, in some cases engagement is seen as a\npsychological state, while in others it emphasizes a participative vein. In\nthis paper, I review and discuss both and propose a new definition to clarify\nthe concept engagement with computers. Thus, engagement is a quality of an\nactive connection between a user and a computing product, either a website or a\nmobile phone app. Studying it requires understanding a set of aspects like the\nuser's affect, motivation and attention, as well as the product's design,\ncontent and composition. Finally, I propose explaining these concepts aligned\nwith engagement and integrate them into a preliminary model to measure the\nmanifestations.\n",
"title": "From Attention to Participation: Reviewing and Modelling Engagement with Computers"
} | null | null | null | null | true | null | 19722 | null | Default | null | null |
null | {
"abstract": " Unexpected clustering in the orbital elements of minor bodies beyond the\nKuiper belt has led to speculations that our solar system actually hosts nine\nplanets, the eight established plus a hypothetical \"Planet Nine\". Several\nrecent studies have shown that a planet with a mass of about 10 Earth masses on\na distant eccentric orbit with perihelion far beyond the Kuiper belt could\ncreate and maintain this clustering. The evolutionary path resulting in an\norbit such as the one suggested for Planet Nine is nevertheless not easily\nexplained. Here we investigate whether a planet scattered away from the\ngiant-planet region could be lifted to an orbit similar to the one suggested\nfor Planet Nine through dynamical friction with a cold, distant planetesimal\nbelt. Recent simulations of planetesimal formation via the streaming\ninstability suggest that planetesimals can readily form beyond 100au. We\nexplore this circularisation by dynamical friction with a set of numerical\nsimulations. We find that a planet that is scattered from the region close to\nNeptune onto an eccentric orbit has a 20-30% chance of obtaining an orbit\nsimilar to that of Planet Nine after 4.6Gyr. Our simulations also result in\nstrong or partial clustering of the planetesimals; however, whether or not this\nclustering is observable depends on the location of the inner edge of the\nplanetesimal belt. If the inner edge is located at 200au the degree of\nclustering amongst observable objects is significant.\n",
"title": "Circularizing Planet Nine through dynamical friction with an extended, cold planetesimal belt"
} | null | null | null | null | true | null | 19723 | null | Default | null | null |
null | {
"abstract": " We introduce two tactics to attack agents trained by deep reinforcement\nlearning algorithms using adversarial examples, namely the strategically-timed\nattack and the enchanting attack. In the strategically-timed attack, the\nadversary aims at minimizing the agent's reward by only attacking the agent at\na small subset of time steps in an episode. Limiting the attack activity to\nthis subset helps prevent detection of the attack by the agent. We propose a\nnovel method to determine when an adversarial example should be crafted and\napplied. In the enchanting attack, the adversary aims at luring the agent to a\ndesignated target state. This is achieved by combining a generative model and a\nplanning algorithm: while the generative model predicts the future states, the\nplanning algorithm generates a preferred sequence of actions for luring the\nagent. A sequence of adversarial examples is then crafted to lure the agent to\ntake the preferred sequence of actions. We apply the two tactics to the agents\ntrained by the state-of-the-art deep reinforcement learning algorithm including\nDQN and A3C. In 5 Atari games, our strategically timed attack reduces as much\nreward as the uniform attack (i.e., attacking at every time step) does by\nattacking the agent 4 times less often. Our enchanting attack lures the agent\ntoward designated target states with a more than 70% success rate. Videos are\navailable at this http URL\n",
"title": "Tactics of Adversarial Attack on Deep Reinforcement Learning Agents"
} | null | null | null | null | true | null | 19724 | null | Default | null | null |
null | {
"abstract": " One of the big challenges in machine learning applications is that training\ndata can be different from the real-world data faced by the algorithm. In\nlanguage modeling, users' language (e.g. in private messaging) could change in\na year and be completely different from what we observe in publicly available\ndata. At the same time, public data can be used for obtaining general knowledge\n(i.e. general model of English). We study approaches to distributed fine-tuning\nof a general model on user private data with the additional requirements of\nmaintaining the quality on the general data and minimization of communication\ncosts. We propose a novel technique that significantly improves prediction\nquality on users' language compared to a general model and outperforms gradient\ncompression methods in terms of communication efficiency. The proposed\nprocedure is fast and leads to an almost 70% perplexity reduction and 8.7\npercentage point improvement in keystroke saving rate on informal English\ntexts. We also show that the range of tasks our approach is applicable to is\nnot limited by language modeling only. Finally, we propose an experimental\nframework for evaluating differential privacy of distributed training of\nlanguage models and show that our approach has good privacy guarantees.\n",
"title": "Differentially Private Distributed Learning for Language Modeling Tasks"
} | null | null | null | null | true | null | 19725 | null | Default | null | null |
null | {
"abstract": " In this paper, we propose a Tensor Train Neighborhood Preserving Embedding\n(TTNPE) to embed multi-dimensional tensor data into low dimensional tensor\nsubspace. Novel approaches to solve the optimization problem in TTNPE are\nproposed. For this embedding, we evaluate novel trade-off gain among\nclassification, computation, and dimensionality reduction (storage) for\nsupervised learning. It is shown that compared to the state-of-the-arts tensor\nembedding methods, TTNPE achieves superior trade-off in classification,\ncomputation, and dimensionality reduction in MNIST handwritten digits and\nWeizmann face datasets.\n",
"title": "Tensor Train Neighborhood Preserving Embedding"
} | null | null | null | null | true | null | 19726 | null | Default | null | null |
null | {
"abstract": " Three dimensional (3D) topology optimization problems always involve huge\nnumbers of Degrees of Freedom (DOFs) in finite element analysis (FEA) and\ndesign variables in numerical optimization, respectively. This will inevitably\nlead to large computational efforts in the solution process. In the present\npaper, an efficient and explicit topology optimization approach which can\nreduce not only the number of design variables but also the number of degrees\nof freedom in FEA is proposed based on the Moving Morphable Voids (MMVs)\nsolution framework. This is achieved by introducing a set of geometry\nparameters (e.g., control points of B-spline surfaces) to describe the boundary\nof a structure explicitly and removing the unnecessary DOFs from the FE model\nat every step of numerical optimization. Numerical examples demonstrate that\nthe proposed approach does can overcome the bottleneck problems associated with\na 3D topology optimization problem in a straightforward way and enhance the\nsolution efficiency significantly.\n",
"title": "Explicit three dimensional topology optimization via Moving Morphable Void (MMV) approach"
} | null | null | null | null | true | null | 19727 | null | Default | null | null |
null | {
"abstract": " Let M be an irreducible Riemannian symmetric space. The index i(M) of M is\nthe minimal codimension of a (non-trivial) totally geodesic submanifold of M.\nThe purpose of this note is to determine the index i(M) for all irreducible\nRiemannian symmetric spaces M of type (II) and (IV).\n",
"title": "The index of compact simple Lie groups"
} | null | null | null | null | true | null | 19728 | null | Default | null | null |
null | {
"abstract": " Microscopy imaging plays a vital role in understanding many biological\nprocesses in development and disease. The recent advances in automation of\nmicroscopes and development of methods and markers for live cell imaging has\nled to rapid growth in the amount of image data being captured. To efficiently\nand reliably extract useful insights from these captured sequences, automated\ncell tracking is essential. This is a challenging problem due to large\nvariation in the appearance and shapes of cells depending on many factors\nincluding imaging methodology, biological characteristics of cells, cell matrix\ncomposition, labeling methodology, etc. Often cell tracking methods require a\nsequence-specific segmentation method and manual tuning of many tracking\nparameters, which limits their applicability to sequences other than those they\nare designed for. In this paper, we propose 1) a deep learning based cell\nproposal method, which proposes candidates for cells along with their scores,\nand 2) a cell tracking method, which links proposals in adjacent frames in a\ngraphical model using edges representing different cellular events and poses\njoint cell detection and tracking as the selection of a subset of cell and edge\nproposals. Our method is completely automated and given enough training data\ncan be applied to a wide variety of microscopy sequences. We evaluate our\nmethod on multiple fluorescence and phase contrast microscopy sequences\ncontaining cells of various shapes and appearances from ISBI cell tracking\nchallenge, and show that our method outperforms existing cell tracking methods.\nCode is available at: this https URL\n",
"title": "Cell Tracking via Proposal Generation and Selection"
} | null | null | null | null | true | null | 19729 | null | Default | null | null |
null | {
"abstract": " In this paper, we study geometric properties of basins of attraction of\nmonotone systems. Our results are based on a combination of monotone systems\ntheory and spectral operator theory. We exploit the framework of the Koopman\noperator, which provides a linear infinite-dimensional description of nonlinear\ndynamical systems and spectral operator-theoretic notions such as eigenvalues\nand eigenfunctions. The sublevel sets of the dominant eigenfunction form a\nfamily of nested forward-invariant sets and the basin of attraction is the\nlargest of these sets. The boundaries of these sets, called isostables, allow\nstudying temporal properties of the system. Our first observation is that the\ndominant eigenfunction is increasing in every variable in the case of monotone\nsystems. This is a strong geometric property which simplifies the computation\nof isostables. We also show how variations in basins of attraction can be\nbounded under parametric uncertainty in the vector field of monotone systems.\nFinally, we study the properties of the parameter set for which a monotone\nsystem is multistable. Our results are illustrated on several systems of two to\nfour dimensions.\n",
"title": "Geometric Properties of Isostables and Basins of Attraction of Monotone Systems"
} | null | null | null | null | true | null | 19730 | null | Default | null | null |
null | {
"abstract": " The generation of anisotropic shapes occurs during morphogenesis of almost\nall organisms. With the recent renewal of the interest in mechanical aspects of\nmorphogenesis, it has become clear that mechanics contributes to anisotropic\nforms in a subtle interaction with various molecular actors. Here, we consider\nplants, fungi, oomycetes, and bacteria, and we review the mechanisms by which\nelongated shapes are generated and maintained. We focus on theoretical models\nof the interplay between growth and mechanics, in relation with experimental\ndata, and discuss how models may help us improve our understanding of the\nunderlying biological mechanisms.\n",
"title": "Elongation and shape changes in organisms with cell walls: a dialogue between experiments and models"
} | null | null | null | null | true | null | 19731 | null | Default | null | null |
null | {
"abstract": " We extend the notion of localic completion of generalised metric spaces by\nSteven Vickers to the setting of generalised uniform spaces. A generalised\nuniform space (gus) is a set X equipped with a family of generalised metrics on\nX, where a generalised metric on X is a map from the product of X to the upper\nreals satisfying zero self-distance law and triangle inequality.\nFor a symmetric generalised uniform space, the localic completion lifts its\ngeneralised uniform structure to a point-free generalised uniform structure.\nThis point-free structure induces a complete generalised uniform structure on\nthe set of formal points of the localic completion that gives the standard\ncompletion of the original gus with Cauchy filters.\nWe extend the localic completion to a full and faithful functor from the\ncategory of locally compact uniform spaces into that of overt locally compact\ncompletely regular formal topologies. Moreover, we give an elementary\ncharacterisation of the cover of the localic completion of a locally compact\nuniform space that simplifies the existing characterisation for metric spaces.\nThese results generalise the corresponding results for metric spaces by Erik\nPalmgren.\nFurthermore, we show that the localic completion of a symmetric gus is\nequivalent to the point-free completion of the uniform formal topology\nassociated with the gus.\nWe work in Aczel's constructive set theory CZF with the Regular Extension\nAxiom. Some of our results also require Countable Choice.\n",
"title": "Localic completion of uniform spaces"
} | null | null | null | null | true | null | 19732 | null | Default | null | null |
null | {
"abstract": " According to magnetohydrodynamics (MHD), the encounter of two collisional\nmagnetized plasmas at high velocity gives rise to shock waves. Investigations\nconducted so far have found that the same conclusion still holds in the case of\ncollisionless plasmas. For the case of a flow-aligned field, MHD stipulates\nthat the field and the fluid are disconnected, so that the shock produced is\nindependent of the field. We present a violation of this MHD prediction when\nconsidering the encounter of two cold pair plasmas along a flow-aligned\nmagnetic field. As the guiding magnetic field grows, isotropization is\nprogressively suppressed, resulting in a strong influence of the field on the\nresulting structure. A micro-physics analysis allows to understand the\nmechanisms at work. Particle-in-cell simulations also support our conclusions\nand show that the results are not restricted to a strictly parallel field.\n",
"title": "Kinetic inhibition of MHD-shocks in the vicinity of a parallel magnetic field"
} | null | null | null | null | true | null | 19733 | null | Default | null | null |
null | {
"abstract": " In this paper, we propose Squeezed Convolutional Variational AutoEncoder\n(SCVAE) for anomaly detection in time series data for Edge Computing in\nIndustrial Internet of Things (IIoT). The proposed model is applied to labeled\ntime series data from UCI datasets for exact performance evaluation, and\napplied to real world data for indirect model performance comparison. In\naddition, by comparing the models before and after applying Fire Modules from\nSqueezeNet, we show that model size and inference times are reduced while\nsimilar levels of performance is maintained.\n",
"title": "Squeezed Convolutional Variational AutoEncoder for Unsupervised Anomaly Detection in Edge Device Industrial Internet of Things"
} | null | null | null | null | true | null | 19734 | null | Default | null | null |
null | {
"abstract": " Most state-of-the-art graph kernels only take local graph properties into\naccount, i.e., the kernel is computed with regard to properties of the\nneighborhood of vertices or other small substructures. On the other hand,\nkernels that do take global graph propertiesinto account may not scale well to\nlarge graph databases. Here we propose to start exploring the space between\nlocal and global graph kernels, striking the balance between both worlds.\nSpecifically, we introduce a novel graph kernel based on the $k$-dimensional\nWeisfeiler-Lehman algorithm. Unfortunately, the $k$-dimensional\nWeisfeiler-Lehman algorithm scales exponentially in $k$. Consequently, we\ndevise a stochastic version of the kernel with provable approximation\nguarantees using conditional Rademacher averages. On bounded-degree graphs, it\ncan even be computed in constant time. We support our theoretical results with\nexperiments on several graph classification benchmarks, showing that our\nkernels often outperform the state-of-the-art in terms of classification\naccuracies.\n",
"title": "Global Weisfeiler-Lehman Graph Kernels"
} | null | null | null | null | true | null | 19735 | null | Default | null | null |
null | {
"abstract": " Quantitative nuclear magnetic resonance imaging (MRI) shifts more and more\ninto the focus of clinical research. Especially determination of relaxation\ntimes without/and with contrast agents becomes the foundation of tissue\ncharacterization, e.g. in cardiac MRI for myocardial fibrosis. Techniques which\nassess longitudinal relaxation times rely on repetitive application of readout\nmodules, which are interrupted by free relaxation periods, e.g. the Modified\nLook-Locker Inversion Recovery = MOLLI sequence. These discontinuous sequences\nreveal an apparent relaxation time, and, by techniques extrapolated from\ncontinuous readout sequences, the real T1 is determined. What is missing is a\nrigorous analysis of the dependence of the apparent relaxation time on its real\npartner, readout sequence parameters and biological parameters as heart rate.\nThis is provided in this paper for the discontinuous balanced steady state free\nprecession (bSSFP) and spoiled gradient echo readouts. It turns out that the\napparente longitudinal relaxation rate is the time average of the relaxation\nrates during the readout module, and free relaxation period. Knowing the heart\nrate our results vice versa allow to determine the real T1 from its measured\napparent partner.\n",
"title": "An analytical Model which Determines the Apparent T1 for Modified Look-Locker Inversion Recovery (MOLLI) -- Analysis of the Longitudinal Relaxation under the Influence of Discontinuous Balanced and Spoiled Gradient Echo Readouts"
} | null | null | null | null | true | null | 19736 | null | Default | null | null |
null | {
"abstract": " We present an updated halo-dependent and halo-independent analysis of viable\nlight WIMP dark matter candidates which could account for the excess observed\nin CDMS-II-Si. We include recent constraints from LUX, PandaX-II, and PICO-60,\nas well as projected sensitivities for XENON1T, SuperCDMS SNOLAB, LZ, DARWIN,\nDarkSide-20k, and PICO-250, on candidates with spin-independent isospin\nconserving and isospin-violating interactions, and either elastic or exothermic\nscattering. We show that there exist dark matter candidates which can explain\nthe CDMS-II-Si data and remain very marginally consistent with the null results\nof all current experiments, however such models are highly tuned, making a dark\nmatter interpretation of CDMS-II-Si very unlikely. We find that these models\ncan only be ruled out in the future by an experiment comparable to LZ or\nPICO-250.\n",
"title": "Updated Constraints on the Dark Matter Interpretation of CDMS-II-Si Data"
} | null | null | null | null | true | null | 19737 | null | Default | null | null |
null | {
"abstract": " Current and upcoming radio interferometric experiments are aiming to make a\nstatistical characterization of the high-redshift 21cm fluctuation signal\nspanning the hydrogen reionization and X-ray heating epochs of the universe.\nHowever, connecting 21cm statistics to underlying physical parameters is\ncomplicated by the theoretical challenge of modeling the relevant physics at\ncomputational speeds quick enough to enable exploration of the high dimensional\nand weakly constrained parameter space. In this work, we use machine learning\nalgorithms to build a fast emulator that mimics expensive simulations of the\n21cm signal across a wide parameter space to high precision. We embed our\nemulator within a Markov-Chain Monte Carlo framework, enabling it to explore\nthe posterior distribution over a large number of model parameters, including\nthose that govern the Epoch of Reionization, the Epoch of X-ray Heating, and\ncosmology. As a worked example, we use our emulator to present an updated\nparameter constraint forecast for the Hydrogen Epoch of Reionization Array\nexperiment, showing that its characterization of a fiducial 21cm power spectrum\nwill considerably narrow the allowed parameter space of reionization and\nheating parameters, and could help strengthen Planck's constraints on\n$\\sigma_8$. We provide both our generalized emulator code and its\nimplementation specifically for 21cm parameter constraints as publicly\navailable software.\n",
"title": "Emulating Simulations of Cosmic Dawn for 21cm Power Spectrum Constraints on Cosmology, Reionization, and X-ray Heating"
} | null | null | null | null | true | null | 19738 | null | Default | null | null |
null | {
"abstract": " Magnetic skyrmions are particle-like objects with topologically-protected\nstability which can be set into motion with an applied current. Using a\nparticle-based model we simulate current-driven magnetic skyrmions interacting\nwith random quenched disorder and examine the skyrmion velocity fluctuations\nparallel and perpendicular to the direction of motion as a function of\nincreasing drive. We show that the Magnus force contribution to skyrmion\ndynamics combined with the random pinning produces an isotropic effective\nshaking temperature. As a result, the skyrmions form a moving crystal at large\ndrives instead of the moving smectic state observed in systems with a\nnegligible Magnus force where the effective shaking temperature is anisotropic.\nWe demonstrate that spectral analysis of the velocity noise fluctuations can be\nused to identify dynamical phase transitions and to extract information about\nthe different dynamic phases, and show how the velocity noise fluctuations are\ncorrelated with changes in the skyrmion Hall angle, transport features, and\nskyrmion lattice structure.\n",
"title": "Fluctuations and Noise Signatures of Driven Magnetic Skyrmions"
} | null | null | null | null | true | null | 19739 | null | Default | null | null |
null | {
"abstract": " In this paper, we introduce metallic maps between metallic Riemannian\nmanifolds, provide an example and obtain certain conditions for such maps to be\ntotally geodesic. We also give a sufficient condition for a map between\nmetallic Riemannian manifolds to be harmonic map. Then we investigate the\nconstancy of certain maps between metallic Riemannian manifolds and various\nmanifolds by imposing the holomorphic-like condition. Moreover, we check the\nreverse case and show that some such maps are constant if there is a condition\nfor this.\n",
"title": "Metallic maps between metallic Riemannian manifolds and constancy of certain maps"
} | null | null | null | null | true | null | 19740 | null | Default | null | null |
null | {
"abstract": " Hashing, or learning binary embeddings of data, is frequently used in nearest\nneighbor retrieval. In this paper, we develop learning to rank formulations for\nhashing, aimed at directly optimizing ranking-based evaluation metrics such as\nAverage Precision (AP) and Normalized Discounted Cumulative Gain (NDCG). We\nfirst observe that the integer-valued Hamming distance often leads to tied\nrankings, and propose to use tie-aware versions of AP and NDCG to evaluate\nhashing for retrieval. Then, to optimize tie-aware ranking metrics, we derive\ntheir continuous relaxations, and perform gradient-based optimization with deep\nneural networks. Our results establish the new state-of-the-art for image\nretrieval by Hamming ranking in common benchmarks.\n",
"title": "Hashing as Tie-Aware Learning to Rank"
} | null | null | null | null | true | null | 19741 | null | Default | null | null |
null | {
"abstract": " A new exact solution of Einstein's field equations on the background of\nparaboloidal spacetime using Karmarkar condition is reported. The physical\nacceptability conditions of the model are investigated and found that the model\nis compatible with a number of compact star candidates like Her X-1, LMC X-4,\nEXO 1785-248, PSR J1903+327, Vela X-1 and PSR J1614-2230. A noteworthy feature\nof the model is that it is geometrically significant and simple in form.\n",
"title": "Models of compact stars on paraboloidal spacetime satisfying Karmarkar condition"
} | null | null | null | null | true | null | 19742 | null | Default | null | null |
null | {
"abstract": " Implicit discourse relation classification is of great challenge due to the\nlack of connectives as strong linguistic cues, which motivates the use of\nannotated implicit connectives to improve the recognition. We propose a feature\nimitation framework in which an implicit relation network is driven to learn\nfrom another neural network with access to connectives, and thus encouraged to\nextract similarly salient features for accurate classification. We develop an\nadversarial model to enable an adaptive imitation scheme through competition\nbetween the implicit network and a rival feature discriminator. Our method\neffectively transfers discriminability of connectives to the implicit features,\nand achieves state-of-the-art performance on the PDTB benchmark.\n",
"title": "Adversarial Connective-exploiting Networks for Implicit Discourse Relation Classification"
} | null | null | null | null | true | null | 19743 | null | Default | null | null |
null | {
"abstract": " The idea of posing a command following or tracking control problem as an\ninput reconstruction problem is explored in the paper. For a class of square\nMIMO systems with known dynamics, by pretending that reference commands are\nactual outputs of the system, input reconstruction methods can be used to\ndetermine control action that will result in a system following desired\nreference commands. A feedback controller which is a combination of an unbiased\nstate estimator and an input reconstructor that ensures unbiased tracking of\nreference commands is proposed. Simulations and real-time implementation are\npresented to demonstrate utility of the proposed idea. Conditions under which\nproposed controller may be used for non-square systems are also discussed.\n",
"title": "An Input Reconstruction Approach for Command Following in Linear MIMO Systems"
} | null | null | [
"Computer Science"
]
| null | true | null | 19744 | null | Validated | null | null |
null | {
"abstract": " Wireless communication plays a vital role in the promising performance of\nconnected and automated vehicle (CAV) technology. This paper proposes a\nVissim-based microscopic traffic simulation framework with an analytical\ndedicated short-range communication (DSRC) module for packet reception. Being\nderived from ns-2, a packet-level network simulator, the DSRC probability\nmodule takes into account the imperfect wireless communication that occurs in\nreal-world deployment. Four managed lane deployment strategies are evaluated\nusing the proposed framework. While the average packet reception rate is above\n93\\% among all tested scenarios, the results reveal that the reliability of the\nvehicle-to-vehicle (V2V) communication can be influenced by the deployment\nstrategies. Additionally, the proposed framework exhibits desirable scalability\nfor traffic simulation and it is able to evaluate transportation-network-level\ndeployment strategies in the near future for CAV technologies.\n",
"title": "Simulation Framework for Cooperative Adaptive Cruise Control with Empirical DSRC Module"
} | null | null | [
"Computer Science"
]
| null | true | null | 19745 | null | Validated | null | null |
null | {
"abstract": " This paper studies the synchronization of a finite number of Kuramoto\noscillators in a frequency-dependent bidirectional tree network. We assume that\nthe coupling strength of each link in each direction is equal to the product of\na common coefficient and the exogenous frequency of its corresponding head\noscillator. We derive a sufficient condition for the common coupling strength\nin order to guarantee frequency synchronization in tree networks. Moreover, we\ndiscuss the dependency of the obtained bound on both the graph structure and\nthe way that exogenous frequencies are distributed. Further, we present an\napplication of the obtained result by means of an event-triggered algorithm for\nachieving frequency synchronization in a star network assuming that the common\ncoupling coefficient is given.\n",
"title": "Synchronization of Kuramoto oscillators in a bidirectional frequency-dependent tree network"
} | null | null | null | null | true | null | 19746 | null | Default | null | null |
null | {
"abstract": " We derive representation theorems for exchangeable distributions on finite\nand infinite graphs using elementary arguments based on geometric and\ngraph-theoretic concepts. Our results elucidate some of the key differences,\nand their implications, between statistical network models that are finitely\nexchangeable and models that define a consistent sequence of probability\ndistributions on graphs of increasing size.\n",
"title": "On Exchangeability in Network Models"
} | null | null | null | null | true | null | 19747 | null | Default | null | null |
null | {
"abstract": " The replicator equation is one of the fundamental tools to study evolutionary\ndynamics in well-mixed populations. This paper contributes to the literature on\nevolutionary graph theory, providing a version of the replicator equation for a\nfamily of connected networks with communities, where nodes in the same\ncommunity have the same degree. This replicator equation is applied to the\nstudy of different classes of games, exploring the impact of the graph\nstructure on the equilibria of the evolutionary dynamics.\n",
"title": "Replicator equation on networks with degree regular communities"
} | null | null | null | null | true | null | 19748 | null | Default | null | null |
null | {
"abstract": " Go gaming is a struggle for territory control between rival, black and white,\nstones on a board. We model the Go dynamics in a game by means of the Ising\nmodel whose interaction coefficients reflect essential rules and tactics\nemployed in Go to build long-term strategies. At any step of the game, the\nenergy functional of the model provides the control degree (strength) of a\nplayer over the board. A close fit between predictions of the model with actual\ngames is obtained.\n",
"title": "Go game formal revealing by Ising model"
} | null | null | null | null | true | null | 19749 | null | Default | null | null |
null | {
"abstract": " Laser writing with ultrashort pulses provides a potential route for the\nmanufacture of three-dimensional wires, waveguides and defects within diamond.\nWe present a transmission electron microscopy (TEM) study of the intrinsic\nstructure of the laser modifications and reveal a complex distribution of\ndefects. Electron energy loss spectroscopy (EELS) indicates that the majority\nof the irradiated region remains as $sp^3$ bonded diamond.\nElectrically-conductive paths are attributed to the formation of multiple\nnano-scale, $sp^2$-bonded graphitic wires and a network of strain-relieving\nmicro-cracks.\n",
"title": "High resolution structural characterisation of laser-induced defect clusters inside diamond"
} | null | null | null | null | true | null | 19750 | null | Default | null | null |
null | {
"abstract": " Ensemble weather predictions require statistical post-processing of\nsystematic errors to obtain reliable and accurate probabilistic forecasts.\nTraditionally, this is accomplished with distributional regression models in\nwhich the parameters of a predictive distribution are estimated from a training\nperiod. We propose a flexible alternative based on neural networks that can\nincorporate nonlinear relationships between arbitrary predictor variables and\nforecast distribution parameters that are automatically learned in a\ndata-driven way rather than requiring pre-specified link functions. In a case\nstudy of 2-meter temperature forecasts at surface stations in Germany, the\nneural network approach significantly outperforms benchmark post-processing\nmethods while being computationally more affordable. Key components to this\nimprovement are the use of auxiliary predictor variables and station-specific\ninformation with the help of embeddings. Furthermore, the trained neural\nnetwork can be used to gain insight into the importance of meteorological\nvariables thereby challenging the notion of neural networks as uninterpretable\nblack boxes. Our approach can easily be extended to other statistical\npost-processing and forecasting problems. We anticipate that recent advances in\ndeep learning combined with the ever-increasing amounts of model and\nobservation data will transform the post-processing of numerical weather\nforecasts in the coming decade.\n",
"title": "Neural networks for post-processing ensemble weather forecasts"
} | null | null | null | null | true | null | 19751 | null | Default | null | null |
null | {
"abstract": " In recent work, Pomerance and Shparlinski have obtained results on the number\nof cycles in the functional graph of the map $x \\mapsto x^a$ in\n$\\mathbb{F}_p^*$. We prove similar results for other families of finite groups.\nIn particular, we obtain estimates for the number of cycles for cyclic groups,\nsymmetric groups, dihedral groups and $SL_2(\\mathbb{F}_q)$. We also show that\nthe cyclic group of order $n$ minimizes the number of cycles among all\nnilpotent groups of order $n$ for a fixed exponent. Finally, we pose several\nproblems.\n",
"title": "Power maps in finite groups"
} | null | null | null | null | true | null | 19752 | null | Default | null | null |
null | {
"abstract": " In this note we prove a conjecture by Li, Qu, Li, and Fu on permutation\ntrinomials over $\\mathbb{F}_3^{2k}$. In addition, new examples and\ngeneralizations of some families of permutation polynomials of\n$\\mathbb{F}_{3^k}$ and $\\mathbb{F}_{5^k}$ are given. We also study permutation\nquadrinomials of type $Ax^{q(q-1)+1} + Bx^{2(q-1)+1} + Cx^{q} + x$. Our method\nis based on the investigation of an algebraic curve associated with a\n{fractional polynomial} over a finite field.\n",
"title": "Permutation polynomials, fractional polynomials, and algebraic curves"
} | null | null | null | null | true | null | 19753 | null | Default | null | null |
null | {
"abstract": " Recently in speaker recognition, performance degradation due to the channel\ndomain mismatched condition has been actively addressed. However, the\nmismatches arising from language is yet to be sufficiently addressed. This\npaper proposes an approach which employs recursive whitening transformation to\nmitigate the language mismatched condition. The proposed method is based on the\nmultiple whitening transformation, which is intended to remove un-whitened\nresidual components in the dataset associated with i-vector length\nnormalization. The experiments were conducted on the Speaker Recognition\nEvaluation 2016 trials of which the task is non-English speaker recognition\nusing development dataset consist of both a large scale out-of-domain (English)\ndataset and an extremely low-quantity in-domain (non-English) dataset. For\nperformance comparison, we develop a state-of- the-art system using deep neural\nnetwork and bottleneck feature, which is based on a phonetically aware model.\nFrom the experimental results, along with other prior studies, effectiveness of\nthe proposed method on language mismatched condition is validated.\n",
"title": "Recursive Whitening Transformation for Speaker Recognition on Language Mismatched Condition"
} | null | null | null | null | true | null | 19754 | null | Default | null | null |
null | {
"abstract": " Employers actively look for talents having not only specific hard skills but\nalso various soft skills. To analyze the soft skill demands on the job market,\nit is important to be able to detect soft skill phrases from job advertisements\nautomatically. However, a naive matching of soft skill phrases can lead to\nfalse positive matches when a soft skill phrase, such as friendly, is used to\ndescribe a company, a team, or another entity, rather than a desired candidate.\nIn this paper, we propose a phrase-matching-based approach which\ndifferentiates between soft skill phrases referring to a candidate vs.\nsomething else. The disambiguation is formulated as a binary text\nclassification problem where the prediction is made for the potential soft\nskill based on the context where it occurs. To inform the model about the soft\nskill for which the prediction is made, we develop several approaches,\nincluding soft skill masking and soft skill tagging.\nWe compare several neural network based approaches, including CNN, LSTM and\nHierarchical Attention Model. The proposed tagging-based input representation\nusing LSTM achieved the highest recall of 83.92% on the job dataset when fixing\na precision to 95%.\n",
"title": "Learning Representations for Soft Skill Matching"
} | null | null | null | null | true | null | 19755 | null | Default | null | null |
null | {
"abstract": " We study a class of anomalies associated with time-reversal and spatial\nreflection symmetry in (2+1)D topological phases of matter. In these systems,\nthe topological quantum numbers of the quasiparticles, such as the fusion rules\nand braiding statistics, possess a $\\mathbb{Z}_2$ symmetry which can be\nassociated with either time-reversal (denoted $\\mathbb{Z}_2^{\\bf T})$ or\nspatial reflections. Under this symmetry, correlation functions of all Wilson\nloop operators in the low energy topological quantum field theory (TQFT) are\ninvariant. However, the theories that we study possess a severe anomaly\nassociated with the failure to consistently localize the symmetry action to the\nquasiparticles, precluding even defining a notion of symmetry\nfractionalization. We present simple sufficient conditions which determine when\n$\\mathbb{Z}_2^{\\bf T}$ symmetry localization anomalies exist. We present an\ninfinite series of TQFTs with such anomalies, some examples of which include\nUSp$(4)_2$ and SO$(4)_4$ Chern-Simons (CS) theory. The theories that we find\nwith these $\\mathbb{Z}_2^{\\bf T}$ anomalies can be obtained by gauging the\nunitary $\\mathbb{Z}_2$ subgroup of a different TQFT with a $\\mathbb{Z}_4^{\\bf\nT}$ symmetry. We show that the anomaly can be resolved in several ways: (1) the\ntrue symmetry of the theory is $\\mathbb{Z}_4^{\\bf T}$, or (2) the theory can be\nconsidered to be a theory of fermions, with ${\\bf T}^2 = (-1)^{N_f}$\ncorresponding to fermion parity. Finally, we demonstrate that theories with the\n$\\mathbb{Z}_2^{\\bf T}$ localization anomaly can be compatible with\n$\\mathbb{Z}_2^{\\bf T}$ if they are \"pseudo-realized\" at the surface of a (3+1)D\nsymmetry-enriched topological phase. The \"pseudo-realization\" refers to the\nfact that the bulk (3+1)D system is described by a dynamical $\\mathbb{Z}_2$\ngauge theory and thus only a subset of the quasiparticles are confined to the\nsurface.\n",
"title": "Time-reversal and spatial reflection symmetry localization anomalies in (2+1)D topological phases of matter"
} | null | null | null | null | true | null | 19756 | null | Default | null | null |
null | {
"abstract": " We present a new model for the redshift-space power spectrum of galaxies and\ndemonstrate its accuracy in modeling the monopole, quadrupole, and hexadecapole\nof the galaxy density field down to scales of $k = 0.4 \\ h\\mathrm{Mpc}^{-1}$.\nThe model describes the clustering of galaxies in the context of a halo model\nand the clustering of the underlying halos in redshift space using a\ncombination of Eulerian perturbation theory and $N$-body simulations. The\nmodeling of redshift-space distortions is done using the so-called distribution\nfunction approach. The final model has 13 free parameters, and each parameter\nis physically motivated rather than a nuisance parameter, which allows the use\nof well-motivated priors. We account for the Finger-of-God effect from centrals\nand both isolated and non-isolated satellites rather than using a single\nvelocity dispersion to describe the combined effect. We test and validate the\naccuracy of the model on several sets of high-fidelity $N$-body simulations, as\nwell as realistic mock catalogs designed to simulate the BOSS DR12 CMASS data\nset. The suite of simulations covers a range of cosmologies and galaxy bias\nmodels, providing a rigorous test of the level of theoretical systematics\npresent in the model. The level of bias in the recovered values of $f \\sigma_8$\nis found to be small. When including scales to $k = 0.4 \\ h\\mathrm{Mpc}^{-1}$,\nwe find 15-30\\% gains in the statistical precision of $f \\sigma_8$ relative to\n$k = 0.2 \\ h\\mathrm{Mpc}^{-1}$ and a roughly 10-15\\% improvement for the\nperpendicular Alcock-Paczynski parameter $\\alpha_\\perp$. Using the BOSS DR12\nCMASS mocks as a benchmark for comparison, we estimate an uncertainty on $f\n\\sigma_8$ that is $\\sim$10-20\\% larger than other similar Fourier-space RSD\nmodels in the literature that use $k \\leq 0.2 \\ h\\mathrm{Mpc}^{-1}$, suggesting\nthat these models likely have a too-limited parametrization.\n",
"title": "Extending the modeling of the anisotropic galaxy power spectrum to $k = 0.4 \\ h\\mathrm{Mpc}^{-1}$"
} | null | null | null | null | true | null | 19757 | null | Default | null | null |
null | {
"abstract": " It is proven that an infinite finitely generated group cannot be elementarily\nequivalent to an ultraproduct of finite groups of a given Prüfer rank.\nFurthermore, it is shown that an infinite finitely generated group of finite\nPrüfer rank is not pseudofinite.\n",
"title": "Finite rank and pseudofinite groups"
} | null | null | null | null | true | null | 19758 | null | Default | null | null |
null | {
"abstract": " The spin Peltier effect (SPE), heat-current generation due to spin-current\ninjection, in various metal (Pt, W, and Au single layers and Pt/Cu\nbilayer)/ferrimagnetic insulator (yttrium iron garnet: YIG) junction systems\nhas been investigated by means of a lock-in thermography (LIT) method. The SPE\nis excited by a spin current across the metal/YIG interface, which is generated\nby applying a charge current to the metallic layer via the spin Hall effect.\nThe LIT method enables the thermal imaging of the SPE free from the\nJoule-heating contribution. Importantly, we observed spin-current-induced\ntemperature modulation not only in the Pt/YIG and W/YIG systems but also in the\nAu/YIG and Pt/Cu/YIG systems, excluding the possible contamination by anomalous\nEttingshausen effects due to proximity-induced ferromagnetism near the\nmetal/YIG interface. As demonstrated in our previous study, the SPE signals are\nconfined only in the vicinity of the metal/YIG interface; we buttress this\nconclusion by reducing a spatial blur due to thermal diffusion in an infrared\nemission layer on the sample surface used for the LIT measurements. We also\nfound that the YIG-thickness dependence of the SPE is similar to that of the\nspin Seebeck effect measured in the same Pt/YIG sample, implying the reciprocal\nrelation between them.\n",
"title": "Thermographic measurements of the spin Peltier effect in metal/yttrium-iron-garnet junction systems"
} | null | null | null | null | true | null | 19759 | null | Default | null | null |
null | {
"abstract": " We discuss the number-theoretic properties of distributions appearing in\nphysical systems when an observable is a quotient of two independent\nexponentially weighted integers. The spectral density of ensemble of linear\npolymer chains distributed with the law $\\sim f^L$ ($0<f<1$), where $L$ is the\nchain length, serves as a particular example. At $f\\to 1$, the spectral density\ncan be expressed through the discontinuous at all rational points, Thomae\n(\"popcorn\") function. We suggest a continuous approximation of the popcorn\nfunction, based on the Dedekind $\\eta$-function near the real axis. Moreover,\nwe provide simple arguments, based on the \"Euclid orchard\" construction, that\ndemonstrate the presence of Lifshitz tails, typical for the 1D Anderson\nlocalization, at the spectral edges. We emphasize that the ultrametric\nstructure of the spectral density is ultimately connected with number-theoretic\nrelations on asymptotic modular functions. We also pay attention to connection\nof the Dedekind $\\eta$-function near the real axis to invariant measures of\nsome continued fractions studied by Borwein and Borwein in 1993.\n",
"title": "Number-theoretic aspects of 1D localization: \"popcorn function\" with Lifshitz tails and its continuous approximation by the Dedekind $η$"
} | null | null | null | null | true | null | 19760 | null | Default | null | null |
null | {
"abstract": " In this paper, we describe a phenomenon, which we named \"super-convergence\",\nwhere neural networks can be trained an order of magnitude faster than with\nstandard training methods. The existence of super-convergence is relevant to\nunderstanding why deep networks generalize well. One of the key elements of\nsuper-convergence is training with one learning rate cycle and a large maximum\nlearning rate. A primary insight that allows super-convergence training is that\nlarge learning rates regularize the training, hence requiring a reduction of\nall other forms of regularization in order to preserve an optimal\nregularization balance. We also derive a simplification of the Hessian Free\noptimization method to compute an estimate of the optimal learning rate.\nExperiments demonstrate super-convergence for Cifar-10/100, MNIST and Imagenet\ndatasets, and resnet, wide-resnet, densenet, and inception architectures. In\naddition, we show that super-convergence provides a greater boost in\nperformance relative to standard training when the amount of labeled training\ndata is limited. The architectures and code to replicate the figures in this\npaper are available at github.com/lnsmith54/super-convergence. See\nthis http URL for an application of\nsuper-convergence to win the DAWNBench challenge (see\nthis https URL).\n",
"title": "Super-Convergence: Very Fast Training of Neural Networks Using Large Learning Rates"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 19761 | null | Validated | null | null |
null | {
"abstract": " In this paper we perform the analysis that leads to the space of initial\nconditions for the Hamiltonian system $q' = p^2 + zq + \\alpha$, $p' = -q^2 - zp\n- \\beta$, studied by the author in an earlier article. By compactifying the\nphase space of the system from $\\mathbb{C}^2$ to $\\mathbb{CP}^2$ three base\npoints arise in the standard coordinate charts covering the complex projective\nspace. Each of these is removed by a sequence of three blow-ups, a construction\nto regularise the system at these points. The resulting space, where the\nexceptional curves introduced after the first and second blow-up are removed,\nis the so-called Okamoto's space of initial conditions for this system which,\nat every point, defines a regular initial value problem in some coordinate\nchart of the space.\n",
"title": "Space of initial conditions for a cubic Hamiltonian system"
} | null | null | null | null | true | null | 19762 | null | Default | null | null |
null | {
"abstract": " This work reports an electronic and micro-structural study of an appealing\nsystem for optoelectronics: tungsten disulphide WS$_2$ on epitaxial graphene\n(EG) on SiC(0001). The WS$_2$ is grown via chemical vapor deposition (CVD) onto\nthe EG. Low-energy electron diffraction (LEED) measurements assign the\nzero-degree orientation as the preferential azimuthal alignment for WS$_2$/EG.\nThe valence-band (VB) structure emerging from this alignment is investigated by\nmeans of photoelectron spectroscopy measurements, with both high space and\nenergy resolution. We find that the spin-orbit splitting of monolayer WS$_2$ on\ngraphene is of 462 meV, larger than what is reported to date for other\nsubstrates. We determine the value of the work function for the WS$_2$/EG to be\n4.5$\\pm$0.1 eV. A large shift of the WS$_2$ VB maximum is observed as well ,\ndue to the lowering of the WS$_2$ work function caused by the donor-like\ninterfacial states of EG. Density functional theory (DFT) calculations carried\nout on a coincidence supercell confirm the experimental band structure to an\nexcellent degree. X-ray photoemission electron microscopy (XPEEM) measurements\nperformed on single WS$_2$ crystals confirm the van der Waals nature of the\ninterface coupling between the two layers. In virtue of its band alignment and\nlarge spin-orbit splitting, this system gains strong appeal for optical\nspin-injection experiments and opto-spintronic applications in general.\n",
"title": "Electronic properties of WS$_2$ on epitaxial graphene on SiC(0001)"
} | null | null | null | null | true | null | 19763 | null | Default | null | null |
null | {
"abstract": " We fabricated NiFe$_\\textrm{2}$O$_\\textrm{x}$ thin films on\nMgAl$_2$O$_4$(001) substrates by reactive dc magnetron co-sputtering varying\nthe oxygen partial pressure during deposition. The fabrication of a variable\nmaterial with oxygen deficiency leads to controllable electrical and optical\nproperties which would be beneficial for the investigations of the transport\nphenomena and would, therefore, promote the use of such materials in spintronic\nand spin caloritronic applications. We used several characterization techniques\nin order to investigate the film properties, focusing on their structural,\nmagnetic, electrical, and optical properties. From the electrical resistivity\nmeasurements we obtained the conduction mechanisms that govern the systems in\nhigh and low temperature regimes, extracting low thermal activation energies\nwhich unveil extrinsic transport mechanisms. The thermal activation energy\ndecreases in the less oxidized samples revealing the pronounced contribution of\na large amount of electronic states localized in the band gap to the electrical\nconductivity. Hall effect measurements showed the mixed-type semiconducting\ncharacter of our films. The optical band gaps were determined via\nultraviolet-visible spectroscopy. They follow a similar trend as the thermal\nactivation energy, with lower band gap values in the less oxidized samples.\n",
"title": "Electrical transport and optical band gap of NiFe$_\\textrm{2}$O$_\\textrm{x}$ thin films"
} | null | null | null | null | true | null | 19764 | null | Default | null | null |
null | {
"abstract": " This paper proposes a new sampling-based nonlinear model predictive control\n(MPC) algorithm, with a bound on complexity quadratic in the prediction horizon\nN and linear in the number of samples. The idea of the proposed algorithm is to\nuse the sequence of predicted inputs from the previous time step as a warm\nstart, and to iteratively update this sequence by changing its elements one by\none, starting from the last predicted input and ending with the first predicted\ninput. This strategy, which resembles the dynamic programming principle, allows\nfor parallelization up to a certain level and yields a suboptimal nonlinear MPC\nalgorithm with guaranteed recursive feasibility, stability and improved cost\nfunction at every iteration, which is suitable for real-time implementation.\nThe complexity of the algorithm per each time step in the prediction horizon\ndepends only on the horizon, the number of samples and parallel threads, and it\nis independent of the measured system state. Comparisons with the fmincon\nnonlinear optimization solver on benchmark examples indicate that as the\nsimulation time progresses, the proposed algorithm converges rapidly to the\n\"optimal\" solution, even when using a small number of samples.\n",
"title": "Towards parallelizable sampling-based Nonlinear Model Predictive Control"
} | null | null | null | null | true | null | 19765 | null | Default | null | null |
null | {
"abstract": " We study the primary entanglement effect on the decoherence of fields reduced\ndensity matrix which are in interaction with another fields or independent mode\nfunctions. We show that the primary entanglement has a significant role in\ndecoherence of the system quantum state. We find that the existence of\nentanglement could couple dynamical equations coming from Schrödinger\nequation. We show if one wants to see no effect of the entanglement parameter\nin decoherence then interaction terms in Hamiltonian can not be independent\nfrom each other. Generally, including the primary entanglement destroys the\nindependence of the interaction terms. Our results could be generalized to\nevery scalar quantum field theory with a well defined quantization in a given\ncurved space time.\n",
"title": "Quantum decoherence from entanglement during inflation"
} | null | null | [
"Physics"
]
| null | true | null | 19766 | null | Validated | null | null |
null | {
"abstract": " Breast density classification is an essential part of breast cancer\nscreening. Although a lot of prior work considered this problem as a task for\nlearning algorithms, to our knowledge, all of them used small and not\nclinically realistic data both for training and evaluation of their models. In\nthis work, we explore the limits of this task with a data set coming from over\n200,000 breast cancer screening exams. We use this data to train and evaluate a\nstrong convolutional neural network classifier. In a reader study, we find that\nour model can perform this task comparably to a human expert.\n",
"title": "Breast density classification with deep convolutional neural networks"
} | null | null | null | null | true | null | 19767 | null | Default | null | null |
null | {
"abstract": " Motivated by the recent progress in analog computing [Science 343, 160\n(2014)], a new approach to perform spatial integration is presented using a\ndielectric slab waveguide. Our approach is indeed based on the fact that the\ntransmission coefficient of a simple dielectric slab waveguide at its mode\nexcitation angle matches the Green's function of first order integration.\nInspired by the mentioned dielectric-based integrator, we further demonstrate\nits graphene-based counterpart. The latter is not only reconfigurable but also\nhighly miniaturized in contrast to the previously reported designs [Opt.\nCommun. 338, 457 (2015)]. Such integrators have the potential to be used in\nultrafast analog computation and signal processing.\n",
"title": "Spatial Integration by a Dielectric Slab Waveguide and its Planar Graphene-based Counterpart"
} | null | null | null | null | true | null | 19768 | null | Default | null | null |
null | {
"abstract": " A crucial challenge in image-based modeling of biomedical data is to identify\ntrends and features that separate normality and pathology. In many cases, the\nmorphology of the imaged object exhibits continuous change as it deviates from\nnormality, and thus a generative model can be trained to model this\nmorphological continuum. Moreover, given side information that correlates to\ncertain trend in morphological change, a latent variable model can be\nregularized such that its latent representation reflects this side information.\nIn this work, we use the Wasserstein Auto-encoder to model this pathology\ncontinuum, and apply the Hilbert-Schmitt Independence Criterion (HSIC) to\nenforce dependency between certain latent features and the provided side\ninformation. We experimentally show that the model can provide disentangled and\ninterpretable latent representations and also generate a continuum of\nmorphological changes that corresponds to change in the side information.\n",
"title": "Modeling the Biological Pathology Continuum with HSIC-regularized Wasserstein Auto-encoders"
} | null | null | null | null | true | null | 19769 | null | Default | null | null |
null | {
"abstract": " We study an optimal control problem arising from a generalization of\nrock-paper-scissors in which the number of strategies may be selected from any\npositive odd number greater than 1 and in which the payoff to the winner is\ncontrolled by a control variable $\\gamma$. Using the replicator dynamics as the\nequations of motion, we show that a quasi-linearization of the problem admits a\nspecial optimal control form in which explicit dynamics for the controller can\nbe identified. We show that all optimal controls must satisfy a specific second\norder differential equation parameterized by the number of strategies in the\ngame. We show that as the number of strategies increases, a limiting case\nadmits a closed form for the open-loop optimal control. In performing our\nanalysis we show necessary conditions on an optimal control problem that allow\nthis analytic approach to function.\n",
"title": "Control Problems with Vanishing Lie Bracket Arising from Complete Odd Circulant Evolutionary Games"
} | null | null | [
"Computer Science"
]
| null | true | null | 19770 | null | Validated | null | null |
null | {
"abstract": " This paper describes our experience of training a team of developers of an\nEast-European phone service provider. The training experience was structured in\ntwo sessions of two days each conducted in different weeks with a gap of about\nfifteen days. The first session was dedicated to the Continuous Integration\nDelivery Pipeline, and the second on Agile methods. We summarize the activity,\nits preparation and delivery and draw some conclusions out of it on our\nmistakes and how future session should be addressed.\n",
"title": "Teaching DevOps in Corporate Environments: An experience report"
} | null | null | null | null | true | null | 19771 | null | Default | null | null |
null | {
"abstract": " The purpose of this article is to analyze the connection between\nEynard-Orantin topological recursion and formal WKB solutions of a\n$\\hbar$-difference equation: $\\Psi(x+\\hbar)=\\left(e^{\\hbar\\frac{d}{dx}}\\right)\n\\Psi(x)=L(x;\\hbar)\\Psi(x)$ with $L(x;\\hbar)\\in GL_2( (\\mathbb{C}(x))[\\hbar])$.\nIn particular, we extend the notion of determinantal formulas and topological\ntype property proposed for formal WKB solutions of $\\hbar$-differential systems\nto this setting. We apply our results to a specific $\\hbar$-difference system\nassociated to the quantum curve of the Gromov-Witten invariants of\n$\\mathbb{P}^1$ for which we are able to prove that the correlation functions\nare reconstructed from the Eynard-Orantin differentials computed from the\ntopological recursion applied to the spectral curve $y=\\cosh^{-1}\\frac{x}{2}$.\nFinally, identifying the large $x$ expansion of the correlation functions,\nproves a recent conjecture made by B. Dubrovin and D. Yang regarding a new\ngenerating series for Gromov-Witten invariants of $\\mathbb{P}^1$.\n",
"title": "WKB solutions of difference equations and reconstruction by the topological recursion"
} | null | null | null | null | true | null | 19772 | null | Default | null | null |
null | {
"abstract": " Covalently linked acene dimers are of interest as candidates for\nintramolecular singlet fission. We report many-electron calculations of the\nenergies and wavefunctions of the optical singlets, the lowest triplet exciton\nand the triplet-triplet biexciton, as well as the final states of excited state\nabsorptions from these states in a family of phenyl-linked pentacene dimers.\nWhile it is difficult to distinguish between the triplet and the\ntriplet-triplet from their transient absorptions in the 500-600 nm region, by\ncomparing theoretical transient absorption spectra against published and\nunpublished experimental transient absorptions in the near and mid infrared we\nconclude that the end product of photoexcitation in these particular\nbipentacenes is the bound triplet-triplet and not free triplets. We predict\nadditional transient absorptions at even longer wavelengths, beyond 1500 nm, to\nthe equivalent of the classic 2$^1$A$_g^-$ in linear polyenes.\n",
"title": "Diagrammatic Exciton Basis Theory of the Photophysics of Pentacene Dimers"
} | null | null | [
"Physics"
]
| null | true | null | 19773 | null | Validated | null | null |
null | {
"abstract": " We discuss an investigation of student difficulties with degenerate\nperturbation theory (DPT) carried out in advanced quantum mechanics courses by\nadministering free-response and multiple-choice questions and conducting\nindividual interviews with students. We find that students share many common\ndifficulties related to this topic. We used the difficulties found via research\nas resources to develop and evaluate a Quantum Interactive Learning Tutorial\n(QuILT) which strives to help students develop a functional understanding of\nDPT. We discuss the development of the DPT QuILT and its preliminary evaluation\nin the undergraduate and graduate courses.\n",
"title": "Developing and evaluating an interactive tutorial on degenerate perturbation theory"
} | null | null | [
"Physics"
]
| null | true | null | 19774 | null | Validated | null | null |
null | {
"abstract": " In this paper, we first design a time optimal control problem for the heat\nequation with sampled-data controls, and then use it to approximate a time\noptimal control problem for the heat equation with distributed controls. Our\ndesign is reasonable from perspective of sampled-data controls. And it might\nprovide a right way for the numerical approach of a time optimal distributed\ncontrol problem, via the corresponding semi-discretized (in time variable) time\noptimal control problem.\nThe study of such a time optimal sampled-data control problem is not easy,\nbecause it may have infinitely many optimal controls. We find connections among\nthis problem, a minimal norm sampled-data control problem and a minimization\nproblem. And obtain some properties on these problems. Based on these, we not\nonly build up error estimates for optimal time and optimal controls between the\ntime optimal sampled-data control problem and the time optimal distributed\ncontrol problem, in terms of the sampling period, but also prove that such\nestimates are optimal in some sense.\n",
"title": "Time optimal sampled-data controls for heat equations"
} | null | null | null | null | true | null | 19775 | null | Default | null | null |
null | {
"abstract": " Knowledge about the graph structure of the Web is important for understanding\nthis complex socio-technical system and for devising proper policies supporting\nits future development. Knowledge about the differences between clean and\nmalicious parts of the Web is important for understanding potential treats to\nits users and for devising protection mechanisms. In this study, we conduct\ndata science methods on a large crawl of surface and deep Web pages with the\naim to increase such knowledge. To accomplish this, we answer the following\nquestions. Which theoretical distributions explain important local\ncharacteristics and network properties of websites? How are these\ncharacteristics and properties different between clean and malicious\n(malware-affected) websites? What is the prediction power of local\ncharacteristics and network properties to classify malware websites? To the\nbest of our knowledge, this is the first large-scale study describing the\ndifferences in global properties between malicious and clean parts of the Web.\nIn other words, our work is building on and bridging the gap between\n\\textit{Web science} that tackles large-scale graph representations and\n\\textit{Web cyber security} that is concerned with malicious activities on the\nWeb. The results presented herein can also help antivirus vendors in devising\napproaches to improve their detection algorithms.\n",
"title": "Malware distributions and graph structure of the Web"
} | null | null | [
"Computer Science"
]
| null | true | null | 19776 | null | Validated | null | null |
null | {
"abstract": " We show that for any convex differentiable loss function, a deep linear\nnetwork has no spurious local minima as long as it is true for the two layer\ncase. When applied to the quadratic loss, our result immediately implies the\npowerful result in [Kawaguchi 2016] that there is no spurious local minima in\ndeep linear networks. Further, with the recent work [Zhou and Liang 2018], we\ncan remove all the assumptions in [Kawaguchi 2016]. Our proof is short and\nelementary. It builds on the recent work of [Laurent and von Brecht 2018] and\nuses a new rank one perturbation argument.\n",
"title": "Depth creates no more spurious local minima"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 19777 | null | Validated | null | null |
null | {
"abstract": " Research on automated vehicles has experienced an explosive growth over the\npast decade. A main obstacle to their practical realization, however, is a\nconvincing safety concept. This question becomes ever more important as more\nsophisticated algorithms are used and the vehicle automation level increases.\nThe field of functional safety offers a systematic approach to identify\npossible sources of risk and to improve the safety of a vehicle. It is based on\npractical experience across the aerospace, process and other industries over\nmultiple decades. This experience is compiled in the functional safety standard\nfor the automotive domain, ISO 26262, which is widely adopted throughout the\nautomotive industry. However, its applicability and relevance for highly\nautomated vehicles is subject to a controversial debate. This paper takes a\ncritical look at the discussion and summarizes the main steps of ISO 26262 for\na safe control design for automated vehicles.\n",
"title": "On the Application of ISO 26262 in Control Design for Automated Vehicles"
} | null | null | null | null | true | null | 19778 | null | Default | null | null |
null | {
"abstract": " This paper deals with a nonhomogeneous scalar parabolic equation with\npossibly degenerate diffusion term; the process has only one stationary state.\nThe equation can be interpreted as modeling collective movements (crowd\ndynamics, for instance). We first prove the existence of semi-wavefront\nsolutions for every wave speed; their properties are investigated. Then, a\nfamily of travelling wave solutions is constructed by a suitable combination of\nthe previous semi-wavefront solutions. Proofs exploit comparison-type\ntechniques and are carried out in the case of one spatial variable; the\nextension to the general case is straightforward.\n",
"title": "Semi-wavefront solutions in models of collective movements with density-dependent diffusivity"
} | null | null | null | null | true | null | 19779 | null | Default | null | null |
null | {
"abstract": " Phase changing materials (PCM) are widely used for optical data recording,\nsensing, all-optical switching, and optical limiting. Our focus here is on the\ncase when the change in the transmission characteristics of the optical\nmaterial is caused by the input light itself. Specifically, the light-induced\nheating triggers the phase transition in the PCM. In this paper, using a\nnumerical example, we demonstrate that incorporating the PCM in a photonic\nstructure can lead to a dramatic modification of the effects of light-induced\nphase transition, as compared to a stand-alone sample of the same PCM. Our\nfocus is on short pulses. We discuss some possible applications of such\nphase-changing photonic structures for optical sensing and limiting.\n",
"title": "Self-Regulated Transport in Photonic Crystals with Phase-Changing Defects"
} | null | null | null | null | true | null | 19780 | null | Default | null | null |
null | {
"abstract": " Sortition, i.e., random appointment for public duty, has been employed by\nsocieties throughout the years, especially for duties related to the judicial\nsystem, as a firewall designated to prevent illegitimate interference between\nparties in a legal case and agents of the legal system. In judicial systems of\nmodern western countries, random procedures are mainly employed to select the\njury, the court and/or the judge in charge of judging a legal case, so that\nthey have a significant role in the course of a case. Therefore, these random\nprocedures must comply with some principles, as statistical soundness; complete\nauditability; open-source programming; and procedural, cryptographical and\ncomputational security. Nevertheless, some of these principles are neglected by\nsome random procedures in judicial systems, that are, in some cases, performed\nin secrecy and are not auditable by the involved parts. The assignment of cases\nin the Brazilian Supreme Court (Supremo Tribunal Federal) is an example of such\nprocedures, for it is performed by a closed-source algorithm, unknown to the\npublic and to the parts involved in the judicial cases, that allegedly assign\nthe cases randomly to the justice chairs based on their caseload.\nIn this context, this article presents a review of how sortition has been\nemployed historically by societies, and discusses how Mathematical Statistics\nmay be applied to random procedures of the judicial system, as it has been\napplied for almost a century on clinical trials, for example. Based on this\ndiscussion, a statistical model for assessing randomness in case assignment is\nproposed and applied to the Brazilian Supreme Court in order to shed light on\nhow this assignment process is performed by the closed-source algorithm.\nGuidelines for random procedures are outlined and topics for further researches\npresented.\n",
"title": "Assessing randomness in case assignment: the case study of the Brazilian Supreme Court"
} | null | null | [
"Statistics"
]
| null | true | null | 19781 | null | Validated | null | null |
null | {
"abstract": " We provide a particle picture representation for the non-symmetric Rosenblatt\nprocess and for Hermite processes of any order, extending the result of\nBojdecki, Gorostiza and Talarczyk in~\\cite{FILT}. We show that these processes\ncan be obtained as limits in the sense of finite-dimensional distributions of\ncertain functionals of a system of particles evolving according to symmetric\nstable Lévy motions. In the case of $k$-Hermite processes the corresponding\nfunctional involves $k$-intersection local time of symmetric stable Lévy\nprocesses\n",
"title": "Particle picture representation of the non-symmetric Rosenblatt process and Hermite processes of any order"
} | null | null | [
"Mathematics"
]
| null | true | null | 19782 | null | Validated | null | null |
null | {
"abstract": " Generating adversarial examples is a critical step for evaluating and\nimproving the robustness of learning machines. So far, most existing methods\nonly work for classification and are not designed to alter the true performance\nmeasure of the problem at hand. We introduce a novel flexible approach named\nHoudini for generating adversarial examples specifically tailored for the final\nperformance measure of the task considered, be it combinatorial and\nnon-decomposable. We successfully apply Houdini to a range of applications such\nas speech recognition, pose estimation and semantic segmentation. In all cases,\nthe attacks based on Houdini achieve higher success rate than those based on\nthe traditional surrogates used to train the models while using a less\nperceptible adversarial perturbation.\n",
"title": "Houdini: Fooling Deep Structured Prediction Models"
} | null | null | null | null | true | null | 19783 | null | Default | null | null |
null | {
"abstract": " We consider the problem of designing efficient regularization algorithms when\nregularization is encoded by a (strongly) convex functional. Unlike classical\npenalization methods based on a relaxation approach, we propose an iterative\nmethod where regularization is achieved via early stopping. Our results show\nthat the proposed procedure achieves the same recovery accuracy as penalization\nmethods, while naturally integrating computational considerations. An empirical\nanalysis on a number of problems provides promising results with respect to the\nstate of the art.\n",
"title": "Don't relax: early stopping for convex regularization"
} | null | null | null | null | true | null | 19784 | null | Default | null | null |
null | {
"abstract": " We present a model that takes into account the coupling between evolutionary\ngame dynamics and social influence. Importantly, social influence and game\ndynamics take place in different domains, which we model as different layers of\na multiplex network. We show that the coupling between these dynamical\nprocesses can lead to cooperation in scenarios where the pure game dynamics\npredicts defection. In addition, we show that the structure of the network\nlayers and the relation between them can further increase cooperation.\nRemarkably, if the layers are related in a certain way, the system can reach a\npolarized metastable state.These findings could explain the prevalence of\npolarization observed in many social dilemmas.\n",
"title": "Interplay between social influence and competitive strategical games in multiplex networks"
} | null | null | null | null | true | null | 19785 | null | Default | null | null |
null | {
"abstract": " The quality of a Neural Machine Translation system depends substantially on\nthe availability of sizable parallel corpora. For low-resource language pairs\nthis is not the case, resulting in poor translation quality. Inspired by work\nin computer vision, we propose a novel data augmentation approach that targets\nlow-frequency words by generating new sentence pairs containing rare words in\nnew, synthetically created contexts. Experimental results on simulated\nlow-resource settings show that our method improves translation quality by up\nto 2.9 BLEU points over the baseline and up to 3.2 BLEU over back-translation.\n",
"title": "Data Augmentation for Low-Resource Neural Machine Translation"
} | null | null | null | null | true | null | 19786 | null | Default | null | null |
null | {
"abstract": " Close-contact melting refers to the process of a heat source melting its way\ninto a phase-change material. Of special interest is the close-contact melting\nvelocity, or more specifically the relative velocity between the heat source\nand the phase-change material. In this work, we present a novel numerical\napproach to simulate quasi-steady, heat flux driven close-contact melting. It\nextends existing approaches found in the literature, and, for the first time,\nallows to study the impact of a spatially varying heat flux distribution. We\nwill start by deriving the governing equations in a Lagrangian reference frame\nfixed to the heat source. Exploiting the narrowness of the melt film enables us\nto reduce the momentum balance to the Reynolds equation, which is coupled to\nthe energy balance via the velocity field. We particularize our derivation for\ntwo simple, yet technically relevant geometries, namely a 3d circular disc and\na 2d planar heat source. An iterative solution procedure for the coupled system\nis described in detail and discussed on the basis of a convergence study.\nFurthermore, we present an extension to allow for rotational melting modes.\nVarious test cases demonstrate the proficiency of our method. In particular, we\nwill utilize the method to assess the efficiency of the close-contact melting\nprocess and to quantify the model error introduced if convective losses are\nneglected. Finally, we will draw conclusions and present an outlook to future\nwork.\n",
"title": "A Lagrangian approach to modeling heat flux driven close-contact melting"
} | null | null | [
"Physics"
]
| null | true | null | 19787 | null | Validated | null | null |
null | {
"abstract": " Complex networks or graphs are ubiquitous in sciences and engineering:\nbiological networks, brain networks, transportation networks, social networks,\nand the World Wide Web, to name a few. Spectral graph theory provides a set of\nuseful techniques and models for understanding `patterns of interconnectedness'\nin a graph. Our prime focus in this paper is on the following question: Is\nthere a unified explanation and description of the fundamental spectral graph\nmethods? There are at least two reasons to be interested in this question.\nFirstly, to gain a much deeper and refined understanding of the basic\nfoundational principles, and secondly, to derive rich consequences with\npractical significance for algorithm design. However, despite half a century of\nresearch, this question remains one of the most formidable open issues, if not\nthe core problem in modern network science. The achievement of this paper is to\ntake a step towards answering this question by discovering a simple, yet\nuniversal statistical logic of spectral graph analysis. The prescribed\nviewpoint appears to be good enough to accommodate almost all existing spectral\ngraph techniques as a consequence of just one single formalism and algorithm.\n",
"title": "Spectral Graph Analysis: A Unified Explanation and Modern Perspectives"
} | null | null | null | null | true | null | 19788 | null | Default | null | null |
null | {
"abstract": " Accurate demand forecasts can help on-line retail organizations better plan\ntheir supply-chain processes. The challenge, however, is the large number of\nassociative factors that result in large, non-stationary shifts in demand,\nwhich traditional time series and regression approaches fail to model. In this\npaper, we propose a Neural Network architecture called AR-MDN, that\nsimultaneously models associative factors, time-series trends and the variance\nin the demand. We first identify several causal features and use a combination\nof feature embeddings, MLP and LSTM to represent them. We then model the output\ndensity as a learned mixture of Gaussian distributions. The AR-MDN can be\ntrained end-to-end without the need for additional supervision. We experiment\non a dataset of an year's worth of data over tens-of-thousands of products from\nFlipkart. The proposed architecture yields a significant improvement in\nforecasting accuracy when compared with existing alternatives.\n",
"title": "ARMDN: Associative and Recurrent Mixture Density Networks for eRetail Demand Forecasting"
} | null | null | null | null | true | null | 19789 | null | Default | null | null |
null | {
"abstract": " In the covariate shift learning scenario, the training and test covariate\ndistributions differ, so that a predictor's average loss over the training and\ntest distributions also differ. In this work, we explore the potential of\nextreme dimension reduction, i.e. to very low dimensions, in improving the\nperformance of importance weighting methods for handling covariate shift, which\nfail in high dimensions due to potentially high train/test covariate divergence\nand the inability to accurately estimate the requisite density ratios. We first\nformulate and solve a problem optimizing over linear subspaces a combination of\ntheir predictive utility and train/test divergence within. Applying it to\nsimulated and real data, we show extreme dimension reduction helps sometimes\nbut not always, due to a bias introduced by dimension reduction.\n",
"title": "Extreme Dimension Reduction for Handling Covariate Shift"
} | null | null | null | null | true | null | 19790 | null | Default | null | null |
null | {
"abstract": " One-dimensional systems obtained as low-energy limits of hybrid\nsuperconductor-topological insulator devices provide means of production,\ntransport, and destruction of Majorana bound states (MBSs) by variations of the\nmagnetic flux. When two or more pairs of MBSs are present in the intermediate\nstate, there is a possibility of a Landau-Zener transition, wherein even a slow\nvariation of the flux leads to production of a quasiparticle pair. We study\nnumerically a version of this process, with four MBSs produced and subsequently\ndestroyed, and find that, quite universally, the probability of quasiparticle\nproduction in it is 50%. This implies that the effect may be a limiting factor\nin applications requiring a high degree of quantum coherence.\n",
"title": "Landau-Zener transitions for Majorana fermions"
} | null | null | null | null | true | null | 19791 | null | Default | null | null |
null | {
"abstract": " Convolution as inner product has been the founding basis of convolutional\nneural networks (CNNs) and the key to end-to-end visual representation\nlearning. Benefiting from deeper architectures, recent CNNs have demonstrated\nincreasingly strong representation abilities. Despite such improvement, the\nincreased depth and larger parameter space have also led to challenges in\nproperly training a network. In light of such challenges, we propose\nhyperspherical convolution (SphereConv), a novel learning framework that gives\nangular representations on hyperspheres. We introduce SphereNet, deep\nhyperspherical convolution networks that are distinct from conventional inner\nproduct based convolutional networks. In particular, SphereNet adopts\nSphereConv as its basic convolution operator and is supervised by generalized\nangular softmax loss - a natural loss formulation under SphereConv. We show\nthat SphereNet can effectively encode discriminative representation and\nalleviate training difficulty, leading to easier optimization, faster\nconvergence and comparable (even better) classification accuracy over\nconvolutional counterparts. We also provide some theoretical insights for the\nadvantages of learning on hyperspheres. In addition, we introduce the learnable\nSphereConv, i.e., a natural improvement over prefixed SphereConv, and\nSphereNorm, i.e., hyperspherical learning as a normalization method.\nExperiments have verified our conclusions.\n",
"title": "Deep Hyperspherical Learning"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 19792 | null | Validated | null | null |
null | {
"abstract": " This paper presents a robotic pick-and-place system that is capable of\ngrasping and recognizing both known and novel objects in cluttered\nenvironments. The key new feature of the system is that it handles a wide range\nof object categories without needing any task-specific training data for novel\nobjects. To achieve this, it first uses a category-agnostic affordance\nprediction algorithm to select and execute among four different grasping\nprimitive behaviors. It then recognizes picked objects with a cross-domain\nimage classification framework that matches observed images to product images.\nSince product images are readily available for a wide range of objects (e.g.,\nfrom the web), the system works out-of-the-box for novel objects without\nrequiring any additional training data. Exhaustive experimental results\ndemonstrate that our multi-affordance grasping achieves high success rates for\na wide variety of objects in clutter, and our recognition algorithm achieves\nhigh accuracy for both known and novel grasped objects. The approach was part\nof the MIT-Princeton Team system that took 1st place in the stowing task at the\n2017 Amazon Robotics Challenge. All code, datasets, and pre-trained models are\navailable online at this http URL\n",
"title": "Robotic Pick-and-Place of Novel Objects in Clutter with Multi-Affordance Grasping and Cross-Domain Image Matching"
} | null | null | [
"Computer Science"
]
| null | true | null | 19793 | null | Validated | null | null |
null | {
"abstract": " We develop various aspects of classical enumerative geometry, including Euler\ncharacteristics and formulas for counting degenerate fibres in a pencil, with\nthe classical numerical formulas being replaced by identitites in the\nGrothendieck-Witt group of quadratic forms with coefficients in the base-field.\n",
"title": "Toward an enumerative geometry with quadratic forms"
} | null | null | null | null | true | null | 19794 | null | Default | null | null |
null | {
"abstract": " The spread of online reviews, ratings and opinions and its growing influence\non people's behavior and decisions boosted the interest to extract meaningful\ninformation from this data deluge. Hence, crowdsourced ratings of products and\nservices gained a critical role in business, governments, and others. We\npropose a new reputation-based ranking system utilizing multipartite rating\nsubnetworks, that clusters users by their similarities, using Kolmogorov\ncomplexity. Our system is novel in that it reflects a diversity of\nopinions/preferences by assigning possibly distinct rankings, for the same\nitem, for different groups of users. We prove the convergence and efficiency of\nthe system and show that it copes better with spamming/spurious users, and it\nis more robust to attacks than state-of-the-art approaches.\n",
"title": "Robust reputation-based ranking on multipartite rating networks"
} | null | null | null | null | true | null | 19795 | null | Default | null | null |
null | {
"abstract": " A dual control problem is presented for the optimal stochastic control of a\nsystem governed by partial differential equations. Relationships between the\noptimal values of the original and the dual problems are investigated and two\nduality theorems are proved. The dual problem serves to provide upper bounds\nfor the optimal and maximum value of the original one or even to give the\noptimal value.\n",
"title": "Duality and upper bounds in optimal stochastic control governed by partial differential equations"
} | null | null | null | null | true | null | 19796 | null | Default | null | null |
null | {
"abstract": " We propose Graph Priority Sampling (GPS), a new paradigm for order-based\nreservoir sampling from massive streams of graph edges. GPS provides a general\nway to weight edge sampling according to auxiliary and/or size variables so as\nto accomplish various estimation goals of graph properties. In the context of\nsubgraph counting, we show how edge sampling weights can be chosen so as to\nminimize the estimation variance of counts of specified sets of subgraphs. In\ndistinction with many prior graph sampling schemes, GPS separates the functions\nof edge sampling and subgraph estimation. We propose two estimation frameworks:\n(1) Post-Stream estimation, to allow GPS to construct a reference sample of\nedges to support retrospective graph queries, and (2) In-Stream estimation, to\nallow GPS to obtain lower variance estimates by incrementally updating the\nsubgraph count estimates during stream processing. Unbiasedness of subgraph\nestimators is established through a new Martingale formulation of graph stream\norder sampling, which shows that subgraph estimators, written as a product of\nconstituent edge estimators are unbiased, even when computed at different\npoints in the stream. The separation of estimation and sampling enables\nsignificant resource savings relative to previous work. We illustrate our\nframework with applications to triangle and wedge counting. We perform a\nlarge-scale experimental study on real-world graphs from various domains and\ntypes. GPS achieves high accuracy with less than 1% error for triangle and\nwedge counting, while storing a small fraction of the graph with average update\ntimes of a few microseconds per edge. Notably, for a large Twitter graph with\nmore than 260M edges, GPS accurately estimates triangle counts with less than\n1% error, while storing only 40K edges.\n",
"title": "On Sampling from Massive Graph Streams"
} | null | null | null | null | true | null | 19797 | null | Default | null | null |
null | {
"abstract": " This paper studies the optimal investment problem with random endowment in an\ninventory-based price impact model with competitive market makers. Our goal is\nto analyze how price impact affects optimal policies, as well as both pricing\nrules and demand schedules for contingent claims. For exponential market makers\npreferences, we establish two effects due to price impact: constrained trading,\nand non-linear hedging costs. To the former, wealth processes in the impact\nmodel are identified with those in a model without impact, but with constrained\ntrading, where the (random) constraint set is generically neither closed nor\nconvex. Regarding hedging, non-linear hedging costs motivate the study of\narbitrage free prices for the claim. We provide three such notions, which\ncoincide in the frictionless case, but which dramatically differ in the\npresence of price impact. Additionally, we show arbitrage opportunities, should\nthey arise from claim prices, can be exploited only for limited position sizes,\nand may be ignored if outweighed by hedging considerations. We also show that\narbitrage inducing prices may arise endogenously in equilibrium, and that\nequilibrium positions are inversely proportional to the market makers'\nrepresentative risk aversion. Therefore, large positions endogenously arise in\nthe limit of either market maker risk neutrality, or a large number of market\nmakers.\n",
"title": "Optimal Investment, Demand and Arbitrage under Price Impact"
} | null | null | null | null | true | null | 19798 | null | Default | null | null |
null | {
"abstract": " The emergence and nature of amplitude mediated chimera states,\nspatio-temporal patterns of co-existing coherent and incoherent regions, are\ninvestigated for a globally coupled system of active and inactive\nGinzburg-Landau oscillators. The existence domain of such states is found to\nshrink and shift in parametric space as the fraction of inactive oscillators is\nincreased. The role of inactive oscillators is found to be two fold - they get\nactivated to form a separate region of coherent oscillations and in addition\ndecrease the common collective frequency of the coherent regions by their\npresence. The dynamical origin of these effects is delineated through a\ndetailed bifurcation analysis of a reduced model equation that is based on a\nmean field approximation. Our results may have practical implications for the\nrobustness of such states in biological or physical systems where age related\ndeterioration in the functionality of components can occur.\n",
"title": "Amplitude Mediated Chimera States with Active and Inactive Oscillators"
} | null | null | [
"Physics"
]
| null | true | null | 19799 | null | Validated | null | null |
null | {
"abstract": " Photons from distant astronomical sources can be used as a classical source\nof randomness to improve fundamental tests of quantum nonlocality,\nwave-particle duality, and local realism through Bell's inequality and\ndelayed-choice quantum eraser tests inspired by Wheeler's cosmic-scale\nMach-Zehnder interferometer gedankenexperiment. Such sources of random numbers\nmay also be useful for information-theoretic applications such as key\ndistribution for quantum cryptography. Building on the design of an\n\"astronomical random-number generator\" developed for the recent \"cosmic Bell\"\nexperiment [Handsteiner et al., Phys. Rev. Lett. 118, 060401 (2017)], in this\npaper we report on the design and characterization of a device that, with\n20-nanosecond latency, outputs a bit based on whether the wavelength of an\nincoming photon is greater than or less than 700 nm. Using the one-meter\ntelescope at the Jet Propulsion Laboratory (JPL) Table Mountain Observatory, we\ngenerated random bits from astronomical photons in both color channels from 50\nstars of varying color and magnitude, and from 12 quasars with redshifts up to\n$z = 3.9$. With stars, we achieved bit rates of $\\sim 1 \\times 10^6$ Hz /\nm$^2$, limited by saturation for our single-photon detectors, and with quasars\nof magnitudes between 12.9 and 16, we achieved rates between $\\sim 10^2$ and $2\n\\times 10^3$ Hz /m$^2$. For bright quasars, the resulting bitstreams exhibit\nsufficiently low amounts of statistical predictability as quantified by the\nmutual information. In addition, a sufficiently high fraction of bits generated\nare of true astronomical origin in order to address both the locality and\nfreedom-of-choice loopholes when used to set the measurement settings in a test\nof the Bell-CHSH inequality.\n",
"title": "Astronomical random numbers for quantum foundations experiments"
} | null | null | null | null | true | null | 19800 | null | Default | null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.