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": " In many modern machine learning applications, the outcome is expensive or\ntime-consuming to collect while the predictor information is easy to obtain.\nSemi-supervised learning (SSL) aims at utilizing large amounts of `unlabeled'\ndata along with small amounts of `labeled' data to improve the efficiency of a\nclassical supervised approach. Though numerous SSL classification and\nprediction procedures have been proposed in recent years, no methods currently\nexist to evaluate the prediction performance of a working regression model. In\nthe context of developing phenotyping algorithms derived from electronic\nmedical records (EMR), we present an efficient two-step estimation procedure\nfor evaluating a binary classifier based on various prediction performance\nmeasures in the semi-supervised (SS) setting. In step I, the labeled data is\nused to obtain a non-parametrically calibrated estimate of the conditional risk\nfunction. In step II, SS estimates of the prediction accuracy parameters are\nconstructed based on the estimated conditional risk function and the unlabeled\ndata. We demonstrate that under mild regularity conditions the proposed\nestimators are consistent and asymptotically normal. Importantly, the\nasymptotic variance of the SS estimators is always smaller than that of the\nsupervised counterparts under correct model specification. We also correct for\npotential overfitting bias in the SS estimators in finite sample with\ncross-validation and develop a perturbation resampling procedure to approximate\ntheir distributions. Our proposals are evaluated through extensive simulation\nstudies and illustrated with two real EMR studies aiming to develop phenotyping\nalgorithms for rheumatoid arthritis and multiple sclerosis.\n",
"title": "Semi-Supervised Approaches to Efficient Evaluation of Model Prediction Performance"
} | null | null | null | null | true | null | 1701 | null | Default | null | null |
null | {
"abstract": " Kuniba, Okado, Takagi and Yamada have found that the time-evolution of the\nTakahashi-Satsuma box-ball system can be linearized by considering rigged\nconfigurations associated with states of the box-ball system. We introduce a\nsimple way to understand the rigged configuration of $\\mathfrak{sl}_2$-type,\nand give an elementary proof of the linearization property. Our approach can be\napplied to a box-ball system with finite carrier, which is related to a\ndiscrete modified KdV equation, and also to the combinatorial $R$-matrix of\n$A_1^{(1)}$-type. We also discuss combinatorial statistics and related\nfermionic formulas associated with the states of the box-ball systems. A\nfermionic-type formula we obtain for the finite carrier case seems to be new.\n",
"title": "Linearization of the box-ball system: an elementary approach"
} | null | null | null | null | true | null | 1702 | null | Default | null | null |
null | {
"abstract": " Scientists and engineers commonly use simulation models to study real systems\nfor which actual experimentation is costly, difficult, or impossible. Many\nsimulations are stochastic in the sense that repeated runs with the same input\nconfiguration will result in different outputs. For expensive or time-consuming\nsimulations, stochastic kriging \\citep{ankenman} is commonly used to generate\npredictions for simulation model outputs subject to uncertainty due to both\nfunction approximation and stochastic variation. Here, we develop and justify a\nfew guidelines for experimental design, which ensure accuracy of stochastic\nkriging emulators. We decompose error in stochastic kriging predictions into\nnominal, numeric, parameter estimation and parameter estimation numeric\ncomponents and provide means to control each in terms of properties of the\nunderlying experimental design. The design properties implied for each source\nof error are weakly conflicting and broad principles are proposed. In brief,\nspace-filling properties \"small fill distance\" and \"large separation distance\"\nshould balance with replication at distinct input configurations, with number\nof replications depending on the relative magnitudes of stochastic and process\nvariability. Non-stationarity implies higher input density in more active\nregions, while regression functions imply a balance with traditional design\nproperties. A few examples are presented to illustrate the results.\n",
"title": "Controlling Sources of Inaccuracy in Stochastic Kriging"
} | null | null | null | null | true | null | 1703 | null | Default | null | null |
null | {
"abstract": " Reinforcement Learning is gaining attention by the wireless networking\ncommunity due to its potential to learn good-performing configurations only\nfrom the observed results. In this work we propose a stateless variation of\nQ-learning, which we apply to exploit spatial reuse in a wireless network. In\nparticular, we allow networks to modify both their transmission power and the\nchannel used solely based on the experienced throughput. We concentrate in a\ncompletely decentralized scenario in which no information about neighbouring\nnodes is available to the learners. Our results show that although the\nalgorithm is able to find the best-performing actions to enhance aggregate\nthroughput, there is high variability in the throughput experienced by the\nindividual networks. We identify the cause of this variability as the\nadversarial setting of our setup, in which the most played actions provide\nintermittent good/poor performance depending on the neighbouring decisions. We\nalso evaluate the effect of the intrinsic learning parameters of the algorithm\non this variability.\n",
"title": "Implications of Decentralized Q-learning Resource Allocation in Wireless Networks"
} | null | null | [
"Computer Science"
]
| null | true | null | 1704 | null | Validated | null | null |
null | {
"abstract": " Non-reversible Markov chain Monte Carlo schemes based on piecewise\ndeterministic Markov processes have been recently introduced in applied\nprobability, automatic control, physics and statistics. Although these\nalgorithms demonstrate experimentally good performance and are accordingly\nincreasingly used in a wide range of applications, geometric ergodicity results\nfor such schemes have only been established so far under very restrictive\nassumptions. We give here verifiable conditions on the target distribution\nunder which the Bouncy Particle Sampler algorithm introduced in \\cite{P_dW_12}\nis geometrically ergodic. This holds whenever the target satisfies a curvature\ncondition and has tails decaying at least as fast as an exponential and at most\nas fast as a Gaussian distribution. This allows us to provide a central limit\ntheorem for the associated ergodic averages. When the target has tails thinner\nthan a Gaussian distribution, we propose an original modification of this\nscheme that is geometrically ergodic. For thick-tailed target distributions,\nsuch as $t$-distributions, we extend the idea pioneered in \\cite{J_G_12} in a\nrandom walk Metropolis context. We apply a change of variable to obtain a\ntransformed target satisfying the tail conditions for geometric ergodicity. By\nsampling the transformed target using the Bouncy Particle Sampler and mapping\nback the Markov process to the original parameterization, we obtain a\ngeometrically ergodic algorithm.\n",
"title": "Exponential Ergodicity of the Bouncy Particle Sampler"
} | null | null | [
"Statistics"
]
| null | true | null | 1705 | null | Validated | null | null |
null | {
"abstract": " These are lecture notes for the course \"MATS4300 Analysis and X-ray\ntomography\" given at the University of Jyväskylä in Fall 2017. The course\nis a broad overview of various tools in analysis that can be used to study\nX-ray tomography. The focus is on tools and ideas, not so much on technical\ndetails and minimal assumptions. Only very basic functional analysis is assumed\nas background. Exercise problems are included.\n",
"title": "Analysis and X-ray tomography"
} | null | null | null | null | true | null | 1706 | null | Default | null | null |
null | {
"abstract": " Recent studies show that widely used deep neural networks (DNNs) are\nvulnerable to carefully crafted adversarial examples. Many advanced algorithms\nhave been proposed to generate adversarial examples by leveraging the\n$\\mathcal{L}_p$ distance for penalizing perturbations. Researchers have\nexplored different defense methods to defend against such adversarial attacks.\nWhile the effectiveness of $\\mathcal{L}_p$ distance as a metric of perceptual\nquality remains an active research area, in this paper we will instead focus on\na different type of perturbation, namely spatial transformation, as opposed to\nmanipulating the pixel values directly as in prior works. Perturbations\ngenerated through spatial transformation could result in large $\\mathcal{L}_p$\ndistance measures, but our extensive experiments show that such spatially\ntransformed adversarial examples are perceptually realistic and more difficult\nto defend against with existing defense systems. This potentially provides a\nnew direction in adversarial example generation and the design of corresponding\ndefenses. We visualize the spatial transformation based perturbation for\ndifferent examples and show that our technique can produce realistic\nadversarial examples with smooth image deformation. Finally, we visualize the\nattention of deep networks with different types of adversarial examples to\nbetter understand how these examples are interpreted.\n",
"title": "Spatially Transformed Adversarial Examples"
} | null | null | null | null | true | null | 1707 | null | Default | null | null |
null | {
"abstract": " We prove that the arrow category of a monoidal model category, equipped with\nthe pushout product monoidal structure and the projective model structure, is a\nmonoidal model category. This answers a question posed by Mark Hovey, and has\nthe important consequence that it allows for the consideration of a monoidal\nproduct in cubical homotopy theory. As illustrations we include numerous\nexamples of non-cofibrantly generated monoidal model categories, including\nchain complexes, small categories, topological spaces, and pro-categories.\n",
"title": "Arrow Categories of Monoidal Model Categories"
} | null | null | null | null | true | null | 1708 | null | Default | null | null |
null | {
"abstract": " Given a suitable ordering of the positive root system associated with a\nsemisimple Lie algebra, there exists a natural correspondence between Verma\nmodules and related polynomial algebras. With this, the Lie algebra action on a\nVerma module can be interpreted as a differential operator action on\npolynomials, and thus on the corresponding truncated formal power series. We\nprove that the space of truncated formal power series is a\ndifferential-operator representation of the Weyl group $W$. We also introduce a\nsystem of partial differential equations to investigate singular vectors in the\nVerma module. It is shown that the solution space of the system in the space of\ntruncated formal power series is the span of $\\{w(1)\\ |\\ w\\in W\\}$. Those\n$w(1)$ that are polynomials correspond to singular vectors in the Verma module.\nThis elementary approach by partial differential equations also gives a new\nproof of the well-known BGG-Verma Theorem.\n",
"title": "Differential-operator representations of Weyl group and singular vectors"
} | null | null | null | null | true | null | 1709 | null | Default | null | null |
null | {
"abstract": " This paper consists of two parts. The first provides a review of the basic\nproperties of integrable and almost-toric systems, with a particular emphasis\non the integral affine structure associated to an integrable system. The second\npart introduces faithful semitoric systems, a generalization of semitoric\nsystems (introduced by Vu Ngoc and classified by Pelayo and Vu Ngoc) that\nprovides the language to develop surgeries on almost-toric systems in dimension\n4. We prove that faithful semitoric systems are natural building blocks of\nalmost-toric systems. Moreover, we show that they enjoy many of the properties\nthat their (proper) semitoric counterparts do.\n",
"title": "Faithful Semitoric Systems"
} | null | null | [
"Physics",
"Mathematics"
]
| null | true | null | 1710 | null | Validated | null | null |
null | {
"abstract": " As online fraudsters invest more resources, including purchasing large pools\nof fake user accounts and dedicated IPs, fraudulent attacks become less obvious\nand their detection becomes increasingly challenging. Existing approaches such\nas average degree maximization suffer from the bias of including more nodes\nthan necessary, resulting in lower accuracy and increased need for manual\nverification. Hence, we propose HoloScope, which uses information from graph\ntopology and temporal spikes to more accurately detect groups of fraudulent\nusers. In terms of graph topology, we introduce \"contrast suspiciousness,\" a\ndynamic weighting approach, which allows us to more accurately detect\nfraudulent blocks, particularly low-density blocks. In terms of temporal\nspikes, HoloScope takes into account the sudden bursts and drops of fraudsters'\nattacking patterns. In addition, we provide theoretical bounds for how much\nthis increases the time cost needed for fraudsters to conduct adversarial\nattacks. Additionally, from the perspective of ratings, HoloScope incorporates\nthe deviation of rating scores in order to catch fraudsters more accurately.\nMoreover, HoloScope has a concise framework and sub-quadratic time complexity,\nmaking the algorithm reproducible and scalable. Extensive experiments showed\nthat HoloScope achieved significant accuracy improvements on synthetic and real\ndata, compared with state-of-the-art fraud detection methods.\n",
"title": "HoloScope: Topology-and-Spike Aware Fraud Detection"
} | null | null | null | null | true | null | 1711 | null | Default | null | null |
null | {
"abstract": " We provide new approximation guarantees for greedy low rank matrix estimation\nunder standard assumptions of restricted strong convexity and smoothness. Our\nnovel analysis also uncovers previously unknown connections between the low\nrank estimation and combinatorial optimization, so much so that our bounds are\nreminiscent of corresponding approximation bounds in submodular maximization.\nAdditionally, we also provide statistical recovery guarantees. Finally, we\npresent empirical comparison of greedy estimation with established baselines on\ntwo important real-world problems.\n",
"title": "On Approximation Guarantees for Greedy Low Rank Optimization"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 1712 | null | Validated | null | null |
null | {
"abstract": " The topology of a power grid affects its dynamic operation and settlement in\nthe electricity market. Real-time topology identification can enable faster\ncontrol action following an emergency scenario like failure of a line. This\narticle discusses a graphical model framework for topology estimation in bulk\npower grids (both loopy transmission and radial distribution) using\nmeasurements of voltage collected from the grid nodes. The graphical model for\nthe probability distribution of nodal voltages in linear power flow models is\nshown to include additional edges along with the operational edges in the true\ngrid. Our proposed estimation algorithms first learn the graphical model and\nsubsequently extract the operational edges using either thresholding or a\nneighborhood counting scheme. For grid topologies containing no three-node\ncycles (two buses do not share a common neighbor), we prove that an exact\nextraction of the operational topology is theoretically guaranteed. This\nincludes a majority of distribution grids that have radial topologies. For\ngrids that include cycles of length three, we provide sufficient conditions\nthat ensure existence of algorithms for exact reconstruction. In particular,\nfor grids with constant impedance per unit length and uniform injection\ncovariances, this observation leads to conditions on geographical placement of\nthe buses. The performance of algorithms is demonstrated in test case\nsimulations.\n",
"title": "Topology Estimation in Bulk Power Grids: Guarantees on Exact Recovery"
} | null | null | [
"Computer Science",
"Mathematics",
"Statistics"
]
| null | true | null | 1713 | null | Validated | null | null |
null | {
"abstract": " We present Wasserstein introspective neural networks (WINN) that are both a\ngenerator and a discriminator within a single model. WINN provides a\nsignificant improvement over the recent introspective neural networks (INN)\nmethod by enhancing INN's generative modeling capability. WINN has three\ninteresting properties: (1) A mathematical connection between the formulation\nof the INN algorithm and that of Wasserstein generative adversarial networks\n(WGAN) is made. (2) The explicit adoption of the Wasserstein distance into INN\nresults in a large enhancement to INN, achieving compelling results even with a\nsingle classifier --- e.g., providing nearly a 20 times reduction in model size\nover INN for unsupervised generative modeling. (3) When applied to supervised\nclassification, WINN also gives rise to improved robustness against adversarial\nexamples in terms of the error reduction. In the experiments, we report\nencouraging results on unsupervised learning problems including texture, face,\nand object modeling, as well as a supervised classification task against\nadversarial attacks.\n",
"title": "Wasserstein Introspective Neural Networks"
} | null | null | null | null | true | null | 1714 | null | Default | null | null |
null | {
"abstract": " Let $\\Omega$ be an unbounded domain in $\\mathbb{R}\\times\\mathbb{R}^{d}.$ A\npositive harmonic function $u$ on $\\Omega$ that vanishes on the boundary of\n$\\Omega$ is called a Martin function. In this note, we show that, when $\\Omega$\nis convex, the superlevel sets of a Martin function are also convex. As a\nconsequence we obtain that if in addition $\\Omega$ is symmetric, then the\nmaximum of any Martin function along a slice $\\Omega\\cap\n(\\{t\\}\\times\\mathbb{R}^d)$ is attained at $(t,0).$\n",
"title": "Convexity of level lines of Martin functions and applications"
} | null | null | null | null | true | null | 1715 | null | Default | null | null |
null | {
"abstract": " We introduce new skein invariants of links based on a procedure where we\nfirst apply the skein relation only to crossings of distinct components, so as\nto produce collections of unlinked knots. We then evaluate the resulting knots\nusing a given invariant. A skein invariant can be computed on each link solely\nby the use of skein relations and a set of initial conditions. The new\nprocedure, remarkably, leads to generalizations of the known skein invariants.\nWe make skein invariants of classical links, $H[R]$, $K[Q]$ and $D[T]$, based\non the invariants of knots, $R$, $Q$ and $T$, denoting the regular isotopy\nversion of the Homflypt polynomial, the Kauffman polynomial and the Dubrovnik\npolynomial. We provide skein theoretic proofs of the well-definedness of these\ninvariants. These invariants are also reformulated into summations of the\ngenerating invariants ($R$, $Q$, $T$) on sublinks of a given link $L$, obtained\nby partitioning $L$ into collections of sublinks.\n",
"title": "New skein invariants of links"
} | null | null | null | null | true | null | 1716 | null | Default | null | null |
null | {
"abstract": " Big data streaming applications require utilization of heterogeneous parallel\ncomputing systems, which may comprise multiple multi-core CPUs and many-core\naccelerating devices such as NVIDIA GPUs and Intel Xeon Phis. Programming such\nsystems require advanced knowledge of several hardware architectures and\ndevice-specific programming models, including OpenMP and CUDA. In this paper,\nwe present HSTREAM, a compiler directive-based language extension to support\nprogramming stream computing applications for heterogeneous parallel computing\nsystems. HSTREAM source-to-source compiler aims to increase the programming\nproductivity by enabling programmers to annotate the parallel regions for\nheterogeneous execution and generate target specific code. The HSTREAM runtime\nautomatically distributes the workload across CPUs and accelerating devices. We\ndemonstrate the usefulness of HSTREAM language extension with various\napplications from the STREAM benchmark. Experimental evaluation results show\nthat HSTREAM can keep the same programming simplicity as OpenMP, and the\ngenerated code can deliver performance beyond what CPUs-only and GPUs-only\nexecutions can deliver.\n",
"title": "HSTREAM: A directive-based language extension for heterogeneous stream computing"
} | null | null | null | null | true | null | 1717 | null | Default | null | null |
null | {
"abstract": " The interest in higher derivatives field theories has its origin mainly in\ntheir influence concerning the renormalization properties of physical models\nand to remove ultraviolet divergences. The noncommutative Podolsky theory is a\nconstrained system that cannot by directly quantized by the canonical way. In\nthis work we have used the Faddeev-Jackiw method in order to obtain the Dirac\nbrackets of the NC Podolsky theory.\n",
"title": "Faddeev-Jackiw approach of the noncommutative spacetime Podolsky electromagnetic theory"
} | null | null | null | null | true | null | 1718 | null | Default | null | null |
null | {
"abstract": " In this work, we study the spin Hall effect and Rashba-Edelstein effect of a\n2D Weyl fermion system in the clean limit using the Kubo formalism. Spin\ntransport is solely due to the spin-torque current in this strongly spin-orbit\ncoupled (SOC) system, and chiral spin-flip scattering off non-SOC scalar\nimpurities, with potential strength $V$ and size $a$, gives rise to a\nskew-scattering mechanism for the spin Hall effect. The key result is that the\nresultant spin-Hall angle has a fixed sign, with $\\theta^{SH} \\sim O\n\\left(\\tfrac{V^2}{v_F^2/a^2} (k_F a)^4 \\right)$ being a strongly-dependent\nfunction of $k_F a$, with $k_F$ and $v_F$ being the Fermi wave-vector and Fermi\nvelocity respectively. This, therefore, allows for the possibility of tuning\nthe SHE by adjusting the Fermi energy or impurity size.\n",
"title": "Spin Transport and Accumulation in 2D Weyl Fermion System"
} | null | null | [
"Physics"
]
| null | true | null | 1719 | null | Validated | null | null |
null | {
"abstract": " This paper explores the application of Koopman operator theory to the control\nof robotic systems. The operator is introduced as a method to generate\ndata-driven models that have utility for model-based control methods. We then\nmotivate the use of the Koopman operator towards augmenting model-based\ncontrol. Specifically, we illustrate how the operator can be used to obtain a\nlinearizable data-driven model for an unknown dynamical process that is useful\nfor model-based control synthesis. Simulated results show that with increasing\ncomplexity in the choice of the basis functions, a closed-loop controller is\nable to invert and stabilize a cart- and VTOL-pendulum systems. Furthermore,\nthe specification of the basis function are shown to be of importance when\ngenerating a Koopman operator for specific robotic systems. Experimental\nresults with the Sphero SPRK robot explore the utility of the Koopman operator\nin a reduced state representation setting where increased complexity in the\nbasis function improve open- and closed-loop controller performance in various\nterrains, including sand.\n",
"title": "Model-Based Control Using Koopman Operators"
} | null | null | null | null | true | null | 1720 | null | Default | null | null |
null | {
"abstract": " Turbulence is a challenging feature common to a wide range of complex\nphenomena. Random fibre lasers are a special class of lasers in which the\nfeedback arises from multiple scattering in a one-dimensional disordered\ncavity-less medium. Here, we report on statistical signatures of turbulence in\nthe distribution of intensity fluctuations in a continuous-wave-pumped\nerbium-based random fibre laser, with random Bragg grating scatterers. The\ndistribution of intensity fluctuations in an extensive data set exhibits three\nqualitatively distinct behaviours: a Gaussian regime below threshold, a mixture\nof two distributions with exponentially decaying tails near the threshold, and\na mixture of distributions with stretched-exponential tails above threshold.\nAll distributions are well described by a hierarchical stochastic model that\nincorporates Kolmogorov's theory of turbulence, which includes energy cascade\nand the intermittence phenomenon. Our findings have implications for explaining\nthe remarkably challenging turbulent behaviour in photonics, using a random\nfibre laser as the experimental platform.\n",
"title": "Turbulence Hierarchy in a Random Fibre Laser"
} | null | null | null | null | true | null | 1721 | null | Default | null | null |
null | {
"abstract": " In the setting of nonparametric regression, we propose and study a\ncombination of stochastic gradient methods with Nyström subsampling, allowing\nmultiple passes over the data and mini-batches. Generalization error bounds for\nthe studied algorithm are provided. Particularly, optimal learning rates are\nderived considering different possible choices of the step-size, the mini-batch\nsize, the number of iterations/passes, and the subsampling level. In comparison\nwith state-of-the-art algorithms such as the classic stochastic gradient\nmethods and kernel ridge regression with Nyström, the studied algorithm has\nadvantages on the computational complexity, while achieving the same optimal\nlearning rates. Moreover, our results indicate that using mini-batches can\nreduce the total computational cost while achieving the same optimal\nstatistical results.\n",
"title": "Optimal Rates for Learning with Nyström Stochastic Gradient Methods"
} | null | null | [
"Computer Science",
"Mathematics",
"Statistics"
]
| null | true | null | 1722 | null | Validated | null | null |
null | {
"abstract": " In this work, we present theoretical results on the convergence of non-convex\naccelerated gradient descent in matrix factorization models. The technique is\napplied to matrix sensing problems with squared loss, for the estimation of a\nrank $r$ optimal solution $X^\\star \\in \\mathbb{R}^{n \\times n}$. We show that\nthe acceleration leads to linear convergence rate, even under non-convex\nsettings where the variable $X$ is represented as $U U^\\top$ for $U \\in\n\\mathbb{R}^{n \\times r}$. Our result has the same dependence on the condition\nnumber of the objective --and the optimal solution-- as that of the recent\nresults on non-accelerated algorithms. However, acceleration is observed in\npractice, both in synthetic examples and in two real applications: neuronal\nmulti-unit activities recovery from single electrode recordings, and quantum\nstate tomography on quantum computing simulators.\n",
"title": "Run Procrustes, Run! On the convergence of accelerated Procrustes Flow"
} | null | null | null | null | true | null | 1723 | null | Default | null | null |
null | {
"abstract": " Certain sufficient homological and ring-theoretical conditions are given for\na Hopf algebra to have bijective antipode with applications to noetherian Hopf\nalgebras regarding their homological behaviors.\n",
"title": "A note on the bijectivity of antipode of a Hopf algebra and its applications"
} | null | null | null | null | true | null | 1724 | null | Default | null | null |
null | {
"abstract": " Let $G$ be an undirected graph. An edge of $G$ dominates itself and all edges\nadjacent to it. A subset $E'$ of edges of $G$ is an edge dominating set of $G$,\nif every edge of the graph is dominated by some edge of $E'$. We say that $E'$\nis a perfect edge dominating set of $G$, if every edge not in $E'$ is dominated\nby exactly one edge of $E'$. The perfect edge dominating problem is to\ndetermine a least cardinality perfect edge dominating set of $G$. For this\nproblem, we describe two NP-completeness proofs, for the classes of claw-free\ngraphs of degree at most 3, and for bounded degree graphs, of maximum degree at\nmost $d \\geq 3$ and large girth. In contrast, we prove that the problem admits\nan $O(n)$ time solution, for cubic claw-free graphs. In addition, we prove a\ncomplexity dichotomy theorem for the perfect edge domination problem, based on\nthe results described in the paper. Finally, we describe a linear time\nalgorithm for finding a minimum weight perfect edge dominating set of a\n$P_5$-free graph. The algorithm is robust, in the sense that, given an\narbitrary graph $G$, either it computes a minimum weight perfect edge\ndominating set of $G$, or it exhibits an induced subgraph of $G$, isomorphic to\na $P_5$.\n",
"title": "Perfect Edge Domination: Hard and Solvable Cases"
} | null | null | null | null | true | null | 1725 | null | Default | null | null |
null | {
"abstract": " Hecke-Hopf algebras were defined by A. Berenstein and D. Kazhdan. We give an\nexplicit presentation of an Hecke-Hopf algebra when the parameter $m_{ij},$\nassociated to any two distinct vertices $i$ and $j$ in the presentation of a\nCoxeter group, equals $4,$ $5$ or $6$. As an application, we give a proof of a\nconjecture of Berenstein and Kazhdan when the Coxeter group is crystallographic\nand non-simply-laced. As another application, we show that another conjecture\nof Berenstein and Kazhdan holds when $m_{ij},$ associated to any two distinct\nvertices $i$ and $j,$ equals $4$ and that the conjecture does not hold when\nsome $m_{ij}$ equals $6$ by giving a counterexample to it.\n",
"title": "On the presentation of Hecke-Hopf algebras for non-simply-laced type"
} | null | null | null | null | true | null | 1726 | null | Default | null | null |
null | {
"abstract": " Length-matching is an important technique to bal- ance delays of bus signals\nin high-performance PCB routing. Existing routers, however, may generate very\ndense meander segments. Signals propagating along these meander segments\nexhibit a speedup effect due to crosstalk between the segments of the same\nwire, thus leading to mismatch of arrival times even under the same physical\nwire length. In this paper, we present a post-processing method to enlarge the\nwidth and the distance of meander segments and hence distribute them more\nevenly on the board so that crosstalk can be reduced. In the proposed\nframework, we model the sharing of available routing areas after removing dense\nmeander segments from the initial routing, as well as the generation of relaxed\nmeander segments and their groups for wire length compensation. This model is\ntransformed into an ILP problem and solved for a balanced distribution of wire\npatterns. In addition, we adjust the locations of long wire segments according\nto wire priorities to swap free spaces toward critical wires that need much\nlength compensation. To reduce the problem space of the ILP model, we also\nintroduce a progressive fixing technique so that wire patterns are grown\ngradually from the edge of the routing toward the center area. Experimental\nresults show that the proposed method can expand meander segments significantly\neven under very tight area constraints, so that the speedup effect can be\nalleviated effectively in high- performance PCB designs.\n",
"title": "ILP-based Alleviation of Dense Meander Segments with Prioritized Shifting and Progressive Fixing in PCB Routing"
} | null | null | null | null | true | null | 1727 | null | Default | null | null |
null | {
"abstract": " The yeast Saccharomyces cerevisiae is one of the best characterized\neukaryotic models. The secretory pathway was the first trafficking pathway\nclearly understood mainly thanks to the work done in the laboratory of Randy\nSchekman in the 1980s. They have isolated yeast sec mutants unable to secrete\nan extracellular enzyme and these SEC genes were identified as encoding key\neffectors of the secretory machinery. For this work, the 2013 Nobel Prize in\nPhysiology and Medicine has been awarded to Randy Schekman; the prize is shared\nwith James Rothman and Thomas S{ü}dhof. Here, we present the different\ntrafficking pathways of yeast S. cerevisiae. At the Golgi apparatus newly\nsynthesized proteins are sorted between those transported to the plasma\nmembrane (PM), or the external medium, via the exocytosis or secretory pathway\n(SEC), and those targeted to the vacuole either through endosomes (vacuolar\nprotein sorting or VPS pathway) or directly (alkaline phosphatase or ALP\npathway). Plasma membrane proteins can be internalized by endocytosis (END) and\ntransported to endosomes where they are sorted between those targeted for\nvacuolar degradation and those redirected to the Golgi (recycling or RCY\npathway). Studies in yeast S. cerevisiae allowed the identification of most of\nthe known effectors, protein complexes, and trafficking pathways in eukaryotic\ncells, and most of them are conserved among eukaryotes.\n",
"title": "Membrane Trafficking in the Yeast Saccharomyces cerevisiae Model"
} | null | null | null | null | true | null | 1728 | null | Default | null | null |
null | {
"abstract": " Advances in deep generative networks have led to impressive results in recent\nyears. Nevertheless, such models can often waste their capacity on the minutiae\nof datasets, presumably due to weak inductive biases in their decoders. This is\nwhere graphics engines may come in handy since they abstract away low-level\ndetails and represent images as high-level programs. Current methods that\ncombine deep learning and renderers are limited by hand-crafted likelihood or\ndistance functions, a need for large amounts of supervision, or difficulties in\nscaling their inference algorithms to richer datasets. To mitigate these\nissues, we present SPIRAL, an adversarially trained agent that generates a\nprogram which is executed by a graphics engine to interpret and sample images.\nThe goal of this agent is to fool a discriminator network that distinguishes\nbetween real and rendered data, trained with a distributed reinforcement\nlearning setup without any supervision. A surprising finding is that using the\ndiscriminator's output as a reward signal is the key to allow the agent to make\nmeaningful progress at matching the desired output rendering. To the best of\nour knowledge, this is the first demonstration of an end-to-end, unsupervised\nand adversarial inverse graphics agent on challenging real world (MNIST,\nOmniglot, CelebA) and synthetic 3D datasets.\n",
"title": "Synthesizing Programs for Images using Reinforced Adversarial Learning"
} | null | null | [
"Statistics"
]
| null | true | null | 1729 | null | Validated | null | null |
null | {
"abstract": " A number of recent papers have provided evidence that practical design\nquestions about neural networks may be tackled theoretically by studying the\nbehavior of random networks. However, until now the tools available for\nanalyzing random neural networks have been relatively ad-hoc. In this work, we\nshow that the distribution of pre-activations in random neural networks can be\nexactly mapped onto lattice models in statistical physics. We argue that\nseveral previous investigations of stochastic networks actually studied a\nparticular factorial approximation to the full lattice model. For random linear\nnetworks and random rectified linear networks we show that the corresponding\nlattice models in the wide network limit may be systematically approximated by\na Gaussian distribution with covariance between the layers of the network. In\neach case, the approximate distribution can be diagonalized by Fourier\ntransformation. We show that this approximation accurately describes the\nresults of numerical simulations of wide random neural networks. Finally, we\ndemonstrate that in each case the large scale behavior of the random networks\ncan be approximated by an effective field theory.\n",
"title": "A Correspondence Between Random Neural Networks and Statistical Field Theory"
} | null | null | [
"Computer Science",
"Physics",
"Statistics"
]
| null | true | null | 1730 | null | Validated | null | null |
null | {
"abstract": " We examine the nature, possible orbits and physical properties of the\nprogenitor of the North-western stellar stream (NWS) in the halo of the\nAndromeda galaxy (M31). The progenitor is assumed to be an accreting dwarf\ngalaxy with globular clusters (GCs). It is, in general, difficult to determine\nthe progenitor's orbit precisely because of many necessary parameters.\nRecently, Veljanoski et al. 2014 reported five GCs whose positions and radial\nvelocities suggest an association with the stream. We use this data to\nconstrain the orbital motions of the progenitor using test-particle\nsimulations. Our simulations split the orbit solutions into two branches\naccording to whether the stream ends up in the foreground or in the background\nof M31. Upcoming observations that will determine the distance to the NWS will\nbe able to reject one of the two branches. In either case, the solutions\nrequire that the pericentric radius of any possible orbit be over 2 kpc. We\nestimate the efficiency of the tidal disruption and confirm the consistency\nwith the assumption for the progenitor being a dwarf galaxy. The progenitor\nrequires the mass $\\ga 2\\times10^6 M_{\\sun}$ and half-light radius $\\ga 30$ pc.\nIn addition, $N$-body simulations successfully reproduce the basic observed\nfeatures of the NWS and the GCs' line-of-sight velocities.\n",
"title": "The nature of the progenitor of the M31 North-western stream: globular clusters as milestones of its orbit"
} | null | null | [
"Physics"
]
| null | true | null | 1731 | null | Validated | null | null |
null | {
"abstract": " In the present note we study certain arrangements of codimension $2$ flats in\nprojective spaces, we call them \"Fermat arrangements\". We describe algebraic\nproperties of their defining ideals. In particular, we show that they provide\ncounterexamples to an expected containment relation between ordinary and\nsymbolic powers of homogeneous ideals.\n",
"title": "On codimension two flats in Fermat-type arrangements"
} | null | null | [
"Mathematics"
]
| null | true | null | 1732 | null | Validated | null | null |
null | {
"abstract": " We investigate the problem of inferring the causal predictors of a response\n$Y$ from a set of $d$ explanatory variables $(X^1,\\dots,X^d)$. Classical\nordinary least squares regression includes all predictors that reduce the\nvariance of $Y$. Using only the causal predictors instead leads to models that\nhave the advantage of remaining invariant under interventions, loosely speaking\nthey lead to invariance across different \"environments\" or \"heterogeneity\npatterns\". More precisely, the conditional distribution of $Y$ given its causal\npredictors remains invariant for all observations. Recent work exploits such a\nstability to infer causal relations from data with different but known\nenvironments. We show that even without having knowledge of the environments or\nheterogeneity pattern, inferring causal relations is possible for time-ordered\n(or any other type of sequentially ordered) data. In particular, this allows\ndetecting instantaneous causal relations in multivariate linear time series\nwhich is usually not the case for Granger causality. Besides novel methodology,\nwe provide statistical confidence bounds and asymptotic detection results for\ninferring causal predictors, and present an application to monetary policy in\nmacroeconomics.\n",
"title": "Invariant Causal Prediction for Sequential Data"
} | null | null | null | null | true | null | 1733 | null | Default | null | null |
null | {
"abstract": " In state space models, smoothing refers to the task of estimating a latent\nstochastic process given noisy measurements related to the process. We propose\nan unbiased estimator of smoothing expectations. The lack-of-bias property has\nmethodological benefits: independent estimators can be generated in parallel,\nand confidence intervals can be constructed from the central limit theorem to\nquantify the approximation error. To design unbiased estimators, we combine a\ngeneric debiasing technique for Markov chains with a Markov chain Monte Carlo\nalgorithm for smoothing. The resulting procedure is widely applicable and we\nshow in numerical experiments that the removal of the bias comes at a\nmanageable increase in variance. We establish the validity of the proposed\nestimators under mild assumptions. Numerical experiments are provided on toy\nmodels, including a setting of highly-informative observations, and a realistic\nLotka-Volterra model with an intractable transition density.\n",
"title": "Smoothing with Couplings of Conditional Particle Filters"
} | null | null | null | null | true | null | 1734 | null | Default | null | null |
null | {
"abstract": " Nonlinear dynamics of the free surface of an ideal incompressible\nnon-conducting fluid with high dielectric constant subjected by strong\nhorizontal electric field is simulated on the base of the method of conformal\ntransformations. It is demonstrated that interaction of counter-propagating\nwaves leads to formation of regions with steep wave front at the fluid surface;\nangles of the boundary inclination tend to {\\pi}/2, and the curvature of\nsurface extremely increases. A significant concentration of the energy of the\nsystem occurs at these points. From the physical point of view, the appearance\nof these singularities corresponds to formation of regions at the fluid surface\nwhere pressure exerted by electric field undergoes a discontinuity and\ndynamical pressure increases almost an order of magnitude.\n",
"title": "Formation of High Pressure Gradients at the Free Surface of a Liquid Dielectric in a Tangential Electric Field"
} | null | null | null | null | true | null | 1735 | null | Default | null | null |
null | {
"abstract": " Given $n$ vectors $\\mathbf{x}_i\\in \\mathbb{R}^d$, we want to fit a linear\nregression model for noisy labels $y_i\\in\\mathbb{R}$. The ridge estimator is a\nclassical solution to this problem. However, when labels are expensive, we are\nforced to select only a small subset of vectors $\\mathbf{x}_i$ for which we\nobtain the labels $y_i$. We propose a new procedure for selecting the subset of\nvectors, such that the ridge estimator obtained from that subset offers strong\nstatistical guarantees in terms of the mean squared prediction error over the\nentire dataset of $n$ labeled vectors. The number of labels needed is\nproportional to the statistical dimension of the problem which is often much\nsmaller than $d$. Our method is an extension of a joint subsampling procedure\ncalled volume sampling. A second major contribution is that we speed up volume\nsampling so that it is essentially as efficient as leverage scores, which is\nthe main i.i.d. subsampling procedure for this task. Finally, we show\ntheoretically and experimentally that volume sampling has a clear advantage\nover any i.i.d. sampling when labels are expensive.\n",
"title": "Subsampling for Ridge Regression via Regularized Volume Sampling"
} | null | null | null | null | true | null | 1736 | null | Default | null | null |
null | {
"abstract": " While being of persistent interest for the integration of lattice-matched\nlaser devices with silicon circuits, the electronic structure of dilute nitride\nIII/V-semiconductors has presented a challenge to ab initio computational\napproaches. The root of this lies in the strong distortion N atoms exert on\nmost host materials. Here, we resolve these issues by combining density\nfunctional theory calculations based on the meta-GGA functional presented by\nTran and Blaha (TB09) with a supercell approach for the dilute nitride Ga(NAs).\nExploring the requirements posed to supercells, we show that the distortion\nfield of a single N atom must be allowed to decrease so far, that it does not\noverlap with its periodic images. This also prevents spurious electronic\ninteractions between translational symmetric atoms, allowing to compute band\ngaps in very good agreement with experimentally derived reference values. These\nresults open up the field of dilute nitride compound semiconductors to\npredictive ab initio calculations.\n",
"title": "Ab initio calculations of the concentration dependent band gap reduction in dilute nitrides"
} | null | null | null | null | true | null | 1737 | null | Default | null | null |
null | {
"abstract": " We define outliers as a set of observations which contradicts the proposed\nmathematical (statistical) model and we discuss the frequently observed types\nof the outliers. Further we explore what changes in the model have to be made\nin order to avoid the occurance of the outliers. We observe that some variants\nof the outliers lead to classical results in probability, such as the law of\nlarge numbers and the concept of heavy tailed distributions.\nKey words: outlier; the law of large numbers; heavy tailed distributions;\nmodel rejection.\n",
"title": "Outliers and related problems"
} | null | null | null | null | true | null | 1738 | null | Default | null | null |
null | {
"abstract": " We propose a DC proximal Newton algorithm for solving nonconvex regularized\nsparse learning problems in high dimensions. Our proposed algorithm integrates\nthe proximal Newton algorithm with multi-stage convex relaxation based on the\ndifference of convex (DC) programming, and enjoys both strong computational and\nstatistical guarantees. Specifically, by leveraging a sophisticated\ncharacterization of sparse modeling structures/assumptions (i.e., local\nrestricted strong convexity and Hessian smoothness), we prove that within each\nstage of convex relaxation, our proposed algorithm achieves (local) quadratic\nconvergence, and eventually obtains a sparse approximate local optimum with\noptimal statistical properties after only a few convex relaxations. Numerical\nexperiments are provided to support our theory.\n",
"title": "On Quadratic Convergence of DC Proximal Newton Algorithm for Nonconvex Sparse Learning in High Dimensions"
} | null | null | [
"Computer Science",
"Mathematics",
"Statistics"
]
| null | true | null | 1739 | null | Validated | null | null |
null | {
"abstract": " In reinforcement learning, agents learn by performing actions and observing\ntheir outcomes. Sometimes, it is desirable for a human operator to\n\\textit{interrupt} an agent in order to prevent dangerous situations from\nhappening. Yet, as part of their learning process, agents may link these\ninterruptions, that impact their reward, to specific states and deliberately\navoid them. The situation is particularly challenging in a multi-agent context\nbecause agents might not only learn from their own past interruptions, but also\nfrom those of other agents. Orseau and Armstrong defined \\emph{safe\ninterruptibility} for one learner, but their work does not naturally extend to\nmulti-agent systems. This paper introduces \\textit{dynamic safe\ninterruptibility}, an alternative definition more suited to decentralized\nlearning problems, and studies this notion in two learning frameworks:\n\\textit{joint action learners} and \\textit{independent learners}. We give\nrealistic sufficient conditions on the learning algorithm to enable dynamic\nsafe interruptibility in the case of joint action learners, yet show that these\nconditions are not sufficient for independent learners. We show however that if\nagents can detect interruptions, it is possible to prune the observations to\nensure dynamic safe interruptibility even for independent learners.\n",
"title": "Dynamic Safe Interruptibility for Decentralized Multi-Agent Reinforcement Learning"
} | null | null | null | null | true | null | 1740 | null | Default | null | null |
null | {
"abstract": " Network integration studies try to assess the impact of future developments,\nsuch as the increase of Renewable Energy Sources or the introduction of Smart\nGrid Technologies, on large-scale network areas. Goals can be to support\nstrategic alignment in the regulatory framework or to adapt the network\nplanning principles of Distribution System Operators. This study outlines an\napproach for the automated distribution system planning that can calculate\nnetwork reconfiguration, reinforcement and extension plans in a fully automated\nfashion. This allows the estimation of the expected cost in massive\nprobabilistic simulations of large numbers of real networks and constitutes a\ncore component of a framework for large-scale network integration studies.\nExemplary case study results are presented that were performed in cooperation\nwith different major distribution system operators. The case studies cover the\nestimation of expected network reinforcement costs, technical and economical\nassessment of smart grid technologies and structural network optimisation.\n",
"title": "Heuristic Optimization for Automated Distribution System Planning in Network Integration Studies"
} | null | null | null | null | true | null | 1741 | null | Default | null | null |
null | {
"abstract": " We report ALMA Cycle 2 observations of 230 GHz (1.3 mm) dust continuum\nemission, and $^{12}$CO, $^{13}$CO, and C$^{18}$O J = 2-1 line emission, from\nthe Upper Scorpius transitional disk [PZ99] J160421.7-213028, with an angular\nresolution of ~0\".25 (35 AU). Armed with these data and existing H-band\nscattered light observations, we measure the size and depth of the disk's\ncentral cavity, and the sharpness of its outer edge, in three components:\nsub-$\\mu$m-sized \"small\" dust traced by scattered light, millimeter-sized \"big\"\ndust traced by the millimeter continuum, and gas traced by line emission. Both\ndust populations feature a cavity of radius $\\sim$70 AU that is depleted by\nfactors of at least 1000 relative to the dust density just outside. The\nmillimeter continuum data are well explained by a cavity with a sharp edge.\nScattered light observations can be fitted with a cavity in small dust that has\neither a sharp edge at 60 AU, or an edge that transitions smoothly over an\nannular width of 10 AU near 60 AU. In gas, the data are consistent with a\ncavity that is smaller, about 15 AU in radius, and whose surface density at 15\nAU is $10^{3\\pm1}$ times smaller than the surface density at 70 AU; the gas\ndensity grades smoothly between these two radii. The CO isotopologue\nobservations rule out a sharp drop in gas surface density at 30 AU or a\ndouble-drop model as found by previous modeling. Future observations are needed\nto assess the nature of these gas and dust cavities, e.g., whether they are\nopened by multiple as-yet-unseen planets or photoevaporation.\n",
"title": "The Sizes and Depletions of the Dust and Gas Cavities in the Transitional Disk J160421.7-213028"
} | null | null | null | null | true | null | 1742 | null | Default | null | null |
null | {
"abstract": " The astonishing success of AlphaGo Zero\\cite{Silver_AlphaGo} invokes a\nworldwide discussion of the future of our human society with a mixed mood of\nhope, anxiousness, excitement and fear. We try to dymystify AlphaGo Zero by a\nqualitative analysis to indicate that AlphaGo Zero can be understood as a\nspecially structured GAN system which is expected to possess an inherent good\nconvergence property. Thus we deduct the success of AlphaGo Zero may not be a\nsign of a new generation of AI.\n",
"title": "Demystifying AlphaGo Zero as AlphaGo GAN"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 1743 | null | Validated | null | null |
null | {
"abstract": " Electron-doped Eu(Fe$_{0.93}$Rh$_{0.07}$)$_2$As$_2$ has been systematically\nstudied by high pressure investigations of the magnetic and electrical\ntransport properties, in order to unravel the complex interplay of\nsuperconductivity and magnetism. The compound reveals an exceedingly broad\nre-entrant transition to the superconducting state between $T_{\\rm{c,on}} =\n19.8$ K and $T_{\\rm{c,0}} = 5.2$ K due to a canted A-type antiferromagnetic\nordering of the Eu$^{2+}$ moments at $T_{\\rm{N}} = 16.6$ K and a re-entrant\nspin glass transition at $T_{\\rm{SG}} = 14.1$ K. At ambient pressure evidences\nfor the coexistence of superconductivity and ferromagnetism could be observed,\nas well as a magnetic-field-induced enhancement of the zero-resistance\ntemperature $T_{\\rm{c,0}}$ up to $7.2$ K with small magnetic fields applied\nparallel to the \\textit{ab}-plane of the crystal. We attribute the\nfield-induced-enhancement of superconductivity to the suppression of the\nferromagnetic component of the Eu$^{2+}$ moments along the \\textit{c}-axis,\nwhich leads to a reduction of the orbital pair breaking effect. Application of\nhydrostatic pressure suppresses the superconducting state around $14$ kbar\nalong with a linear temperature dependence of the resistivity, implying that a\nnon-Fermi liquid region is located at the boundary of the superconducting\nphase. At intermediate pressure, an additional feature in the resistivity\ncurves is identified, which can be suppressed by external magnetic fields and\ncompetes with the superconducting phase. We suggest that the effect of negative\npressure by the chemical Rh substitution in\nEu(Fe$_{0.93}$Rh$_{0.07}$)$_2$As$_2$ is partially reversed, leading to a\nre-activation of the spin density wave.\n",
"title": "Effects of pressure and magnetic field on the re-entrant superconductor Eu(Fe$_{0.93}$Rh$_{0.07}$)$_2$As$_2$"
} | null | null | null | null | true | null | 1744 | null | Default | null | null |
null | {
"abstract": " Chapter 16 in High-Luminosity Large Hadron Collider (HL-LHC) : Preliminary\nDesign Report. The Large Hadron Collider (LHC) is one of the largest scientific\ninstruments ever built. Since opening up a new energy frontier for exploration\nin 2010, it has gathered a global user community of about 7,000 scientists\nworking in fundamental particle physics and the physics of hadronic matter at\nextreme temperature and density. To sustain and extend its discovery potential,\nthe LHC will need a major upgrade in the 2020s. This will increase its\nluminosity (rate of collisions) by a factor of five beyond the original design\nvalue and the integrated luminosity (total collisions created) by a factor ten.\nThe LHC is already a highly complex and exquisitely optimised machine so this\nupgrade must be carefully conceived and will require about ten years to\nimplement. The new configuration, known as High Luminosity LHC (HL-LHC), will\nrely on a number of key innovations that push accelerator technology beyond its\npresent limits. Among these are cutting-edge 11-12 tesla superconducting\nmagnets, compact superconducting cavities for beam rotation with ultra-precise\nphase control, new technology and physical processes for beam collimation and\n300 metre-long high-power superconducting links with negligible energy\ndissipation. The present document describes the technologies and components\nthat will be used to realise the project and is intended to serve as the basis\nfor the detailed engineering design of HL-LHC.\n",
"title": "Commissioning and Operation"
} | null | null | null | null | true | null | 1745 | null | Default | null | null |
null | {
"abstract": " It is shown that the relativistic quantum mechanics of a single fermion can\nbe developed only on the basis of the standard representation of the Dirac\nbispinor. As in the nonrelativistic quantum mechanics, the arbitrariness in\ndefining the bispinor, as a four-component wave function, is restricted by its\nmultiplication by an arbitrary phase factor. We reveal the role of the large\nand small components of the bispinor, establish their link in the\nnonrelativistic limit with the Pauli spinor, as well as explain the role of\nstates with negative energies. The Klein tunneling is treated here as a\nphysical phenomenon analogous to the propagation of the electromagnetic wave in\na medium with negative dielectric permittivity and permeability. For the case\nof localized stationary states we define the effective one-particle operators\nwhich act in the space of the large component but contain the contributions of\nboth components. The effective operator of energy is presented in a compact\nanalytical form.\n",
"title": "Only in the standard representation the Dirac theory is a quantum theory of a single fermion"
} | null | null | null | null | true | null | 1746 | null | Default | null | null |
null | {
"abstract": " A new type of absorbing boundary conditions for molecular dynamics\nsimulations are presented. The exact boundary conditions for crystalline solids\nwith harmonic approximation are expressed as a dynamic Dirichlet- to-Neumann\n(DtN) map. It connects the displacement of the atoms at the boundary to the\ntraction on these atoms. The DtN map is valid for a domain with general\ngeometry. To avoid evaluating the time convo- lution of the dynamic DtN map, we\napproximate the associated kernel function by rational functions in the Laplace\ndomain. The parameters in the approximations are determined by interpolations.\nThe explicit forms of the zeroth, first, and second order approximations will\nbe presented. The stability of the molecular dynamics model, supplemented with\nthese absorbing boundary conditions is established. Two numerical simulations\nare performed to demonstrate the effectiveness of the methods.\n",
"title": "Stable absorbing boundary conditions for molecular dynamics in general domains"
} | null | null | [
"Physics"
]
| null | true | null | 1747 | null | Validated | null | null |
null | {
"abstract": " This paper deals with the homotopy theory of differential graded operads. We\nendow the Koszul dual category of curved conilpotent cooperads, where the\nnotion of quasi-isomorphism barely makes sense, with a model category structure\nQuillen equivalent to that of operads. This allows us to describe the homotopy\nproperties of differential graded operads in a simpler and richer way, using\nobstruction methods.\n",
"title": "Algebraic operads up to homotopy"
} | null | null | null | null | true | null | 1748 | null | Default | null | null |
null | {
"abstract": " We consider the nonlinear Kalman filtering problem using Kullback-Leibler\n(KL) and $\\alpha$-divergence measures as optimization criteria. Unlike linear\nKalman filters, nonlinear Kalman filters do not have closed form Gaussian\nposteriors because of a lack of conjugacy due to the nonlinearity in the\nlikelihood. In this paper we propose novel algorithms to optimize the forward\nand reverse forms of the KL divergence, as well as the alpha-divergence which\ncontains these two as limiting cases. Unlike previous approaches, our\nalgorithms do not make approximations to the divergences being optimized, but\nuse Monte Carlo integration techniques to derive unbiased algorithms for direct\noptimization. We assess performance on radar and sensor tracking, and options\npricing problems, showing general improvement over the UKF and EKF, as well as\ncompetitive performance with particle filtering.\n",
"title": "Nonlinear Kalman Filtering with Divergence Minimization"
} | null | null | null | null | true | null | 1749 | null | Default | null | null |
null | {
"abstract": " We prove that if $X---> X^+$ is a threefold terminal flip, then\n$c_1(X).c_2(X)\\leq c_1(X^+).c_2(X^+)$ where $c_1(X)$ and $c_2(X)$ denote the\nChern classes. This gives the affirmative answer to a Question by Xie\n\\cite{Xie2}. We obtain the similar but weaker result in the case of divisorial\ncontraction to curves.\n",
"title": "On Chern number inequality in dimension 3"
} | null | null | null | null | true | null | 1750 | null | Default | null | null |
null | {
"abstract": " The Helioseismic and Magnetic Imager (HMI) provides continuum images and\nmagnetograms with a cadence better than one per minute. It has been\ncontinuously observing the Sun 24 hours a day for the past 7 years. The obvious\ntrade-off between full disk observations and spatial resolution makes HMI not\nenough to analyze the smallest-scale events in the solar atmosphere. Our aim is\nto develop a new method to enhance HMI data, simultaneously deconvolving and\nsuper-resolving images and magnetograms. The resulting images will mimic\nobservations with a diffraction-limited telescope twice the diameter of HMI.\nOur method, which we call Enhance, is based on two deep fully convolutional\nneural networks that input patches of HMI observations and output deconvolved\nand super-resolved data. The neural networks are trained on synthetic data\nobtained from simulations of the emergence of solar active regions. We have\nobtained deconvolved and supper-resolved HMI images. To solve this ill-defined\nproblem with infinite solutions we have used a neural network approach to add\nprior information from the simulations. We test Enhance against Hinode data\nthat has been degraded to a 28 cm diameter telescope showing very good\nconsistency. The code is open source.\n",
"title": "Enhancing SDO/HMI images using deep learning"
} | null | null | null | null | true | null | 1751 | null | Default | null | null |
null | {
"abstract": " In contact with a superconductor, a normal metal modifies its properties due\nto Andreev reflection. In the current work, the local density of states (LDOS)\nof superconductor - normal metal Mo$_{78}$Ge$_{22}$ - Au bilayers are studied\nby means of STM applied from the Au side. Three bilayers have been prepared on\nsilicate glass substrate consisting of 100, 10 and 5 nm MoGe thin films covered\nalways by 5 nm Au layer. The tunneling spectra were measured at temperatures\nfrom 0.5 K to 7 K. The two-dimensional cross-correlation between topography and\nnormalized zero-bias conductance (ZBC) indicates a proximity effect between 100\nand 10 nm MoGe thin films and Au layer where a superconducting gap slightly\nsmaller than that of bulk MoGe is observed. The effect of the thinnest 5 nm\nMoGe layer on Au leads to much smaller gap moreover the LDOS reveals almost\ncompletely suppressed coherence peaks. This is attributed to a strong\npair-breaking effect of spin-flip processes at the interface between MoGe films\nand the substrate.\n",
"title": "Suppression of the superconductivity in ultrathin amorphous Mo$_{78}$Ge$_{22}$ thin films observed by STM"
} | null | null | null | null | true | null | 1752 | null | Default | null | null |
null | {
"abstract": " Functional data analysis is typically conducted within the $L^2$-Hilbert\nspace framework. There is by now a fully developed statistical toolbox allowing\nfor the principled application of the functional data machinery to real-world\nproblems, often based on dimension reduction techniques such as functional\nprincipal component analysis. At the same time, there have recently been a\nnumber of publications that sidestep dimension reduction steps and focus on a\nfully functional $L^2$-methodology. This paper goes one step further and\ndevelops data analysis methodology for functional time series in the space of\nall continuous functions. The work is motivated by the fact that objects with\nrather different shapes may still have a small $L^2$-distance and are therefore\nidentified as similar when using an $L^2$-metric. However, in applications it\nis often desirable to use metrics reflecting the visualization of the curves in\nthe statistical analysis. The methodological contributions are focused on\ndeveloping two-sample and change-point tests as well as confidence bands, as\nthese procedures appear do be conducive to the proposed setting. Particular\ninterest is put on relevant differences; that is, on not trying to test for\nexact equality, but rather for pre-specified deviations under the null\nhypothesis.\nThe procedures are justified through large-sample theory. To ensure\npracticability, non-standard bootstrap procedures are developed and\ninvestigated addressing particular features that arise in the problem of\ntesting relevant hypotheses. The finite sample properties are explored through\na simulation study and an application to annual temperature profiles.\n",
"title": "Functional data analysis in the Banach space of continuous functions"
} | null | null | null | null | true | null | 1753 | null | Default | null | null |
null | {
"abstract": " In this work we explore a straightforward variational Bayes scheme for\nRecurrent Neural Networks. Firstly, we show that a simple adaptation of\ntruncated backpropagation through time can yield good quality uncertainty\nestimates and superior regularisation at only a small extra computational cost\nduring training, also reducing the amount of parameters by 80\\%. Secondly, we\ndemonstrate how a novel kind of posterior approximation yields further\nimprovements to the performance of Bayesian RNNs. We incorporate local gradient\ninformation into the approximate posterior to sharpen it around the current\nbatch statistics. We show how this technique is not exclusive to recurrent\nneural networks and can be applied more widely to train Bayesian neural\nnetworks. We also empirically demonstrate how Bayesian RNNs are superior to\ntraditional RNNs on a language modelling benchmark and an image captioning\ntask, as well as showing how each of these methods improve our model over a\nvariety of other schemes for training them. We also introduce a new benchmark\nfor studying uncertainty for language models so future methods can be easily\ncompared.\n",
"title": "Bayesian Recurrent Neural Networks"
} | null | null | null | null | true | null | 1754 | null | Default | null | null |
null | {
"abstract": " Galaxy cross-correlations with high-fidelity redshift samples hold the\npotential to precisely calibrate systematic photometric redshift uncertainties\narising from the unavailability of complete and representative training and\nvalidation samples of galaxies. However, application of this technique in the\nDark Energy Survey (DES) is hampered by the relatively low number density,\nsmall area, and modest redshift overlap between photometric and spectroscopic\nsamples. We propose instead using photometric catalogs with reliable\nphotometric redshifts for photo-z calibration via cross-correlations. We verify\nthe viability of our proposal using redMaPPer clusters from the Sloan Digital\nSky Survey (SDSS) to successfully recover the redshift distribution of SDSS\nspectroscopic galaxies. We demonstrate how to combine photo-z with\ncross-correlation data to calibrate photometric redshift biases while\nmarginalizing over possible clustering bias evolution in either the calibration\nor unknown photometric samples. We apply our method to DES Science Verification\n(DES SV) data in order to constrain the photometric redshift distribution of a\ngalaxy sample selected for weak lensing studies, constraining the mean of the\ntomographic redshift distributions to a statistical uncertainty of $\\Delta z\n\\sim \\pm 0.01$. We forecast that our proposal can in principle control\nphotometric redshift uncertainties in DES weak lensing experiments at a level\nnear the intrinsic statistical noise of the experiment over the range of\nredshifts where redMaPPer clusters are available. Our results provide strong\nmotivation to launch a program to fully characterize the systematic errors from\nbias evolution and photo-z shapes in our calibration procedure.\n",
"title": "Cross-Correlation Redshift Calibration Without Spectroscopic Calibration Samples in DES Science Verification Data"
} | null | null | null | null | true | null | 1755 | null | Default | null | null |
null | {
"abstract": " We initiate the study of the completely bounded multipliers of the Haagerup\ntensor product $A(G)\\otimes_{\\rm h} A(G)$ of two copies of the Fourier algebra\n$A(G)$ of a locally compact group $G$. If $E$ is a closed subset of $G$ we let\n$E^{\\sharp} = \\{(s,t) : st\\in E\\}$ and show that if $E^{\\sharp}$ is a set of\nspectral synthesis for $A(G)\\otimes_{\\rm h} A(G)$ then $E$ is a set of local\nspectral synthesis for $A(G)$. Conversely, we prove that if $E$ is a set of\nspectral synthesis for $A(G)$ and $G$ is a Moore group then $E^{\\sharp}$ is a\nset of spectral synthesis for $A(G)\\otimes_{\\rm h} A(G)$. Using the natural\nidentification of the space of all completely bounded weak* continuous\n$VN(G)'$-bimodule maps with the dual of $A(G)\\otimes_{\\rm h} A(G)$, we show\nthat, in the case $G$ is weakly amenable, such a map leaves the multiplication\nalgebra of $L^{\\infty}(G)$ invariant if and only if its support is contained in\nthe antidiagonal of $G$.\n",
"title": "Completely bounded bimodule maps and spectral synthesis"
} | null | null | null | null | true | null | 1756 | null | Default | null | null |
null | {
"abstract": " Black-box risk scoring models permeate our lives, yet are typically\nproprietary or opaque. We propose Distill-and-Compare, a model distillation and\ncomparison approach to audit such models. To gain insight into black-box\nmodels, we treat them as teachers, training transparent student models to mimic\nthe risk scores assigned by black-box models. We compare the student model\ntrained with distillation to a second un-distilled transparent model trained on\nground-truth outcomes, and use differences between the two models to gain\ninsight into the black-box model. Our approach can be applied in a realistic\nsetting, without probing the black-box model API. We demonstrate the approach\non four public data sets: COMPAS, Stop-and-Frisk, Chicago Police, and Lending\nClub. We also propose a statistical test to determine if a data set is missing\nkey features used to train the black-box model. Our test finds that the\nProPublica data is likely missing key feature(s) used in COMPAS.\n",
"title": "Distill-and-Compare: Auditing Black-Box Models Using Transparent Model Distillation"
} | null | null | null | null | true | null | 1757 | null | Default | null | null |
null | {
"abstract": " To improve the efficiency of elderly assessments, an influence-based fast\npreceding questionnaire model (FPQM) is proposed. Compared with traditional\nassessments, the FPQM optimizes questionnaires by reordering their attributes.\nThe values of low-ranking attributes can be predicted by the values of the\nhigh-ranking attributes. Therefore, the number of attributes can be reduced\nwithout redesigning the questionnaires. A new function for calculating the\ninfluence of the attributes is proposed based on probability theory. Reordering\nand reducing algorithms are given based on the attributes' influences. The\nmodel is verified through a practical application. The practice in an\nelderly-care company shows that the FPQM can reduce the number of attributes by\n90.56% with a prediction accuracy of 98.39%. Compared with other methods, such\nas the Expert Knowledge, Rough Set and C4.5 methods, the FPQM achieves the best\nperformance. In addition, the FPQM can also be applied to other questionnaires.\n",
"title": "An influence-based fast preceding questionnaire model for elderly assessments"
} | null | null | null | null | true | null | 1758 | null | Default | null | null |
null | {
"abstract": " We present a GPU-accelerated version of a high-order discontinuous Galerkin\ndiscretization of the unsteady incompressible Navier-Stokes equations. The\nequations are discretized in time using a semi-implicit scheme with explicit\ntreatment of the nonlinear term and implicit treatment of the split Stokes\noperators. The pressure system is solved with a conjugate gradient method\ntogether with a fully GPU-accelerated multigrid preconditioner which is\ndesigned to minimize memory requirements and to increase overall performance. A\nsemi-Lagrangian subcycling advection algorithm is used to shift the\ncomputational load per timestep away from the pressure Poisson solve by\nallowing larger timestep sizes in exchange for an increased number of advection\nsteps. Numerical results confirm we achieve the design order accuracy in time\nand space. We optimize the performance of the most time-consuming kernels by\ntuning the fine-grain parallelism, memory utilization, and maximizing\nbandwidth. To assess overall performance we present an empirically calibrated\nroofline performance model for a target GPU to explain the achieved efficiency.\nWe demonstrate that, in the most cases, the kernels used in the solver are\nclose to their empirically predicted roofline performance.\n",
"title": "A GPU Accelerated Discontinuous Galerkin Incompressible Flow Solver"
} | null | null | null | null | true | null | 1759 | null | Default | null | null |
null | {
"abstract": " The Auger Engineering Radio Array (AERA) aims at the detection of air showers\ninduced by high-energy cosmic rays. As an extension of the Pierre Auger\nObservatory, it measures complementary information to the particle detectors,\nfluorescence telescopes and to the muon scintillators of the Auger Muons and\nInfill for the Ground Array (AMIGA). AERA is sensitive to all fundamental\nparameters of an extensive air shower such as the arrival direction, energy and\ndepth of shower maximum. Since the radio emission is induced purely by the\nelectromagnetic component of the shower, in combination with the AMIGA muon\ncounters, AERA is perfect for separate measurements of the electrons and muons\nin the shower, if combined with a muon counting detector like AMIGA. In\naddition to the depth of the shower maximum, the ratio of the electron and muon\nnumber serves as a measure of the primary particle mass.\n",
"title": "The Auger Engineering Radio Array and multi-hybrid cosmic ray detection (TAUP 2015)"
} | null | null | null | null | true | null | 1760 | null | Default | null | null |
null | {
"abstract": " Painting is an art form that has long functioned as a major channel for the\ncreative expression and communication of humans, its evolution taking place\nunder an interplay with the science, technology, and social environments of the\ntimes. Therefore, understanding the process based on comprehensive data could\nshed light on how humans acted and manifested creatively under changing\nconditions. Yet, there exist few systematic frameworks that characterize the\nprocess for painting, which would require robust statistical methods for\ndefining painting characteristics and identifying human's creative\ndevelopments, and data of high quality and sufficient quantity. Here we propose\nthat the color contrast of a painting image signifying the heterogeneity in\ninter-pixel chromatic distance can be a useful representation of its style,\nintegrating both the color and geometry. From the color contrasts of paintings\nfrom a large-scale, comprehensive archive of 179,853 high-quality images\nspanning several centuries we characterize the temporal evolutionary patterns\nof paintings, and present a deep study of an extraordinary expansion in\ncreative diversity and individuality that came to define the modern era.\n",
"title": "Historic Emergence of Diversity in Painting: Heterogeneity in Chromatic Distance in Images and Characterization of Massive Painting Data Set"
} | null | null | null | null | true | null | 1761 | null | Default | null | null |
null | {
"abstract": " Glassy dynamics is intermittent, as particles suddenly jump out of the cage\nformed by their neighbours, and heterogeneous, as these jumps are not uniformly\ndistributed across the system. Relating these features of the dynamics to the\ndiverse local environments explored by the particles is essential to\nrationalize the relaxation process. Here we investigate this issue\ncharacterizing the local environment of a particle with the amplitude of its\nshort time vibrational motion, as determined by segmenting in cages and jumps\nthe particle trajectories. Both simulations of supercooled liquids and\nexperiments on colloidal suspensions show that particles in large cages are\nlikely to jump after a small time-lag, and that, on average, the cage enlarges\nshortly before the particle jumps. At large time-lags, the cage has essentially\na constant value, which is smaller for longer-lasting cages. Finally, we\nclarify how this coupling between cage size and duration controls the average\nbehaviour and opens the way to a better understanding of the relaxation process\nin glass--forming liquids.\n",
"title": "Cage Size and Jump Precursors in Glass-Forming Liquids: Experiment and Simulations"
} | null | null | null | null | true | null | 1762 | null | Default | null | null |
null | {
"abstract": " Hidden Markov model based various phoneme recognition methods for Bengali\nlanguage is reviewed. Automatic phoneme recognition for Bengali language using\nmultilayer neural network is reviewed. Usefulness of multilayer neural network\nover single layer neural network is discussed. Bangla phonetic feature table\nconstruction and enhancement for Bengali speech recognition is also discussed.\nComparison among these methods is discussed.\n",
"title": "A Comprehensive Survey on Bengali Phoneme Recognition"
} | null | null | null | null | true | null | 1763 | null | Default | null | null |
null | {
"abstract": " In this paper, we prove that the arithmetic automorphic periods for $GL_{n}$\nover a CM field factorize through the infinite places. This generalizes a\nconjecture of Shimura in 1983, and is predicted by the Langlands correspondence\nbetween automorphic representations and motives.\n",
"title": "Factorization of arithmetic automorphic periods"
} | null | null | null | null | true | null | 1764 | null | Default | null | null |
null | {
"abstract": " In this chapter we analyze the multiple ionization by impact of |Z|=1\nprojectiles: electrons, positrons, protons and antiprotons. Differences and\nsimilarities among the cross sections by these four projectiles allows us to\nhave an insight on the physics involved. Mass and charge effects, energy\nthresholds, and relative importance of collisional and post-collisional\nprocesses are discussed. For this purpose, we performed a detailed\ntheoretical-experimental comparison for single up to quintuple ionization of\nNe, Ar, Kr and Xe by particles and antiparticles. We include an extensive\ncompilation of the available data for the sixteen collisional systems, and the\ntheoretical cross sections by means of the continuum distorted wave eikonal\ninitial state approximation. We underline here that post-collisional ionization\nis decisive to describe multiple ionization by light projectiles, covering\nalmost the whole energy range, from threshold to high energies. The\nnormalization of positron and antiproton measurements to electron impact ones,\nthe lack of data in certain cases, and the future prospects are presented and\ndiscussed.\n",
"title": "Multielectronic processes in particle and antiparticle collisions with rare gases"
} | null | null | [
"Physics"
]
| null | true | null | 1765 | null | Validated | null | null |
null | {
"abstract": " In this paper, we reconsider a circular cylinder horizontally floating on an\nunbounded reservoir in a gravitational field directed downwards, which was\nstudied by Bhatnargar and Finn in 2006. We follow their approach but with some\nmodifications. We establish the relation between the total energy relative to\nthe undisturbed state and the total force. There is a monotone relation between\nthe height of the centre and the wetting angle. We study the number of\nequilibria, the floating configurations and their stability for all parameter\nvalues. We find that the system admits at most two equilibrium points for\narbitrary contact angle, the one with smaller wetting angle is stable and the\none with larger wetting angle is unstable. The initial model has a limitation\nthat the fluid interfaces may intersect. We show that the stable equilibrium\npoint never lies in the intersection region, while the unstable equilibrium\npoint may lie in the intersection region.\n",
"title": "A Floating Cylinder on An Unbounded Bath"
} | null | null | null | null | true | null | 1766 | null | Default | null | null |
null | {
"abstract": " This paper considers the problem of switching between two periodic motions,\nalso known as limit cycles, to create agile running motions. For each limit\ncycle, we use a control Lyapunov function to estimate the region of attraction\nat the apex of the flight phase. We switch controllers at the apex, only if the\ncurrent state of the robot is within the region of attraction of the subsequent\nlimit cycle. If the intersection between two limit cycles is the null set, then\nwe construct additional limit cycles till we are able to achieve sufficient\noverlap of the region of attraction between sequential limit cycles.\nAdditionally, we impose an exponential convergence condition on the control\nLyapunov function that allows us to rapidly transition between limit cycles.\nUsing the approach we demonstrate switching between 5 limit cycles in about 5\nsteps with the speed changing from 2 m/s to 5 m/s.\n",
"title": "Switching between Limit Cycles in a Model of Running Using Exponentially Stabilizing Discrete Control Lyapunov Function"
} | null | null | null | null | true | null | 1767 | null | Default | null | null |
null | {
"abstract": " As more aspects of social interaction are digitally recorded, there is a\ngrowing need to develop privacy-preserving data analysis methods. Social\nscientists will be more likely to adopt these methods if doing so entails\nminimal change to their current methodology. Toward that end, we present a\ngeneral and modular method for privatizing Bayesian inference for Poisson\nfactorization, a broad class of models that contains some of the most widely\nused models in the social sciences. Our method satisfies local differential\nprivacy, which ensures that no single centralized server need ever store the\nnon-privatized data. To formulate our local-privacy guarantees, we introduce\nand focus on limited-precision local privacy---the local privacy analog of\nlimited-precision differential privacy (Flood et al., 2013). We present two\ncase studies, one involving social networks and one involving text corpora,\nthat test our method's ability to form the posterior distribution over latent\nvariables under different levels of noise, and demonstrate our method's utility\nover a naïve approach, wherein inference proceeds as usual, treating the\nprivatized data as if it were not privatized.\n",
"title": "Locally Private Bayesian Inference for Count Models"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 1768 | null | Validated | null | null |
null | {
"abstract": " We report the application of femtosecond four-wave mixing (FWM) to the study\nof carrier transport in solution-processed CH3NH3PbI3. The diffusion\ncoefficient was extracted through direct detection of the lateral diffusion of\ncarriers utilizing the transient grating technique, coupled with simultaneous\nmeasurement of decay kinetics exploiting the versatility of the boxcar\nexcitation beam geometry. The observation of exponential decay of the transient\ngrating versus interpulse delay indicates diffusive transport with negligible\ntrapping within the first nanosecond following excitation. The in-plane\ntransport geometry in our experiments enabled the diffusion length to be\ncompared directly with the grain size, indicating that carriers move across\nmultiple grain boundaries prior to recombination. Our experiments illustrate\nthe broad utility of FWM spectroscopy for rapid characterization of macroscopic\nfilm transport properties.\n",
"title": "Carrier Diffusion in Thin-Film CH3NH3PbI3 Perovskite Measured using Four-Wave Mixing"
} | null | null | null | null | true | null | 1769 | null | Default | null | null |
null | {
"abstract": " We introduce computable actions of computable groups and prove the following\nversions of effective Birkhoff's ergodic theorem. Let $\\Gamma$ be a computable\namenable group, then there always exists a canonically computable tempered\ntwo-sided F{\\o}lner sequence $(F_n)_{n \\geq\n1}$ in $\\Gamma$. For a computable, measure-preserving, ergodic action of\n$\\Gamma$ on a Cantor space $\\{0,1\\}^{\\mathbb N}$ endowed with a computable\nprobability measure $\\mu$, it is shown that for every bounded lower\nsemicomputable function $f$ on $\\{0,1\\}^{\\mathbb N}$ and for every Martin-Löf\nrandom $\\omega \\in \\{0,1\\}^{\\mathbb N}$ the equality \\[ \\lim\\limits_{n \\to\n\\infty} \\frac{1}{|F_n|} \\sum\\limits_{g \\in F_n} f(g \\cdot \\omega) = \\int\\limits\nf d \\mu \\] holds, where the averages are taken with respect to a canonically\ncomputable tempered two-sided F{\\o}lner sequence $(F_n)_{n \\geq\n1}$. We also prove the same identity for all lower semicomputable $f$'s in\nthe special case when $\\Gamma$ is a computable group of polynomial growth and\n$F_n:=\\mathrm{B}(n)$ is the F{\\o}lner sequence of balls around the neutral\nelement of $\\Gamma$.\n",
"title": "On effective Birkhoff's ergodic theorem for computable actions of amenable groups"
} | null | null | null | null | true | null | 1770 | null | Default | null | null |
null | {
"abstract": " Programming is a valuable skill in the labor market, making the\nunderrepresentation of women in computing an increasingly important issue.\nOnline question and answer platforms serve a dual purpose in this field: they\nform a body of knowledge useful as a reference and learning tool, and they\nprovide opportunities for individuals to demonstrate credible, verifiable\nexpertise. Issues, such as male-oriented site design or overrepresentation of\nmen among the site's elite may therefore compound the issue of women's\nunderrepresentation in IT. In this paper we audit the differences in behavior\nand outcomes between men and women on Stack Overflow, the most popular of these\nQ&A sites. We observe significant differences in how men and women participate\nin the platform and how successful they are. For example, the average woman has\nroughly half of the reputation points, the primary measure of success on the\nsite, of the average man. Using an Oaxaca-Blinder decomposition, an econometric\ntechnique commonly applied to analyze differences in wages between groups, we\nfind that most of the gap in success between men and women can be explained by\ndifferences in their activity on the site and differences in how these\nactivities are rewarded. Specifically, 1) men give more answers than women and\n2) are rewarded more for their answers on average, even when controlling for\npossible confounders such as tenure or buy-in to the site. Women ask more\nquestions and gain more reward per question. We conclude with a hypothetical\nredesign of the site's scoring system based on these behavioral differences,\ncutting the reputation gap in half.\n",
"title": "Gender Differences in Participation and Reward on Stack Overflow"
} | null | null | null | null | true | null | 1771 | null | Default | null | null |
null | {
"abstract": " We express two CR invariant surface area elements in terms of quantities in\npseudohermitian geometry. We deduce the Euler-Lagrange equations of the\nassociated energy functionals. Many solutions are given and discussed. In\nrelation to the singular CR Yamabe problem, we show that one of the energy\nfunctionals appears as the coefficient (up to a constant multiple) of the log\nterm in the associated volume renormalization.\n",
"title": "Invariant surface area functionals and singular Yamabe problem in 3-dimensional CR geometry"
} | null | null | null | null | true | null | 1772 | null | Default | null | null |
null | {
"abstract": " In this article we address the general approach for calculating dynamical\ndipole polarizabilities of small quantum systems, based on a sum-over-states\nformula involving in principle the entire energy spectrum of the system. We\ncomplement this method by a few-parameter model involving a limited number of\neffective transitions, allowing for a compact and accurate representation of\nboth the isotropic and anisotropic components of the polarizability. We apply\nthe method to the series of ten heteronuclear molecules composed of two of\n($^7$Li,$^{23}$Na,$^{39}$K,$^{87}$Rb,$^{133}$Cs) alkali-metal atoms. We rely on\nboth up-to-date spectroscopically-determined potential energy curves for the\nlowest electronic states, and on our systematic studies of these systems\nperformed during the last decade for higher excited states and for permanent\nand transition dipole moments. Such a compilation is timely for the\ncontinuously growing researches on ultracold polar molecules. Indeed the\nknowledge of the dynamic dipole polarizabilities is crucial to model the\noptical response of molecules when trapped in optical lattices, and to\ndetermine optimal lattice frequencies ensuring optimal transfer to the absolute\nground state of initially weakly-bound molecules. When they exist, we determine\nthe so-called \"magic frequencies\" where the ac-Stark shift and thus the viewed\ntrap depth, is the same for both weakly-bound and ground-state molecules.\n",
"title": "Dynamic dipole polarizabilities of heteronuclear alkali dimers: optical response, trapping and control of ultracold molecules"
} | null | null | null | null | true | null | 1773 | null | Default | null | null |
null | {
"abstract": " We show that every $H$-minor-free graph has a light $(1+\\epsilon)$-spanner,\nresolving an open problem of Grigni and Sissokho and proving a conjecture of\nGrigni and Hung. Our lightness bound is\n\\[O\\left(\\frac{\\sigma_H}{\\epsilon^3}\\log \\frac{1}{\\epsilon}\\right)\\] where\n$\\sigma_H = |V(H)|\\sqrt{\\log |V(H)|}$ is the sparsity coefficient of\n$H$-minor-free graphs. That is, it has a practical dependency on the size of\nthe minor $H$. Our result also implies that the polynomial time approximation\nscheme (PTAS) for the Travelling Salesperson Problem (TSP) in $H$-minor-free\ngraphs by Demaine, Hajiaghayi and Kawarabayashi is an efficient PTAS whose\nrunning time is $2^{O_H\\left(\\frac{1}{\\epsilon^4}\\log\n\\frac{1}{\\epsilon}\\right)}n^{O(1)}$ where $O_H$ ignores dependencies on the\nsize of $H$. Our techniques significantly deviate from existing lines of\nresearch on spanners for $H$-minor-free graphs, but build upon the work of\nChechik and Wulff-Nilsen for spanners of general graphs.\n",
"title": "Minor-free graphs have light spanners"
} | null | null | null | null | true | null | 1774 | null | Default | null | null |
null | {
"abstract": " Generative Adversarial Networks (GANs) have gathered a lot of attention from\nthe computer vision community, yielding impressive results for image\ngeneration. Advances in the adversarial generation of natural language from\nnoise however are not commensurate with the progress made in generating images,\nand still lag far behind likelihood based methods. In this paper, we take a\nstep towards generating natural language with a GAN objective alone. We\nintroduce a simple baseline that addresses the discrete output space problem\nwithout relying on gradient estimators and show that it is able to achieve\nstate-of-the-art results on a Chinese poem generation dataset. We present\nquantitative results on generating sentences from context-free and\nprobabilistic context-free grammars, and qualitative language modeling results.\nA conditional version is also described that can generate sequences conditioned\non sentence characteristics.\n",
"title": "Adversarial Generation of Natural Language"
} | null | null | null | null | true | null | 1775 | null | Default | null | null |
null | {
"abstract": " We present the first good evidence for exocomet transits of a host star in\ncontinuum light in data from the Kepler mission. The Kepler star in question,\nKIC 3542116, is of spectral type F2V and is quite bright at K_p = 10. The\ntransits have a distinct asymmetric shape with a steeper ingress and slower\negress that can be ascribed to objects with a trailing dust tail passing over\nthe stellar disk. There are three deeper transits with depths of ~0.1% that\nlast for about a day, and three that are several times more shallow and of\nshorter duration. The transits were found via an exhaustive visual search of\nthe entire Kepler photometric data set, which we describe in some detail. We\nreview the methods we use to validate the Kepler data showing the comet\ntransits, and rule out instrumental artefacts as sources of the signals. We fit\nthe transits with a simple dust-tail model, and find that a transverse comet\nspeed of ~35-50 km/s and a minimum amount of dust present in the tail of ~10^16\ng are required to explain the larger transits. For a dust replenishment time of\n~10 days, and a comet lifetime of only ~300 days, this implies a total cometary\nmass of > 3 x 10^17 g, or about the mass of Halley's comet. We also discuss the\nnumber of comets and orbital geometry that would be necessary to explain the\nsix transits detected over the four years of Kepler prime-field observations.\nFinally, we also report the discovery of a single comet-shaped transit in KIC\n11084727 with very similar transit and host-star properties.\n",
"title": "Likely Transiting Exocomets Detected by Kepler"
} | null | null | [
"Physics"
]
| null | true | null | 1776 | null | Validated | null | null |
null | {
"abstract": " We analyze the charge- and spin response functions of rare-earth nickelates\nRNiO3 and their heterostructures using random-phase approximation in a two-band\nHubbard model. The inter-orbital charge fluctuation is found to be the driving\nmechanism for the rock-salt type bond order in bulk RNiO3, and good agreement\nof the ordering temperature with experimental values is achieved for all RNiO3\nusing realistic crystal structures and interaction parameters. We further show\nthat magnetic ordering in bulk is not driven by the spin fluctuation and should\nbe instead explained as ordering of localized moments. This picture changes for\nlow-dimensional heterostructures, where the charge fluctuation is suppressed\nand overtaken by the enhanced spin instability, which results in a\nspin-density-wave ground state observed in recent experiments. Predictions for\nspectroscopy allow for further experimental testing of our claims.\n",
"title": "Origins of bond and spin order in rare-earth nickelate bulk and heterostructures"
} | null | null | null | null | true | null | 1777 | null | Default | null | null |
null | {
"abstract": " We introduce and study a notion of canonical set theoretical truth, which\nmeans truth in a `canonical model', i.e. a transitive class model that is\nuniquely characterized by some $\\in$-formula. We show that this notion of truth\nis `informative', i.e. there are statements that hold in all canonical models\nbut do not follow from ZFC, such as Reitz' ground model axiom or the\nnonexistence of measurable cardinals. We also show that ZF+$V=L[\\mathbb{R}]$+AD\nhas no canonical models. On the other hand, we show that there are canonical\nmodels for `every real has sharp'. Moreover, we consider `theory-canonical'\nstatements that only fix a transitive class model of ZFC up to elementary\nequivalence and show that it is consistent relative to large cardinals that\nthere are theory-canonical models with measurable cardinals and that\ntheory-canonicity is still informative in the sense explained above.\n",
"title": "Canonical Truth"
} | null | null | null | null | true | null | 1778 | null | Default | null | null |
null | {
"abstract": " As the number of Internet of Things (IoT) devices keeps increasing, data is\nrequired to be communicated and processed by these devices at unprecedented\nrates. Cooperation among wireless devices by exploiting Device-to-Device (D2D)\nconnections is promising, where aggregated resources in a cooperative setup can\nbe utilized by all devices, which would increase the total utility of the\nsetup. In this paper, we focus on the resource allocation problem for\ncooperating IoT devices with multiple heterogeneous applications. In\nparticular, we develop Application-Aware Cooperative Time allocation (AACT)\nframework, which optimizes the time that each application utilizes the\naggregated system resources by taking into account heterogeneous device\nconstraints and application requirements. AACT is grounded on the concept of\nRolling Horizon Control (RHC) where decisions are made by iteratively solving a\nconvex optimization problem over a moving control window of estimated system\nparameters. The simulation results demonstrate significant performance gains.\n",
"title": "AACT: Application-Aware Cooperative Time Allocation for Internet of Things"
} | null | null | null | null | true | null | 1779 | null | Default | null | null |
null | {
"abstract": " PARAFAC2 has demonstrated success in modeling irregular tensors, where the\ntensor dimensions vary across one of the modes. An example scenario is modeling\ntreatments across a set of patients with the varying number of medical\nencounters over time. Despite recent improvements on unconstrained PARAFAC2,\nits model factors are usually dense and sensitive to noise which limits their\ninterpretability. As a result, the following open challenges remain: a) various\nmodeling constraints, such as temporal smoothness, sparsity and non-negativity,\nare needed to be imposed for interpretable temporal modeling and b) a scalable\napproach is required to support those constraints efficiently for large\ndatasets. To tackle these challenges, we propose a {\\it CO}nstrained {\\it\nPA}RAFAC2 (COPA) method, which carefully incorporates optimization constraints\nsuch as temporal smoothness, sparsity, and non-negativity in the resulting\nfactors. To efficiently support all those constraints, COPA adopts a hybrid\noptimization framework using alternating optimization and alternating direction\nmethod of multiplier (AO-ADMM). As evaluated on large electronic health record\n(EHR) datasets with hundreds of thousands of patients, COPA achieves\nsignificant speedups (up to 36 times faster) over prior PARAFAC2 approaches\nthat only attempt to handle a subset of the constraints that COPA enables.\nOverall, our method outperforms all the baselines attempting to handle a subset\nof the constraints in terms of speed, while achieving the same level of\naccuracy. Through a case study on temporal phenotyping of medically complex\nchildren, we demonstrate how the constraints imposed by COPA reveal concise\nphenotypes and meaningful temporal profiles of patients. The clinical\ninterpretation of both the phenotypes and the temporal profiles was confirmed\nby a medical expert.\n",
"title": "COPA: Constrained PARAFAC2 for Sparse & Large Datasets"
} | null | null | null | null | true | null | 1780 | null | Default | null | null |
null | {
"abstract": " We model the intracluster medium as a weakly collisional plasma that is a\nbinary mixture of the hydrogen and the helium ions, along with free electrons.\nWhen, owing to the helium sedimentation, the gradient of the mean molecular\nweight (or equivalently, composition or helium ions' concentration) of the\nplasma is not negligible, it can have appreciable influence on the stability\ncriteria of the thermal convective instabilities, e.g., the heat-flux-buoyancy\ninstability and the magnetothermal instability (MTI). These instabilities are\nconsequences of the anisotropic heat conduction occurring preferentially along\nthe magnetic field lines. In this paper, without ignoring the magnetic tension,\nwe first present the mathematical criterion for the onset of composition\ngradient modified MTI. Subsequently, we relax the commonly adopted equilibrium\nstate in which the plasma is at rest, and assume that the plasma is in a\nsheared state which may be due to differential rotation. We discuss how the\nconcentration gradient affects the coupling between the Kelvin--Helmholtz\ninstability and the MTI in rendering the plasma unstable or stable. We derive\nexact stability criterion by working with the sharp boundary case in which the\nphysical variables---temperature, mean molecular weight, density, and magnetic\nfield---change discontinuously from one constant value to another on crossing\nthe boundary. Finally, we perform the linear stability analysis for the case of\nthe differentially rotating plasma that is thermally and compositionally\nstratified as well. By assuming axisymmetric perturbations, we find the\ncorresponding dispersion relation and the explicit mathematical expression\ndetermining the onset of the modified MTI.\n",
"title": "Effect of Composition Gradient on Magnetothermal Instability Modified by Shear and Rotation"
} | null | null | null | null | true | null | 1781 | null | Default | null | null |
null | {
"abstract": " An event structure is a mathematical abstraction modeling concepts as\ncausality, conflict and concurrency between events. While many other\nmathematical structures, including groups, topological spaces, rings, abound\nwith algorithms and formulas to generate, enumerate and count particular sets\nof their members, no algorithm or formulas are known to generate or count all\nthe possible event structures over a finite set of events. We present an\nalgorithm to generate such a family, along with a functional implementation\nverified using Isabelle/HOL. As byproducts, we obtain a verified enumeration of\nall possible preorders and partial orders. While the integer sequences counting\npreorders and partial orders are already listed on OEIS (On-line Encyclopedia\nof Integer Sequences), the one counting event structures is not. We therefore\nused our algorithm to submit a formally verified addition, which has been\nsuccessfully reviewed and is now part of the OEIS.\n",
"title": "A Verified Algorithm Enumerating Event Structures"
} | null | null | null | null | true | null | 1782 | null | Default | null | null |
null | {
"abstract": " Many real-world networks are known to exhibit facts that counter our\nknowledge prescribed by the theories on network creation and communication\npatterns. A common prerequisite in network analysis is that information on\nnodes and links will be complete because network topologies are extremely\nsensitive to missing information of this kind. Therefore, many real-world\nnetworks that fail to meet this criterion under random sampling may be\ndiscarded.\nIn this paper we offer a framework for interpreting the missing observations\nin network data under the hypothesis that these observations are not missing at\nrandom. We demonstrate the methodology with a case study of a financial trade\nnetwork, where the awareness of agents to the data collection procedure by a\nself-interested observer may result in strategic revealing or withholding of\ninformation. The non-random missingness has been overlooked despite the\npossibility of this being an important feature of the processes by which the\nnetwork is generated. The analysis demonstrates that strategic information\nwithholding may be a valid general phenomenon in complex systems. The evidence\nis sufficient to support the existence of an influential observer and to offer\na compelling dynamic mechanism for the creation of the network.\n",
"title": "Missing Data as Part of the Social Behavior in Real-World Financial Complex Systems"
} | null | null | null | null | true | null | 1783 | null | Default | null | null |
null | {
"abstract": " We study interacting Majorana fermions in two dimensions as a low-energy\neffective model of a vortex lattice in two-dimensional time-reversal-invariant\ntopological superconductors. For that purpose, we implement ab-initio quantum\nMonte Carlo simulation to the Majorana fermion system in which the\npath-integral measure is given by a semi-positive Pfaffian. We discuss\nspontaneous breaking of time-reversal symmetry at finite temperature.\n",
"title": "Quantum Monte Carlo simulation of a two-dimensional Majorana lattice model"
} | null | null | null | null | true | null | 1784 | null | Default | null | null |
null | {
"abstract": " We present a new class of polynomial-time algorithms for submodular function\nminimization (SFM), as well as a unified framework to obtain strongly\npolynomial SFM algorithms. Our new algorithms are based on simple iterative\nmethods for the minimum-norm problem, such as the conditional gradient and the\nFujishige-Wolfe algorithms. We exhibit two techniques to turn simple iterative\nmethods into polynomial-time algorithms.\nFirstly, we use the geometric rescaling technique, which has recently gained\nattention in linear programming. We adapt this technique to SFM and obtain a\nweakly polynomial bound $O((n^4\\cdot EO + n^5)\\log (n L))$.\nSecondly, we exhibit a general combinatorial black-box approach to turn any\nstrongly polynomial $\\varepsilon L$-approximate SFM oracle into a strongly\npolynomial exact SFM algorithm. This framework can be applied to a wide range\nof combinatorial and continuous algorithms, including pseudo-polynomial ones.\nIn particular, we can obtain strongly polynomial algorithms by a repeated\napplication of the conditional gradient or of the Fujishige-Wolfe algorithm.\nCombined with the geometric rescaling technique, the black-box approach\nprovides a $O((n^5\\cdot EO + n^6)\\log^2 n)$ algorithm. Finally, we show that\none of the techniques we develop in the paper can also be combined with the\ncutting-plane method of Lee, Sidford, and Wong, yielding a simplified variant\nof their $O(n^3 \\log^2 n \\cdot EO + n^4\\log^{O(1)} n)$ algorithm.\n",
"title": "Geometric Rescaling Algorithms for Submodular Function Minimization"
} | null | null | null | null | true | null | 1785 | null | Default | null | null |
null | {
"abstract": " PT-symmetry in optics is a condition whereby the real and imaginary parts of\nthe refractive index across a photonic structure are deliberately balanced.\nThis balance can lead to a host of novel optical phenomena, such as\nunidirectional invisibility, loss-induced lasing, single-mode lasing from\nmultimode resonators, and non-reciprocal effects in conjunction with\nnonlinearities. Because PT-symmetry has been thought of as fragile,\nexperimental realizations to date have been usually restricted to on-chip\nmicro-devices. Here, we demonstrate that certain features of PT-symmetry are\nsufficiently robust to survive the statistical fluctuations associated with a\nmacroscopic optical cavity. We construct optical-fiber-based coupled-cavities\nin excess of a kilometer in length (the free spectral range is less than 0.8\nfm) with balanced gain and loss in two sub-cavities and examine the lasing\ndynamics. In such a macroscopic system, fluctuations can lead to a\ncavity-detuning exceeding the free spectral range. Nevertheless, by varying the\ngain-loss contrast, we observe that both the lasing threshold and the growth of\nthe laser power follow the predicted behavior of a stable PT-symmetric\nstructure. Furthermore, a statistical symmetry-breaking point is observed upon\nvarying the cavity loss. These findings indicate that PT-symmetry is a more\nrobust optical phenomenon than previously expected, and points to potential\napplications in optical fiber networks and fiber lasers.\n",
"title": "Statistical PT-symmetric lasing in an optical fiber network"
} | null | null | null | null | true | null | 1786 | null | Default | null | null |
null | {
"abstract": " Single phase, uniform size (~9 nm) Cobalt Ferrite (CFO) nanoparticles have\nbeen synthesized by hydrothermal synthesis using oleic acid as a surfactant.\nThe as synthesized oleic acid coated CFO (OA-CFO) nanoparticles were well\ndispersible in nonpolar solvents but not dispersible in water. The OA-CFO\nnanoparticles have been successfully transformed to highly water dispersible\ncitric acid coated CFO (CA-CFO) nanoparticles using a novel single step ligand\nexchange process by mechanochemical milling, in which small chain citric acid\nmolecules replace the original large chain oleic acid molecules available on\nCFO nanoparticles. The contact angle measurement shows that OA-CFO\nnanoparticles are hydrophobic whereas CA-CFO nanoparticles are superhydrophilic\nin nature. The potentiality of as synthesized OA-CFO and mechanochemically\ntransformed CA-CFO nanoparticles for the demulsification of highly stabilized\nwater-in-oil and oil-in-water emulsions has been demonstrated.\n",
"title": "Transforming Single Domain Magnetic CoFe2O4 Nanoparticles from Hydrophobic to Hydrophilic By Novel Mechanochemical Ligand Exchange"
} | null | null | null | null | true | null | 1787 | null | Default | null | null |
null | {
"abstract": " The Mid-Infrared Instrument (MIRI) for the {\\em James Webb Space Telescope}\n(JWST) will revolutionize our understanding of infrared stellar populations in\nthe Local Volume. Using the rich {\\em Spitzer}-IRS spectroscopic data-set and\nspectral classifications from the Surveying the Agents of Galaxy Evolution\n(SAGE)-Spectroscopic survey of over a thousand objects in the Magellanic\nClouds, the Grid of Red supergiant and Asymptotic giant branch star ModelS\n({\\sc grams}), and the grid of YSO models by Robitaille et al. (2006), we\ncalculate the expected flux-densities and colors in the MIRI broadband filters\nfor prominent infrared stellar populations. We use these fluxes to explore the\n{\\em JWST}/MIRI colours and magnitudes for composite stellar population studies\nof Local Volume galaxies. MIRI colour classification schemes are presented;\nthese diagrams provide a powerful means of identifying young stellar objects,\nevolved stars and extragalactic background galaxies in Local Volume galaxies\nwith a high degree of confidence. Finally, we examine which filter combinations\nare best for selecting populations of sources based on their JWST colours.\n",
"title": "Probing the dusty stellar populations of the Local Volume Galaxies with JWST/MIRI"
} | null | null | [
"Physics"
]
| null | true | null | 1788 | null | Validated | null | null |
null | {
"abstract": " This paper describes an Open Source Software (OSS) project: PythonRobotics.\nThis is a collection of robotics algorithms implemented in the Python\nprogramming language. The focus of the project is on autonomous navigation, and\nthe goal is for beginners in robotics to understand the basic ideas behind each\nalgorithm. In this project, the algorithms which are practical and widely used\nin both academia and industry are selected. Each sample code is written in\nPython3 and only depends on some standard modules for readability and ease of\nuse. It includes intuitive animations to understand the behavior of the\nsimulation.\n",
"title": "PythonRobotics: a Python code collection of robotics algorithms"
} | null | null | null | null | true | null | 1789 | null | Default | null | null |
null | {
"abstract": " This paper is concerned with qualitative properties of bounded steady flows\nof an ideal incompressible fluid with no stagnation point in the\ntwo-dimensional plane R^2. We show that any such flow is a shear flow, that is,\nit is parallel to some constant vector. The proof of this Liouville-type result\nis firstly based on the study of the geometric properties of the level curves\nof the stream function and secondly on the derivation of some estimates on the\nat most logarithmic growth of the argument of the flow. These estimates lead to\nthe conclusion that the streamlines of the flow are all parallel lines.\n",
"title": "A Liouville theorem for the Euler equations in the plane"
} | null | null | null | null | true | null | 1790 | null | Default | null | null |
null | {
"abstract": " We have developed an Electron Tracking Compton Camera (ETCC), which provides\na well-defined Point Spread Function (PSF) by reconstructing a direction of\neach gamma as a point and realizes simultaneous measurement of brightness and\nspectrum of MeV gamma-rays for the first time. Here, we present the results of\nour on-site pilot gamma-imaging-spectroscopy with ETCC at three contaminated\nlocations in the vicinity of the Fukushima Daiichi Nuclear Power Plants in\nJapan in 2014. The obtained distribution of brightness (or emissivity) with\nremote-sensing observations is unambiguously converted into the dose\ndistribution. We confirm that the dose distribution is consistent with the one\ntaken by conventional mapping measurements with a dosimeter physically placed\nat each grid point. Furthermore, its imaging spectroscopy, boosted by\nCompton-edge-free spectra, reveals complex radioactive features in a\nquantitative manner around each individual target point in the\nbackground-dominated environment. Notably, we successfully identify a \"micro\nhot spot\" of residual caesium contamination even in an already decontaminated\narea. These results show that the ETCC performs exactly as the geometrical\noptics predicts, demonstrates its versatility in the field radiation\nmeasurement, and reveals potentials for application in many fields, including\nthe nuclear industry, medical field, and astronomy.\n",
"title": "First On-Site True Gamma-Ray Imaging-Spectroscopy of Contamination near Fukushima Plant"
} | null | null | null | null | true | null | 1791 | null | Default | null | null |
null | {
"abstract": " Using large-scale simulations based on matrix product state and quantum Monte\nCarlo techniques, we study the superfluid to Bose glass-transition for\none-dimensional attractive hard-core bosons at zero temperature, across the\nfull regime from weak to strong disorder. As a function of interaction and\ndisorder strength, we identify a Berezinskii-Kosterlitz-Thouless critical line\nwith two different regimes. At small attraction where critical disorder is weak\ncompared to the bandwidth, the critical Luttinger parameter $K_c$ takes its\nuniversal Giamarchi-Schulz value $K_{c}=3/2$. Conversely, a non-universal\n$K_c>3/2$ emerges for stronger attraction where weak-link physics is relevant.\nIn this strong disorder regime, the transition is characterized by self-similar\npower-law distributed weak links with a continuously varying characteristic\nexponent $\\alpha$.\n",
"title": "Weak Versus Strong Disorder Superfluid-Bose Glass Transition in One Dimension"
} | null | null | null | null | true | null | 1792 | null | Default | null | null |
null | {
"abstract": " {\\it Ellsberg thought experiments} and empirical confirmation of Ellsberg\npreferences pose serious challenges to {\\it subjective expected utility theory}\n(SEUT). We have recently elaborated a quantum-theoretic framework for human\ndecisions under uncertainty which satisfactorily copes with the Ellsberg\nparadox and other puzzles of SEUT. We apply here the quantum-theoretic\nframework to the {\\it Ellsberg two-urn example}, showing that the paradox can\nbe explained by assuming a state change of the conceptual entity that is the\nobject of the decision ({\\it decision-making}, or {\\it DM}, {\\it entity}) and\nrepresenting subjective probabilities by quantum probabilities. We also model\nthe empirical data we collected in a DM test on human participants within the\ntheoretic framework above. The obtained results are relevant, as they provide a\nline to model real life, e.g., financial and medical, decisions that show the\nsame empirical patterns as the two-urn experiment.\n",
"title": "Quantum Structures in Human Decision-making: Towards Quantum Expected Utility"
} | null | null | null | null | true | null | 1793 | null | Default | null | null |
null | {
"abstract": " Event learning is one of the most important problems in AI. However,\nnotwithstanding significant research efforts, it is still a very complex task,\nespecially when the events involve the interaction of humans or agents with\nother objects, as it requires modeling human kinematics and object movements.\nThis study proposes a methodology for learning complex human-object interaction\n(HOI) events, involving the recording, annotation and classification of event\ninteractions. For annotation, we allow multiple interpretations of a motion\ncapture by slicing over its temporal span, for classification, we use\nLong-Short Term Memory (LSTM) sequential models with Conditional Randon Field\n(CRF) for constraints of outputs. Using a setup involving captures of\nhuman-object interaction as three dimensional inputs, we argue that this\napproach could be used for event types involving complex spatio-temporal\ndynamics.\n",
"title": "Fine-grained Event Learning of Human-Object Interaction with LSTM-CRF"
} | null | null | null | null | true | null | 1794 | null | Default | null | null |
null | {
"abstract": " Review of the third edition of \"Interferometry and Synthesis in Radio\nAstronomy\" by Thompson, Moran and Swenson\n",
"title": "Book Review Interferometry and Synthesis in Radio Astronomy - 3rd Ed"
} | null | null | null | null | true | null | 1795 | null | Default | null | null |
null | {
"abstract": " Data augmentation, a technique in which a training set is expanded with\nclass-preserving transformations, is ubiquitous in modern machine learning\npipelines. In this paper, we seek to establish a theoretical framework for\nunderstanding modern data augmentation techniques. We start by showing that for\nkernel classifiers, data augmentation can be approximated by first-order\nfeature averaging and second-order variance regularization components. We\nconnect this general approximation framework to prior work in invariant\nkernels, tangent propagation, and robust optimization. Next, we explicitly\ntackle the compositional aspect of modern data augmentation techniques,\nproposing a novel model of data augmentation as a Markov process. Under this\nmodel, we show that performing $k$-nearest neighbors with data augmentation is\nasymptotically equivalent to a kernel classifier. Finally, we illustrate ways\nin which our theoretical framework can be leveraged to accelerate machine\nlearning workflows in practice, including reducing the amount of computation\nneeded to train on augmented data, and predicting the utility of a\ntransformation prior to training.\n",
"title": "A Kernel Theory of Modern Data Augmentation"
} | null | null | null | null | true | null | 1796 | null | Default | null | null |
null | {
"abstract": " Transition metal oxides are well known for their complex magnetic and\nelectrical properties. When brought together in heterostructure geometries,\nthey show particular promise for spintronics and colossal magnetoresistance\napplications. In this letter, we propose a new mechanism for the coupling\nbetween layers of itinerant ferromagnetic materials in heterostructures. The\ncoupling is mediated by charge carriers that strive to maximally delocalize\nthrough the heterostructure to gain kinetic energy. In doing so, they force a\nferromagnetic or antiferromagnetic coupling between the constituent layers. To\nillustrate this, we focus on heterostructures composed of SrRuO$_3$ and\nLa$_{1-x}$A$_{x}$MnO$_3$ (A=Ca/Sr). Our mechanism is consistent with\nantiferromagnetic alignment that is known to occur in multilayers of\nSrRuO$_3$-La$_{1-x}$A$_{x}$MnO$_3$. To support our assertion, we present a\nminimal Kondo-lattice model which reproduces the known magnetization properties\nof such multilayers. In addition, we discuss a quantum well model for\nheterostructures and argue that the spin-dependent density of states determines\nthe nature of the coupling. As a smoking gun signature, we propose that\nbilayers with the same constituents will oscillate between ferromagnetic and\nantiferromagnetic coupling upon tuning the relative thicknesses of the layers.\n",
"title": "Carrier driven coupling in ferromagnetic oxide heterostructures"
} | null | null | null | null | true | null | 1797 | null | Default | null | null |
null | {
"abstract": " An important problem in training deep networks with high capacity is to\nensure that the trained network works well when presented with new inputs\noutside the training dataset. Dropout is an effective regularization technique\nto boost the network generalization in which a random subset of the elements of\nthe given data and the extracted features are set to zero during the training\nprocess. In this paper, a new randomized regularization technique in which we\nwithhold a random part of the data without necessarily turning off the\nneurons/data-elements is proposed. In the proposed method, of which the\nconventional dropout is shown to be a special case, random data dropout is\nperformed in an arbitrary basis, hence the designation Generalized Dropout. We\nalso present a framework whereby the proposed technique can be applied\nefficiently to convolutional neural networks. The presented numerical\nexperiments demonstrate that the proposed technique yields notable performance\ngain. Generalized Dropout provides new insight into the idea of dropout, shows\nthat we can achieve different performance gains by using different bases\nmatrices, and opens up a new research question as of how to choose optimal\nbases matrices that achieve maximal performance gain.\n",
"title": "Data Dropout in Arbitrary Basis for Deep Network Regularization"
} | null | null | null | null | true | null | 1798 | null | Default | null | null |
null | {
"abstract": " Recent large cancer studies have measured somatic alterations in an\nunprecedented number of tumours. These large datasets allow the identification\nof cancer-related sets of genetic alterations by identifying relevant\ncombinatorial patterns. Among such patterns, mutual exclusivity has been\nemployed by several recent methods that have shown its effectivenes in\ncharacterizing gene sets associated to cancer. Mutual exclusivity arises\nbecause of the complementarity, at the functional level, of alterations in\ngenes which are part of a group (e.g., a pathway) performing a given function.\nThe availability of quantitative target profiles, from genetic perturbations or\nfrom clinical phenotypes, provides additional information that can be leveraged\nto improve the identification of cancer related gene sets by discovering groups\nwith complementary functional associations with such targets.\nIn this work we study the problem of finding groups of mutually exclusive\nalterations associated with a quantitative (functional) target. We propose a\ncombinatorial formulation for the problem, and prove that the associated\ncomputation problem is computationally hard. We design two algorithms to solve\nthe problem and implement them in our tool UNCOVER. We provide analytic\nevidence of the effectiveness of UNCOVER in finding high-quality solutions and\nshow experimentally that UNCOVER finds sets of alterations significantly\nassociated with functional targets in a variety of scenarios. In addition, our\nalgorithms are much faster than the state-of-the-art, allowing the analysis of\nlarge datasets of thousands of target profiles from cancer cell lines. We show\nthat on one such dataset from project Achilles our methods identify several\nsignificant gene sets with complementary functional associations with targets.\n",
"title": "Efficient algorithms to discover alterations with complementary functional association in cancer"
} | null | null | null | null | true | null | 1799 | null | Default | null | null |
null | {
"abstract": " The paper introduces Laplace-type operators for functions defined on the\ntangent space of a Finsler Lie algebroid, using a volume form on the\nprolongation of the algebroid. It also presents the construction of a\nhorizontal Laplace operator for forms defined on the prolongation of the\nalgebroid. All of the Laplace operators considered in the paper are also\nlocally expressed using the Chern-Finsler connection of the algebroid.\n",
"title": "Laplace operators on holomorphic Lie algebroids"
} | null | null | null | null | true | null | 1800 | null | Default | null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.