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": " We study polynomial generalizations of the Kontsevich automorphisms acting on\nthe skew-field of formal rational expressions in two non-commuting variables.\nOur main result is the Laurentness and pseudo-positivity of iterations of these\nautomorphisms. The resulting expressions are described combinatorially using a\ngeneralization of the combinatorics of compatible pairs in a maximal Dyck path\ndeveloped by Lee, Li, and Zelevinsky. By specializing to quasi-commuting\nvariables we obtain pseudo-positive expressions for rank 2 quantum generalized\ncluster variables. In the case that all internal exchange coefficients are\nzero, this quantum specialization provides a combinatorial construction of\ncounting polynomials for Grassmannians of submodules in exceptional\nrepresentations of valued quivers with two vertices.\n",
"title": "Rank Two Non-Commutative Laurent Phenomenon and Pseudo-Positivity"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3901
| null |
Validated
| null | null |
null |
{
"abstract": " Coordinate descent methods employ random partial updates of decision\nvariables in order to solve huge-scale convex optimization problems. In this\nwork, we introduce new adaptive rules for the random selection of their\nupdates. By adaptive, we mean that our selection rules are based on the dual\nresidual or the primal-dual gap estimates and can change at each iteration. We\ntheoretically characterize the performance of our selection rules and\ndemonstrate improvements over the state-of-the-art, and extend our theory and\nalgorithms to general convex objectives. Numerical evidence with hinge-loss\nsupport vector machines and Lasso confirm that the practice follows the theory.\n",
"title": "Faster Coordinate Descent via Adaptive Importance Sampling"
}
| null | null | null | null | true | null |
3902
| null |
Default
| null | null |
null |
{
"abstract": " Stochasticity and limited precision of synaptic weights in neural network\nmodels are key aspects of both biological and hardware modeling of learning\nprocesses. Here we show that a neural network model with stochastic binary\nweights naturally gives prominence to exponentially rare dense regions of\nsolutions with a number of desirable properties such as robustness and good\ngeneralization performance, while typical solutions are isolated and hard to\nfind. Binary solutions of the standard perceptron problem are obtained from a\nsimple gradient descent procedure on a set of real values parametrizing a\nprobability distribution over the binary synapses. Both analytical and\nnumerical results are presented. An algorithmic extension aimed at training\ndiscrete deep neural networks is also investigated.\n",
"title": "On the role of synaptic stochasticity in training low-precision neural networks"
}
| null | null | null | null | true | null |
3903
| null |
Default
| null | null |
null |
{
"abstract": " A congruence is a surface in the Grassmannian $\\mathrm{Gr}(1,\\mathbb{P}^3)$\nof lines in projective $3$-space. To a space curve $C$, we associate the Chow\nhypersurface in $\\mathrm{Gr}(1,\\mathbb{P}^3)$ consisting of all lines which\nintersect $C$. We compute the singular locus of this hypersurface, which\ncontains the congruence of all secants to $C$. A surface $S$ in $\\mathbb{P}^3$\ndefines the Hurwitz hypersurface in $\\mathrm{Gr}(1,\\mathbb{P}^3)$ of all lines\nwhich are tangent to $S$. We show that its singular locus has two components\nfor general enough $S$: the congruence of bitangents and the congruence of\ninflectional tangents. We give new proofs for the bidegrees of the secant,\nbitangent and inflectional congruences, using geometric techniques such as\nduality, polar loci and projections. We also study the singularities of these\ncongruences.\n",
"title": "Secants, bitangents, and their congruences"
}
| null | null | null | null | true | null |
3904
| null |
Default
| null | null |
null |
{
"abstract": " We investigate the behavior of the deviation of the estimator for the density\nof states (DOS) with respect to the exact solution in the course of Wang-Landau\nand Stochastic Approximation Monte Carlo (SAMC) simulations of the\ntwo-dimensional Ising model. We find that the deviation saturates in the\nWang-Landau case. This can be cured by adjusting the refinement scheme. To this\nend, the 1/t-modification of the Wang-Landau algorithm has been suggested. A\nsimilar choice of refinement scheme is employed in the SAMC algorithm. The\nconvergence behavior of all three algorithms is examined. It turns out that the\nconvergence of the SAMC algorithm is very sensitive to the onset of the\nrefinement. Finally, the internal energy and specific heat of the Ising model\nare calculated from the SAMC DOS and compared to exact values.\n",
"title": "Convergence of Stochastic Approximation Monte Carlo and modified Wang-Landau algorithms: Tests for the Ising model"
}
| null | null | null | null | true | null |
3905
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we prove the pointwise convergence and the rate of pointwise\nconvergence for a family of singular integral operators in two-dimensional\nsetting in the following form: \\begin{equation*} L_{\\lambda }\\left(\nf;x,y\\right) =\\underset{D}{\\iint }f\\left( t,s\\right) K_{\\lambda }\\left(\nt-x,s-y\\right) dsdt,\\text{ }\\left( x,y\\right) \\in D, \\end{equation*} where\n$D=\\left \\langle a,b\\right \\rangle \\times \\left \\langle c,d\\right \\rangle $ is\nan arbitrary closed, semi-closed or open rectangle in $\\mathbb{R}^{2}$ and $%\n\\lambda \\in \\Lambda ,$ $\\Lambda $ is a set of non-negative indices with\naccumulation point $\\lambda_{0}$. Also, we provide an example to support these\ntheoretical results. In contrast to previous works, the kernel function\n$K_{\\lambda }\\left( t,s\\right) $ does not have to be even, positive or 2$\\pi\n-$periodic.\n",
"title": "On the approximation by convolution type double singular integral operators"
}
| null | null | null | null | true | null |
3906
| null |
Default
| null | null |
null |
{
"abstract": " We give a criterion which characterizes a homogeneous real multi-variate\npolynomial to have the property that all sufficiently large powers of the\npolynomial (as well as their products with any given positive homogeneous\npolynomial) have positive coefficients. Our result generalizes a result of De\nAngelis, which corresponds to the case of homogeneous bi-variate polynomials,\nas well as a classical result of Pólya, which corresponds to the case of a\nspecific linear polynomial. As an application, we also give a characterization\nof certain polynomial beta functions, which are the spectral radius functions\nof the defining matrix functions of Markov chains.\n",
"title": "Characterization of polynomials whose large powers have all positive coefficients"
}
| null | null | null | null | true | null |
3907
| null |
Default
| null | null |
null |
{
"abstract": " The Moon often appears larger near the perceptual horizon and smaller high in\nthe sky though the visual angle subtended is invariant. We show how this\nillusion results from the optimization of a projective geometrical frame for\nconscious perception through free energy minimization, as articulated in the\nProjective Consciousness Model. The model accounts for all documented\nmodulations of the illusion without anomalies (e.g., the size-distance\nparadox), surpasses other theories in explanatory power, makes sense of inter-\nand intra-subjective variability vis-a-vis the illusion, and yields new\nquantitative and qualitative predictions.\n",
"title": "The Moon Illusion explained by the Projective Consciousness Model"
}
| null | null | null | null | true | null |
3908
| null |
Default
| null | null |
null |
{
"abstract": " We study field-driven magnetic domain wall dynamics in garnet strips by\nlarge-scale three-dimensional micromagnetic simulations. The domain wall\npropagation velocity as a function of the applied field exhibits a low-field\nlinear part terminated by a sudden velocity drop at a threshold field\nmagnitude, related to the onset of excitations of internal degrees of freedom\nof the domain wall magnetization. By considering a wide range of strip\nthicknesses from 30 nm to 1.89 $\\mu$m, we find a non-monotonic thickness\ndependence of the threshold field for the onset of this instability, proceeding\nvia nucleation and propagation of Bloch lines within the domain wall. We\nidentify a critical strip thickness above which the velocity drop is due to\nnucleation of horizontal Bloch lines, while for thinner strips and depending on\nthe boundary conditions employed, either generation of vertical Bloch lines, or\nclose-to-uniform precession of the domain wall internal magnetization takes\nplace. For strips of intermediate thicknesses, the vertical Bloch lines assume\na deformed structure due to demagnetizing fields at the strip surfaces,\nbreaking the symmetry between the top and bottom faces of the strip, and\nresulting in circulating Bloch line dynamics along the perimeter of the domain\nwall.\n",
"title": "Bloch line dynamics within moving domain walls in 3D ferromagnets"
}
| null | null | null | null | true | null |
3909
| null |
Default
| null | null |
null |
{
"abstract": " We show that the UCT problem for separable, nuclear $\\mathrm C^*$-algebras\nrelies only on whether the UCT holds for crossed products of certain finite\ncyclic group actions on the Razak-Jacelon algebra. This observation is\nanalogous to and in fact recovers a characterization of the UCT problem in\nterms of finite group actions on the Cuntz algebra $\\mathcal O_2$ established\nin previous work by the authors. Although based on a similar approach, the new\nconceptual ingredients in the finite context are the recent advances in the\nclassification of stably projectionless $\\mathrm C^*$-algebras, as well as a\nknown characterization of the UCT problem in terms of certain tracially AF\n$\\mathrm C^*$-algebras due to Dadarlat.\n",
"title": "Approaching the UCT problem via crossed products of the Razak-Jacelon algebra"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3910
| null |
Validated
| null | null |
null |
{
"abstract": " Recent outbreaks of Ebola, H1N1 and other infectious diseases have shown that\nthe assumptions underlying the established theory of epidemics management are\ntoo idealistic. For an improvement of procedures and organizations involved in\nfighting epidemics, extended models of epidemics management are required. The\nnecessary extensions consist in a representation of the management loop and the\npotential frictions influencing the loop. The effects of the non-deterministic\nfrictions can be taken into account by including the measures of robustness and\nrisk in the assessment of management options. Thus, besides of the increased\nstructural complexity resulting from the model extensions, the computational\ncomplexity of the task of epidemics management - interpreted as an optimization\nproblem - is increased as well. This is a serious obstacle for analyzing the\nmodel and may require an additional pre-processing enabling a simplification of\nthe analysis process. The paper closes with an outlook discussing some\nforthcoming problems.\n",
"title": "Some Remarks about the Complexity of Epidemics Management"
}
| null | null | null | null | true | null |
3911
| null |
Default
| null | null |
null |
{
"abstract": " We investigate the entanglement properties of an infinite class of excited\nstates in the quantum Lifshitz model (QLM). The presence of a conformal quantum\ncritical point in the QLM makes it unusually tractable for a model above one\nspatial dimension, enabling the ground state entanglement entropy for an\narbitrary domain to be expressed in terms of geometrical and topological\nquantities. Here we extend this result to excited states and find that the\nentanglement can be naturally written in terms of quantities which we dub\n\"entanglement propagator amplitudes\" (EPAs). EPAs are geometrical probabilities\nthat we explicitly calculate and interpret. A comparison of lattice and\ncontinuum results demonstrates that EPAs are universal. This work shows that\nthe QLM is an example of a 2+1d field theory where the universal behavior of\nexcited-state entanglement may be computed analytically.\n",
"title": "Entanglement Entropy in Excited States of the Quantum Lifshitz Model"
}
| null | null |
[
"Physics"
] | null | true | null |
3912
| null |
Validated
| null | null |
null |
{
"abstract": " We present a nonlocal electrostatic formulation of nonuniform ions and water\nmolecules with interstitial voids that uses a Fermi-like distribution to\naccount for steric and correlation effects in electrolyte solutions. The\nformulation is based on the volume exclusion of hard spheres leading to a\nsteric potential and Maxwell's displacement field with Yukawa-type interactions\nresulting in a nonlocal electric potential. The classical Poisson-Boltzmann\nmodel fails to describe steric and correlation effects important in a variety\nof chemical and biological systems, especially in high field or large\nconcentration conditions found in and near binding sites, ion channels, and\nelectrodes. Steric effects and correlations are apparent when we compare\nnonlocal Poisson-Fermi results to Poisson-Boltzmann calculations in electric\ndouble layer and to experimental measurements on the selectivity of potassium\nchannels for K+ over Na+. The present theory links atomic scale descriptions of\nthe crystallized KcsA channel with macroscopic bulk conditions. Atomic\nstructures and macroscopic conditions determine complex functions of great\nimportance in biology, nanotechnology, and electrochemistry.\n",
"title": "Poisson-Fermi Formulation of Nonlocal Electrostatics in Electrolyte Solutions"
}
| null | null | null | null | true | null |
3913
| null |
Default
| null | null |
null |
{
"abstract": " Transfer learning borrows knowledge from a source domain to facilitate\nlearning in a target domain. Two primary issues to be addressed in transfer\nlearning are what and how to transfer. For a pair of domains, adopting\ndifferent transfer learning algorithms results in different knowledge\ntransferred between them. To discover the optimal transfer learning algorithm\nthat maximally improves the learning performance in the target domain,\nresearchers have to exhaustively explore all existing transfer learning\nalgorithms, which is computationally intractable. As a trade-off, a sub-optimal\nalgorithm is selected, which requires considerable expertise in an ad-hoc way.\nMeanwhile, it is widely accepted in educational psychology that human beings\nimprove transfer learning skills of deciding what to transfer through\nmeta-cognitive reflection on inductive transfer learning practices. Motivated\nby this, we propose a novel transfer learning framework known as Learning to\nTransfer (L2T) to automatically determine what and how to transfer are the best\nby leveraging previous transfer learning experiences. We establish the L2T\nframework in two stages: 1) we first learn a reflection function encrypting\ntransfer learning skills from experiences; and 2) we infer what and how to\ntransfer for a newly arrived pair of domains by optimizing the reflection\nfunction. Extensive experiments demonstrate the L2T's superiority over several\nstate-of-the-art transfer learning algorithms and its effectiveness on\ndiscovering more transferable knowledge.\n",
"title": "Learning to Transfer"
}
| null | null | null | null | true | null |
3914
| null |
Default
| null | null |
null |
{
"abstract": " Although the property of strong metric subregularity of set-valued mappings\nhas been present in the literature under various names and with various\ndefinitions for more than two decades, it has attracted much less attention\nthan its older \"siblings\", the metric regularity and the strong metric\nregularity. The purpose of this paper is to show that the strong metric\nsubregularity shares the main features of these two most popular regularity\nproperties and is not less instrumental in applications. We show that the\nstrong metric subregularity of a mapping F acting between metric spaces is\nstable under perturbations of the form f + F, where f is a function with a\nsmall calmness constant. This result is parallel to the Lyusternik-Graves\ntheorem for metric regularity and to the Robinson theorem for strong\nregularity, where the perturbations are represented by a function f with a\nsmall Lipschitz constant. Then we study perturbation stability of the same kind\nfor mappings acting between Banach spaces, where f is not necessarily\ndifferentiable but admits a set-valued derivative-like approximation. Strong\nmetric q-subregularity is also considered, where q is a positive real constant\nappearing as exponent in the definition. Rockafellar's criterion for strong\nmetric subregularity involving injectivity of the graphical derivative is\nextended to mappings acting in infinite-dimensional spaces. A sufficient\ncondition for strong metric subregularity is established in terms of\nsurjectivity of the Frechet coderivative. Various versions of Newton's method\nfor solving generalized equations are considered including inexact and\nsemismooth methods, for which superlinear convergence is shown under strong\nmetric subregularity.\n",
"title": "Strong Metric Subregularity of Mappings in Variational Analysis and Optimization"
}
| null | null | null | null | true | null |
3915
| null |
Default
| null | null |
null |
{
"abstract": " We present the SILVERRUSH program strategy and clustering properties\ninvestigated with $\\sim 2,000$ Ly$\\alpha$ emitters at $z=5.7$ and $6.6$ found\nin the early data of the Hyper Suprime-Cam (HSC) Subaru Strategic Program\nsurvey exploiting the carefully designed narrowband filters. We derive angular\ncorrelation functions with the unprecedentedly large samples of LAEs at $z=6-7$\nover the large total area of $14-21$ deg$^2$ corresponding to $0.3-0.5$\ncomoving Gpc$^2$. We obtain the average large-scale bias values of $b_{\\rm\navg}=4.1\\pm 0.2$ ($4.5\\pm 0.6$) at $z=5.7$ ($z=6.6$) for $\\gtrsim L^*$ LAEs,\nindicating the weak evolution of LAE clustering from $z=5.7$ to $6.6$. We\ncompare the LAE clustering results with two independent theoretical models that\nsuggest an increase of an LAE clustering signal by the patchy ionized bubbles\nat the epoch of reionization (EoR), and estimate the neutral hydrogen fraction\nto be $x_{\\rm HI}=0.15^{+0.15}_{-0.15}$ at $z=6.6$. Based on the halo\noccupation distribution models, we find that the $\\gtrsim L^*$ LAEs are hosted\nby the dark-matter halos with the average mass of $\\log (\\left < M_{\\rm h}\n\\right >/M_\\odot) =11.1^{+0.2}_{-0.4}$ ($10.8^{+0.3}_{-0.5}$) at $z=5.7$\n($6.6$) with a Ly$\\alpha$ duty cycle of 1 % or less, where the results of\n$z=6.6$ LAEs may be slightly biased, due to the increase of the clustering\nsignal at the EoR. Our clustering analysis reveals the low-mass nature of\n$\\gtrsim L^*$ LAEs at $z=6-7$, and that these LAEs probably evolve into massive\nsuper-$L^*$ galaxies in the present-day universe.\n",
"title": "Systematic Identification of LAEs for Visible Exploration and Reionization Research Using Subaru HSC (SILVERRUSH). I. Program Strategy and Clustering Properties of ~2,000 Lya Emitters at z=6-7 over the 0.3-0.5 Gpc$^2$ Survey Area"
}
| null | null | null | null | true | null |
3916
| null |
Default
| null | null |
null |
{
"abstract": " Magnetosphere at ion kinetic scales, or mini-magnetosphere, possesses unusual\nfeatures as predicted by numerical simulations. However, there are practically\nno data on the subject from space observations and the data which are available\nare far too incomplete. In the present work we describe results of laboratory\nexperiment on interaction of plasma flow with magnetic dipole with parameters\nsuch that ion inertia length is smaller than a size of observed magnetosphere.\nA detailed structure of non-coplanar or out-of-plane component of magnetic\nfield has been obtained in meridian plane. Independence of this component on\ndipole moment reversal, as was reported in previous work, has been verified. In\nthe tail distinct lobes and central current sheet have been observed. It was\nfound that lobe regions adjacent to boundary layer are dominated by\nnon-coplanar component of magnetic field. Tail-ward oriented electric current\nin plasma associated with that component appears to be equal to ion current in\nthe frontal part of magnetosphere and in the tail current sheet implying that\nelectrons are stationary in those regions while ions flow by. Obtained data\nstrongly support the proposed model of mini-magnetosphere based on two-fluid\neffects as described by the Hall term.\n",
"title": "Experimental study of mini-magnetosphere"
}
| null | null | null | null | true | null |
3917
| null |
Default
| null | null |
null |
{
"abstract": " We study soliton solutions of matrix Kadomtsev-Petviashvili (KP) equations in\na tropical limit, in which their support at fixed time is a planar graph and\npolarizations are attached to its constituting lines. There is a subclass of\n\"pure line soliton solutions\" for which we find that, in this limit, the\ndistribution of polarizations is fully determined by a Yang-Baxter map. For a\nvector KP equation, this map is given by an R-matrix, whereas it is a\nnon-linear map in case of a more general matrix KP equation. We also consider\nthe corresponding Korteweg-deVries (KdV) reduction. Furthermore, exploiting the\nfine structure of soliton interactions in the tropical limit, we obtain a new\nsolution of the tetrahedron (or Zamolodchikov) equation. Moreover, a solution\nof the functional tetrahedron equation arises from the parameter-dependence of\nthe vector KP R-matrix.\n",
"title": "Matrix KP: tropical limit and Yang-Baxter maps"
}
| null | null | null | null | true | null |
3918
| null |
Default
| null | null |
null |
{
"abstract": " Low-Power Wide-Area Networks (LPWANs) are being successfully used for the\nmonitoring of large-scale systems that are delay-tolerant and which have\nlow-bandwidth requirements. The next step would be instrumenting these for the\ncontrol of Cyber-Physical Systems (CPSs) distributed over large areas which\nrequire more bandwidth, bounded delays and higher reliability or at least more\nrigorous guarantees therein. This paper presents LPWA-MAC, a novel Low Power\nWide-Area network MAC protocol, that ensures bounded end-to-end delays, high\nchannel utility and supports many of the different traffic patterns and\ndata-rates typical of CPS.\n",
"title": "Poster Abstract: LPWA-MAC - a Low Power Wide Area network MAC protocol for cyber-physical systems"
}
| null | null |
[
"Computer Science"
] | null | true | null |
3919
| null |
Validated
| null | null |
null |
{
"abstract": " A compacted tree is a graph created from a binary tree such that repeatedly\noccurring subtrees in the original tree are represented by pointers to existing\nones, and hence every subtree is unique. Such representations form a special\nclass of directed acyclic graphs. We are interested in the asymptotic number of\ncompacted trees of given size, where the size of a compacted tree is given by\nthe number of its internal nodes. Due to its superexponential growth this\nproblem poses many difficulties. Therefore we restrict our investigations to\ncompacted trees of bounded right height, which is the maximal number of edges\ngoing to the right on any path from the root to a leaf.\nWe solve the asymptotic counting problem for this class as well as a closely\nrelated, further simplified class.\nFor this purpose, we develop a calculus on exponential generating functions\nfor compacted trees of bounded right height and for relaxed trees of bounded\nright height, which differ from compacted trees by dropping the above described\nuniqueness condition. This enables us to derive a recursively defined sequence\nof differential equations for the exponential generating functions. The\ncoefficients can then be determined by performing a singularity analysis of the\nsolutions of these differential equations.\nOur main results are the computation of the asymptotic numbers of relaxed as\nwell as compacted trees of bounded right height and given size, when the size\ntends to infinity.\n",
"title": "Asymptotic Enumeration of Compacted Binary Trees"
}
| null | null | null | null | true | null |
3920
| null |
Default
| null | null |
null |
{
"abstract": " The PiXeL detector (PXL) for the Heavy Flavor Tracker (HFT) of the STAR\nexperiment at RHIC is the first application of the state-of-the-art thin\nMonolithic Active Pixel Sensors (MAPS) technology in a collider environment.\nCustom built pixel sensors, their readout electronics and the detector\nmechanical structure are described in detail. Selected detector design aspects\nand production steps are presented. The detector operations during the three\nyears of data taking (2014-2016) and the overall performance exceeding the\ndesign specifications are discussed in the conclusive sections of this paper.\n",
"title": "The STAR MAPS-based PiXeL detector"
}
| null | null |
[
"Physics"
] | null | true | null |
3921
| null |
Validated
| null | null |
null |
{
"abstract": " Let $C({\\bf n})$ be a complete intersection monomial curve in the\n4-dimensional affine space. In this paper we study the complete intersection\nproperty of the monomial curve $C({\\bf n}+w{\\bf v})$, where $w>0$ is an integer\nand ${\\bf v} \\in \\mathbb{N}^{4}$. Also we investigate the Cohen-Macaulayness of\nthe tangent cone of $C({\\bf n}+w{\\bf v})$.\n",
"title": "Complete intersection monomial curves and the Cohen-Macaulayness of their tangent cones"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3922
| null |
Validated
| null | null |
null |
{
"abstract": " The dynamics of a quantum vortex torus knot ${\\cal T}_{P,Q}$ and similar\nknots in an atomic Bose-Einstein condensate at zero temperature in the\nThomas-Fermi regime has been considered in the hydrodynamic approximation. The\ncondensate has a spatially nonuniform equilibrium density profile $\\rho(z,r)$\ndue to an external axisymmetric potential. It is assumed that $z_*=0$, $r_*=1$\nis a maximum point for function $r\\rho(z,r)$, with $\\delta\n(r\\rho)\\approx-(\\alpha-\\epsilon) z^2/2 -(\\alpha+\\epsilon) (\\delta r)^2/2$ at\nsmall $z$ and $\\delta r$. Configuration of knot in the cylindrical coordinates\nis specified by a complex $2\\pi P$-periodic function\n$A(\\varphi,t)=Z(\\varphi,t)+i [R(\\varphi,t)-1]$. In the case $|A|\\ll 1$ the\nsystem is described by relatively simple approximate equations for re-scaled\nfunctions $W_n(\\varphi)\\propto A(2\\pi n+\\varphi)$, where $n=0,\\dots,P-1$, and\n$iW_{n,t}=-(W_{n,\\varphi\\varphi}+\\alpha W_n -\\epsilon W_n^*)/2-\\sum_{j\\neq\nn}1/(W_n^*-W_j^*)$. At $\\epsilon=0$, numerical examples of stable solutions as\n$W_n=\\theta_n(\\varphi-\\gamma t)\\exp(-i\\omega t)$ with non-trivial topology have\nbeen found for $P=3$. Besides that, dynamics of various non-stationary knots\nwith $P=3$ was simulated, and in some cases a tendency towards a finite-time\nsingularity has been detected. For $P=2$ at small $\\epsilon\\neq 0$, rotating\naround $z$ axis configurations of the form $(W_0-W_1)\\approx\nB_0\\exp(i\\zeta)+\\epsilon C(B_0,\\alpha)\\exp(-i\\zeta) + \\epsilon\nD(B_0,\\alpha)\\exp(3i\\zeta)$ have been investigated, where $B_0>0$ is an\narbitrary constant, $\\zeta=k_0\\varphi -\\Omega_0 t+\\zeta_0$, $k_0=Q/2$,\n$\\Omega_0=(k_0^2-\\alpha)/2-2/B_0^2$. In the parameter space $(\\alpha, B_0)$,\nwide stability regions for such solutions have been found. In unstable bands, a\nrecurrence of the vortex knot to a weakly excited state has been noted to be\npossible.\n",
"title": "Stable and unstable vortex knots in a trapped Bose-Einstein condensate"
}
| null | null | null | null | true | null |
3923
| null |
Default
| null | null |
null |
{
"abstract": " Quantum-dot cellular automata (QCA) is a likely candidate for future low\npower nano-scale electronic devices. Sequential circuits in QCA attract more\nattention due to its numerous application in digital industry. On the other\nhand, configurable devices provide low device cost and efficient utilization of\ndevice area. Since the fundamental building block of any sequential logic\ncircuit is flip flop, hence constructing configurable, multi-purpose QCA\nflip-flops are one of the prime importance of current research. This work\nproposes a design of configurable flip-flop (CFF) which is the first of its\nkind in QCA domain. The proposed flip-flop can be configured to D, T and JK\nflip-flop by configuring its control inputs. In addition, to make more\nefficient configurable flip-flop, a clock pulse generator (CPG) is designed\nwhich can trigger all types of edges (falling, rising and dual) of a clock. The\nsame CFF design is used to realize an edge configurable (dual/rising/falling)\nflip- flop with the help of CPG. The biggest advantage of using edge\nconfigurable (dual/rising/falling) flip-flop is that it can be used in 9\ndifferent ways using the same single circuit. All the proposed designs are\nverified using QCADesigner simulator.\n",
"title": "Design of Configurable Sequential Circuits in Quantum-dot Cellular Automata"
}
| null | null | null | null | true | null |
3924
| null |
Default
| null | null |
null |
{
"abstract": " Two major momentum-based techniques that have achieved tremendous success in\noptimization are Polyak's heavy ball method and Nesterov's accelerated\ngradient. A crucial step in all momentum-based methods is the choice of the\nmomentum parameter $m$ which is always suggested to be set to less than $1$.\nAlthough the choice of $m < 1$ is justified only under very strong theoretical\nassumptions, it works well in practice even when the assumptions do not\nnecessarily hold. In this paper, we propose a new momentum based method\n$\\textit{ADINE}$, which relaxes the constraint of $m < 1$ and allows the\nlearning algorithm to use adaptive higher momentum. We motivate our hypothesis\non $m$ by experimentally verifying that a higher momentum ($\\ge 1$) can help\nescape saddles much faster. Using this motivation, we propose our method\n$\\textit{ADINE}$ that helps weigh the previous updates more (by setting the\nmomentum parameter $> 1$), evaluate our proposed algorithm on deep neural\nnetworks and show that $\\textit{ADINE}$ helps the learning algorithm to\nconverge much faster without compromising on the generalization error.\n",
"title": "ADINE: An Adaptive Momentum Method for Stochastic Gradient Descent"
}
| null | null | null | null | true | null |
3925
| null |
Default
| null | null |
null |
{
"abstract": " We give characterizations of a finite group $G$ acting symplectically on a\nrational surface ($\\mathbb{C}P^2$ blown up at two or more points). In\nparticular, we obtain a symplectic version of the dichotomy of $G$-conic\nbundles versus $G$-del Pezzo surfaces for the corresponding $G$-rational\nsurfaces, analogous to a classical result in algebraic geometry. Besides the\ncharacterizations of the group $G$ (which is completely determined for the case\nof $\\mathbb{C}P^2\\# N\\overline{\\mathbb{C}P^2}$, $N=2,3,4$), we also investigate\nthe equivariant symplectic minimality and equivariant symplectic cone of a\ngiven $G$-rational surface.\n",
"title": "Symplectic rational $G$-surfaces and equivariant symplectic cones"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3926
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper, the biderivations without the skew-symmetric condition of the\ntwisted Heisenberg-Virasoro algebra are presented. We find some non-inner and\nnon-skew-symmetric biderivations. As applications, the characterizations of the\nforms of linear commuting maps and the commutative post-Lie algebra structures\non the twisted Heisenberg-Virasoro algebra are given. It also is proved that\nevery biderivation of the graded twisted Heisenberg-Virasoro left-symmetric\nalgebra is trivial.\n",
"title": "Biderivations of the twisted Heisenberg-Virasoro algebra and their applications"
}
| null | null | null | null | true | null |
3927
| null |
Default
| null | null |
null |
{
"abstract": " When the residents of Flint learned that lead had contaminated their water\nsystem, the local government made water-testing kits available to them free of\ncharge. The city government published the results of these tests, creating a\nvaluable dataset that is key to understanding the causes and extent of the lead\ncontamination event in Flint. This is the nation's largest dataset on lead in a\nmunicipal water system.\nIn this paper, we predict the lead contamination for each household's water\nsupply, and we study several related aspects of Flint's water troubles, many of\nwhich generalize well beyond this one city. For example, we show that elevated\nlead risks can be (weakly) predicted from observable home attributes. Then we\nexplore the factors associated with elevated lead. These risk assessments were\ndeveloped in part via a crowd sourced prediction challenge at the University of\nMichigan. To inform Flint residents of these assessments, they have been\nincorporated into a web and mobile application funded by \\texttt{Google.org}.\nWe also explore questions of self-selection in the residential testing program,\nexamining which factors are linked to when and how frequently residents\nvoluntarily sample their water.\n",
"title": "A Data Science Approach to Understanding Residential Water Contamination in Flint"
}
| null | null | null | null | true | null |
3928
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we introduce a stochastic projected subgradient method for\nweakly convex (i.e., uniformly prox-regular) nonsmooth, nonconvex functions---a\nwide class of functions which includes the additive and convex composite\nclasses. At a high-level, the method is an inexact proximal point iteration in\nwhich the strongly convex proximal subproblems are quickly solved with a\nspecialized stochastic projected subgradient method. The primary contribution\nof this paper is a simple proof that the proposed algorithm converges at the\nsame rate as the stochastic gradient method for smooth nonconvex problems. This\nresult appears to be the first convergence rate analysis of a stochastic (or\neven deterministic) subgradient method for the class of weakly convex\nfunctions.\n",
"title": "Proximally Guided Stochastic Subgradient Method for Nonsmooth, Nonconvex Problems"
}
| null | null | null | null | true | null |
3929
| null |
Default
| null | null |
null |
{
"abstract": " We compute the integral of a function or the expectation of a random variable\nwith minimal cost and use, for our new algorithm and for upper bounds of the\ncomplexity, i.i.d. samples. Under certain assumptions it is possible to select\na sample size based on a variance estimation, or -- more generally -- based on\nan estimation of a (central absolute) $p$-moment. That way one can guarantee a\nsmall absolute error with high probability, the problem is thus called\nsolvable. The expected cost of the method depends on the $p$-moment of the\nrandom variable, which can be arbitrarily large.\nIn order to prove the optimality of our algorithm we also provide lower\nbounds. These bounds apply not only to methods based on i.i.d. samples but also\nto general randomized algorithms. They show that -- up to constants -- the cost\nof the algorithm is optimal in terms of accuracy, confidence level, and norm of\nthe particular input random variable. Since the considered classes of random\nvariables or integrands are very large, the worst case cost would be infinite.\nNevertheless one can define adaptive stopping rules such that for each input\nthe expected cost is finite.\nWe contrast these positive results with examples of integration problems that\nare not solvable.\n",
"title": "Solvable Integration Problems and Optimal Sample Size Selection"
}
| null | null | null | null | true | null |
3930
| null |
Default
| null | null |
null |
{
"abstract": " We address the problem of \\emph{instance label stability} in multiple\ninstance learning (MIL) classifiers. These classifiers are trained only on\nglobally annotated images (bags), but often can provide fine-grained\nannotations for image pixels or patches (instances). This is interesting for\ncomputer aided diagnosis (CAD) and other medical image analysis tasks for which\nonly a coarse labeling is provided. Unfortunately, the instance labels may be\nunstable. This means that a slight change in training data could potentially\nlead to abnormalities being detected in different parts of the image, which is\nundesirable from a CAD point of view. Despite MIL gaining popularity in the CAD\nliterature, this issue has not yet been addressed. We investigate the stability\nof instance labels provided by several MIL classifiers on 5 different datasets,\nof which 3 are medical image datasets (breast histopathology, diabetic\nretinopathy and computed tomography lung images). We propose an unsupervised\nmeasure to evaluate instance stability, and demonstrate that a\nperformance-stability trade-off can be made when comparing MIL classifiers.\n",
"title": "Label Stability in Multiple Instance Learning"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
3931
| null |
Validated
| null | null |
null |
{
"abstract": " Computational procedures to foresee the 3D structure of aptamers are in\ncontinuous progress. They constitute a crucial input to research, mainly when\nthe crystallographic counterpart of the structures in silico produced is not\npresent. At now, many codes are able to perform structure and binding\nprediction, although their ability in scoring the results remains rather weak.\nIn this paper, we propose a novel procedure to complement the ranking outcomes\nof free docking code, by applying it to a set of anti-angiopoietin aptamers,\nwhose performances are known. We rank the in silico produced configurations,\nadopting a maximum likelihood estimate, based on their topological and\nelectrical properties. From the analysis, two principal kinds of conformers are\nidentified, whose ability to mimick the binding features of the natural\nreceptor is discussed. The procedure is easily generalizable to many biological\nbiomolecules, useful for increasing chances of success in designing\nhigh-specificity biosensors (aptasensors).\n",
"title": "New indicators for assessing the quality of in silico produced biomolecules: the case study of the aptamer-Angiopoietin-2 complex"
}
| null | null | null | null | true | null |
3932
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we propose a finite element method for solving elliptic\nequations with the observational Dirichlet boundary data which may subject to\nrandom noises. The method is based on the weak formulation of Lagrangian\nmultiplier. We show the convergence of the random finite element error in\nexpectation and, when the noise is sub-Gaussian, in the Orlicz 2- norm which\nimplies the probability that the finite element error estimates are violated\ndecays exponentially. Numerical examples are included.\n",
"title": "A finite element method for elliptic problems with observational boundary data"
}
| null | null | null | null | true | null |
3933
| null |
Default
| null | null |
null |
{
"abstract": " This paper introduces a new probabilistic architecture called Sum-Product\nGraphical Model (SPGM). SPGMs combine traits from Sum-Product Networks (SPNs)\nand Graphical Models (GMs): Like SPNs, SPGMs always enable tractable inference\nusing a class of models that incorporate context specific independence. Like\nGMs, SPGMs provide a high-level model interpretation in terms of conditional\nindependence assumptions and corresponding factorizations. Thus, the new\narchitecture represents a class of probability distributions that combines, for\nthe first time, the semantics of graphical models with the evaluation\nefficiency of SPNs. We also propose a novel algorithm for learning both the\nstructure and the parameters of SPGMs. A comparative empirical evaluation\ndemonstrates competitive performances of our approach in density estimation.\n",
"title": "Sum-Product Graphical Models"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
3934
| null |
Validated
| null | null |
null |
{
"abstract": " The energetic particle environment on the Martian surface is influenced by\nsolar and heliospheric modulation and changes in the local atmospheric pressure\n(or column depth). The Radiation Assessment Detector (RAD) on board the Mars\nScience Laboratory rover Curiosity on the surface of Mars has been measuring\nthis effect for over four Earth years (about two Martian years). The\nanticorrelation between the recorded surface Galactic Cosmic Ray-induced dose\nrates and pressure changes has been investigated by Rafkin et al. (2014) and\nthe long-term solar modulation has also been empirically analyzed and modeled\nby Guo et al. (2015). This paper employs the newly updated HZETRN2015 code to\nmodel the Martian atmospheric shielding effect on the accumulated dose rates\nand the change of this effect under different solar modulation and atmospheric\nconditions. The modeled results are compared with the most up-to-date (from 14\nAugust 2012 to 29 June 2016) observations of the RAD instrument on the surface\nof Mars. Both model and measurements agree reasonably well and show the\natmospheric shielding effect under weak solar modulation conditions and the\ndecline of this effect as solar modulation becomes stronger. This result is\nimportant for better risk estimations of future human explorations to Mars\nunder different heliospheric and Martian atmospheric conditions.\n",
"title": "Dependence of the Martian radiation environment on atmospheric depth: Modeling and measurement"
}
| null | null | null | null | true | null |
3935
| null |
Default
| null | null |
null |
{
"abstract": " The growth in variety and volume of OLTP (Online Transaction Processing)\napplications poses a challenge to OLTP systems to meet performance and cost\ndemands in the existing hardware landscape. These applications are highly\ninteractive (latency sensitive) and require update consistency. They target\ncommodity hardware for deployment and demand scalability in throughput with\nincreasing clients and data. Currently, OLTP systems used by these applications\nprovide trade-offs in performance and ease of development over a variety of\napplications. In order to bridge the gap between performance and ease of\ndevelopment, we propose an intuitive, high-level programming model which allows\nOLTP applications to be modeled as a cluster of application logic units. By\nextending transactions guaranteeing full ACID semantics to provide the proposed\nmodel, we maintain ease of application development. The model allows the\napplication developer to reason about program performance, and to influence it\nwithout the involvement of OLTP system designers (database designers) and/or\nDBAs. As a result, the database designer is free to focus on efficient running\nof programs to ensure optimal cluster resource utilization.\n",
"title": "Transactional Partitioning: A New Abstraction for Main-Memory Databases"
}
| null | null | null | null | true | null |
3936
| null |
Default
| null | null |
null |
{
"abstract": " Until recently, social media was seen to promote democratic discourse on\nsocial and political issues. However, this powerful communication platform has\ncome under scrutiny for allowing hostile actors to exploit online discussions\nin an attempt to manipulate public opinion. A case in point is the ongoing U.S.\nCongress' investigation of Russian interference in the 2016 U.S. election\ncampaign, with Russia accused of using trolls (malicious accounts created to\nmanipulate) and bots to spread misinformation and politically biased\ninformation. In this study, we explore the effects of this manipulation\ncampaign, taking a closer look at users who re-shared the posts produced on\nTwitter by the Russian troll accounts publicly disclosed by U.S. Congress\ninvestigation. We collected a dataset with over 43 million election-related\nposts shared on Twitter between September 16 and October 21, 2016, by about 5.7\nmillion distinct users. This dataset included accounts associated with the\nidentified Russian trolls. We use label propagation to infer the ideology of\nall users based on the news sources they shared. This method enables us to\nclassify a large number of users as liberal or conservative with precision and\nrecall above 90%. Conservatives retweeted Russian trolls about 31 times more\noften than liberals and produced 36x more tweets. Additionally, most retweets\nof troll content originated from two Southern states: Tennessee and Texas.\nUsing state-of-the-art bot detection techniques, we estimated that about 4.9%\nand 6.2% of liberal and conservative users respectively were bots. Text\nanalysis on the content shared by trolls reveals that they had a mostly\nconservative, pro-Trump agenda. Although an ideologically broad swath of\nTwitter users was exposed to Russian Trolls in the period leading up to the\n2016 U.S. Presidential election, it was mainly conservatives who helped amplify\ntheir message.\n",
"title": "Analyzing the Digital Traces of Political Manipulation: The 2016 Russian Interference Twitter Campaign"
}
| null | null | null | null | true | null |
3937
| null |
Default
| null | null |
null |
{
"abstract": " The local event detection is to use posting messages with geotags on social\nnetworks to reveal the related ongoing events and their locations. Recent\nstudies have demonstrated that the geo-tagged tweet stream serves as an\nunprecedentedly valuable source for local event detection. Nevertheless, how to\neffectively extract local events from large geo-tagged tweet streams in real\ntime remains challenging. A robust and efficient cloud-based real-time local\nevent detection software system would benefit various aspects in the real-life\nsociety, from shopping recommendation for customer service providers to\ndisaster alarming for emergency departments. We use the preliminary research\nGeoBurst as a starting point, which proposed a novel method to detect local\nevents. GeoBurst+ leverages a novel cross-modal authority measure to identify\nseveral pivots in the query window. Such pivots reveal different geo-topical\nactivities and naturally attract related tweets to form candidate events. It\nfurther summarises the continuous stream and compares the candidates against\nthe historical summaries to pinpoint truly interesting local events. We mainly\nimplement a website demonstration system Event-Radar with an improved algorithm\nto show the real-time local events online for public interests. Better still,\nas the query window shifts, our method can update the event list with little\ntime cost, thus achieving continuous monitoring of the stream.\n",
"title": "Event-Radar: Real-time Local Event Detection System for Geo-Tagged Tweet Streams"
}
| null | null | null | null | true | null |
3938
| null |
Default
| null | null |
null |
{
"abstract": " We prove that, given a closure function the smallest preimage of a closed set\ncan be calculated in polynomial time in the number of closed sets. This\nconfirms a conjecture of Albenque and Knauer and implies that there is a\npolynomial time algorithm to compute the convex hull-number of a graph, when\nall its convex subgraphs are given as input. We then show that computing if the\nsmallest preimage of a closed set is logarithmic in the size of the ground set\nis LOGSNP-complete if only the ground set is given. A special instance of this\nproblem is computing the dimension of a poset given its linear extension graph,\nthat was conjectured to be in P.\nThe intent to show that the latter problem is LOGSNP-complete leads to\nseveral interesting questions and to the definition of the isometric hull,\ni.e., a smallest isometric subgraph containing a given set of vertices $S$.\nWhile for $|S|=2$ an isometric hull is just a shortest path, we show that\ncomputing the isometric hull of a set of vertices is NP-complete even if\n$|S|=3$. Finally, we consider the problem of computing the isometric\nhull-number of a graph and show that computing it is $\\Sigma^P_2$ complete.\n",
"title": "Computing metric hulls in graphs"
}
| null | null | null | null | true | null |
3939
| null |
Default
| null | null |
null |
{
"abstract": " Evolutionary games on graphs describe how strategic interactions and\npopulation structure determine evolutionary success, quantified by the\nprobability that a single mutant takes over a population. Graph structures,\ncompared to the well-mixed case, can act as amplifiers or suppressors of\nselection by increasing or decreasing the fixation probability of a beneficial\nmutant. Properties of the associated mean fixation times can be more intricate,\nespecially when selection is strong. The intuition is that fixation of a\nbeneficial mutant happens fast (in a dominance game), that fixation takes very\nlong (in a coexistence game), and that strong selection eliminates demographic\nnoise. Here we show that these intuitions can be misleading in structured\npopulations. We analyze mean fixation times on the cycle graph under strong\nfrequency-dependent selection for two different microscopic evolutionary update\nrules (death-birth and birth-death). We establish exact analytical results for\nfixation times under strong selection, and show that there are coexistence\ngames in which fixation occurs in time polynomial in population size. Depending\non the underlying game, we observe inherence of demographic noise even under\nstrong selection, if the process is driven by random death before selection for\nbirth of an offspring (death-birth update). In contrast, if selection for an\noffspring occurs before random removal (birth-death update), strong selection\ncan remove demographic noise almost entirely.\n",
"title": "Evolutionary games on cycles with strong selection"
}
| null | null | null | null | true | null |
3940
| null |
Default
| null | null |
null |
{
"abstract": " We focus on the analysis of planar shapes and solid objects having thin\nfeatures and propose a new mathematical model to characterize them. Based on\nour model, that we call an epsilon-shape, we show how thin parts can be\neffectively and efficiently detected by an algorithm, and propose a novel\napproach to thicken these features while leaving all the other parts of the\nshape unchanged. When compared with state-of-the-art solutions, our proposal\nproves to be particularly flexible, efficient and stable, and does not require\nany unintuitive parameter to fine-tune the process. Furthermore, our method is\nable to detect thin features both in the object and in its complement, thus\nproviding a useful tool to detect thin cavities and narrow channels. We discuss\nthe importance of this kind of analysis in the design of robust structures and\nin the creation of geometry to be fabricated with modern additive manufacturing\ntechnology.\n",
"title": "Epsilon-shapes: characterizing, detecting and thickening thin features in geometric models"
}
| null | null |
[
"Computer Science"
] | null | true | null |
3941
| null |
Validated
| null | null |
null |
{
"abstract": " The learning of domain-invariant representations in the context of domain\nadaptation with neural networks is considered. We propose a new regularization\nmethod that minimizes the discrepancy between domain-specific latent feature\nrepresentations directly in the hidden activation space. Although some standard\ndistribution matching approaches exist that can be interpreted as the matching\nof weighted sums of moments, e.g. Maximum Mean Discrepancy (MMD), an explicit\norder-wise matching of higher order moments has not been considered before. We\npropose to match the higher order central moments of probability distributions\nby means of order-wise moment differences. Our model does not require\ncomputationally expensive distance and kernel matrix computations. We utilize\nthe equivalent representation of probability distributions by moment sequences\nto define a new distance function, called Central Moment Discrepancy (CMD). We\nprove that CMD is a metric on the set of probability distributions on a compact\ninterval. We further prove that convergence of probability distributions on\ncompact intervals w.r.t. the new metric implies convergence in distribution of\nthe respective random variables. We test our approach on two different\nbenchmark data sets for object recognition (Office) and sentiment analysis of\nproduct reviews (Amazon reviews). CMD achieves a new state-of-the-art\nperformance on most domain adaptation tasks of Office and outperforms networks\ntrained with MMD, Variational Fair Autoencoders and Domain Adversarial Neural\nNetworks on Amazon reviews. In addition, a post-hoc parameter sensitivity\nanalysis shows that the new approach is stable w.r.t. parameter changes in a\ncertain interval. The source code of the experiments is publicly available.\n",
"title": "Central Moment Discrepancy (CMD) for Domain-Invariant Representation Learning"
}
| null | null | null | null | true | null |
3942
| null |
Default
| null | null |
null |
{
"abstract": " Encoder-decoder networks using convolutional neural network (CNN)\narchitecture have been extensively used in deep learning literatures thanks to\nits excellent performance for various inverse problems in computer vision,\nmedical imaging, etc. However, it is still difficult to obtain coherent\ngeometric view why such an architecture gives the desired performance. Inspired\nby recent theoretical understanding on generalizability, expressivity and\noptimization landscape of neural networks, as well as the theory of\nconvolutional framelets, here we provide a unified theoretical framework that\nleads to a better understanding of geometry of encoder-decoder CNNs. Our\nunified mathematical framework shows that encoder-decoder CNN architecture is\nclosely related to nonlinear basis representation using combinatorial\nconvolution frames, whose expressibility increases exponentially with the\nnetwork depth. We also demonstrate the importance of skipped connection in\nterms of expressibility, and optimization landscape.\n",
"title": "Understanding Geometry of Encoder-Decoder CNNs"
}
| null | null | null | null | true | null |
3943
| null |
Default
| null | null |
null |
{
"abstract": " ShuffleNet is a state-of-the-art light weight convolutional neural network\narchitecture. Its basic operations include group, channel-wise convolution and\nchannel shuffling. However, channel shuffling is manually designed empirically.\nMathematically, shuffling is a multiplication by a permutation matrix. In this\npaper, we propose to automate channel shuffling by learning permutation\nmatrices in network training. We introduce an exact Lipschitz continuous\nnon-convex penalty so that it can be incorporated in the stochastic gradient\ndescent to approximate permutation at high precision. Exact permutations are\nobtained by simple rounding at the end of training and are used in inference.\nThe resulting network, referred to as AutoShuffleNet, achieved improved\nclassification accuracies on CIFAR-10 and ImageNet data sets. In addition, we\nfound experimentally that the standard convex relaxation of permutation\nmatrices into stochastic matrices leads to poor performance. We prove\ntheoretically the exactness (error bounds) in recovering permutation matrices\nwhen our penalty function is zero (very small). We present examples of\npermutation optimization through graph matching and two-layer neural network\nmodels where the loss functions are calculated in closed analytical form. In\nthe examples, convex relaxation failed to capture permutations whereas our\npenalty succeeded.\n",
"title": "AutoShuffleNet: Learning Permutation Matrices via an Exact Lipschitz Continuous Penalty in Deep Convolutional Neural Networks"
}
| null | null | null | null | true | null |
3944
| null |
Default
| null | null |
null |
{
"abstract": " Once a failure is observed, the primary concern of the developer is to\nidentify what caused it in order to repair the code that induced the incorrect\nbehavior. Until a permanent repair is afforded, code repair patches are\ninvaluable. The aim of this work is to devise an automated patch generation\ntechnique that proceeds as follows: Step1) It identifies a set of\nfailure-causing control dependence chains that are minimal in terms of number\nand length. Step2) It identifies a set of predicates within the chains along\nwith associated execution instances, such that negating the predicates at the\ngiven instances would exhibit correct behavior. Step3) For each candidate\npredicate, it creates a classifier that dictates when the predicate should be\nnegated to yield correct program behavior. Step4) Prior to each candidate\npredicate, the faulty program is injected with a call to its corresponding\nclassifier passing it the program state and getting a return value predictively\nindicating whether to negate the predicate or not. The role of the classifiers\nis to ensure that: 1) the predicates are not negated during passing runs; and\n2) the predicates are negated at the appropriate instances within failing runs.\nWe implemented our patch generation approach for the Java platform and\nevaluated our toolset using 148 defects from the Introclass and Siemens\nbenchmarks. The toolset identified 56 full patches and another 46 partial\npatches, and the classification accuracy averaged 84%.\n",
"title": "ACDC: Altering Control Dependence Chains for Automated Patch Generation"
}
| null | null | null | null | true | null |
3945
| null |
Default
| null | null |
null |
{
"abstract": " This volume contains a final and revised selection of papers presented at the\nEighth Workshop on Intersection Types and Related Systems (ITRS 2016), held on\nJune 26, 2016 in Porto, in affiliation with FSCD 2016.\n",
"title": "Proceedings Eighth Workshop on Intersection Types and Related Systems"
}
| null | null | null | null | true | null |
3946
| null |
Default
| null | null |
null |
{
"abstract": " Recent observations of lensed galaxies at cosmological distances have\ndetected individual stars that are extremely magnified when crossing the\ncaustics of lensing clusters. In idealized cluster lenses with smooth mass\ndistributions, two images of a star of radius $R$ approaching a caustic\nbrighten as $t^{-1/2}$ and reach a peak magnification $\\sim 10^{6}\\, (10\\,\nR_{\\odot}/R)^{1/2}$ before merging on the critical curve. We show that a mass\nfraction ($\\kappa_\\star \\gtrsim \\, 10^{-4.5}$) in microlenses inevitably\ndisrupts the smooth caustic into a network of corrugated microcaustics, and\nproduces light curves with numerous peaks. Using analytical calculations and\nnumerical simulations, we derive the characteristic width of the network,\ncaustic-crossing frequencies, and peak magnifications. For the lens parameters\nof a recent detection and a population of intracluster stars with $\\kappa_\\star\n\\sim 0.01$, we find a source-plane width of $\\sim 20 \\, {\\rm pc}$ for the\ncaustic network, which spans $0.2 \\, {\\rm arcsec}$ on the image plane. A source\nstar takes $\\sim 2\\times 10^4$ years to cross this width, with a total of $\\sim\n6 \\times 10^4$ crossings, each one lasting for $\\sim 5\\,{\\rm\nhr}\\,(R/10\\,R_\\odot)$ with typical peak magnifications of $\\sim 10^{4} \\left(\nR/ 10\\,R_\\odot \\right)^{-1/2}$. The exquisite sensitivity of caustic-crossing\nevents to the granularity of the lens-mass distribution makes them ideal probes\nof dark matter components, such as compact halo objects and ultralight axion\ndark matter.\n",
"title": "Microlensing of Extremely Magnified Stars near Caustics of Galaxy Clusters"
}
| null | null |
[
"Physics"
] | null | true | null |
3947
| null |
Validated
| null | null |
null |
{
"abstract": " The graph Laplacian is a standard tool in data science, machine learning, and\nimage processing. The corresponding matrix inherits the complex structure of\nthe underlying network and is in certain applications densely populated. This\nmakes computations, in particular matrix-vector products, with the graph\nLaplacian a hard task. A typical application is the computation of a number of\nits eigenvalues and eigenvectors. Standard methods become infeasible as the\nnumber of nodes in the graph is too large. We propose the use of the fast\nsummation based on the nonequispaced fast Fourier transform (NFFT) to perform\nthe dense matrix-vector product with the graph Laplacian fast without ever\nforming the whole matrix. The enormous flexibility of the NFFT algorithm allows\nus to embed the accelerated multiplication into Lanczos-based eigenvalues\nroutines or iterative linear system solvers and even consider other than the\nstandard Gaussian kernels. We illustrate the feasibility of our approach on a\nnumber of test problems from image segmentation to semi-supervised learning\nbased on graph-based PDEs. In particular, we compare our approach with the\nNyström method. Moreover, we present and test an enhanced, hybrid version of\nthe Nyström method, which internally uses the NFFT.\n",
"title": "NFFT meets Krylov methods: Fast matrix-vector products for the graph Laplacian of fully connected networks"
}
| null | null |
[
"Statistics"
] | null | true | null |
3948
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper we propose a modified version of the simulated annealing\nalgorithm for solving a stochastic global optimization problem. More precisely,\nwe address the problem of finding a global minimizer of a function with noisy\nevaluations. We provide a rate of convergence and its optimized parametrization\nto ensure a minimal number of evaluations for a given accuracy and a confidence\nlevel close to 1. This work is completed with a set of numerical\nexperimentations and assesses the practical performance both on benchmark test\ncases and on real world examples.\n",
"title": "Convergence rate of a simulated annealing algorithm with noisy observations"
}
| null | null | null | null | true | null |
3949
| null |
Default
| null | null |
null |
{
"abstract": " Major histocompatibility complex class two (MHC-II) molecules are\ntrans-membrane proteins and key components of the cellular immune system. Upon\nrecognition of foreign peptides expressed on the MHC-II binding groove, helper\nT cells mount an immune response against invading pathogens. Therefore,\nmechanistic identification and knowledge of physico-chemical features that\ngovern interactions between peptides and MHC-II molecules is useful for the\ndesign of effective epitope-based vaccines, as well as for understanding of\nimmune responses. In this paper, we present a comprehensive trans-allelic\nprediction model, a generalized version of our previous biophysical model, that\ncan predict peptide interactions for all three human MHC-II loci (HLA-DR,\nHLA-DP and HLA-DQ), using both peptide sequence data and structural information\nof MHC-II molecules. The advantage of this approach over other machine learning\nmodels is that it offers a simple and plausible physical explanation for\npeptide-MHC-II interactions. We train the model using a benchmark experimental\ndataset, and measure its predictive performance using novel data. Despite its\nrelative simplicity, we find that the model has comparable performance to the\nstate-of-the-art method. Focusing on the physical bases of peptide-MHC binding,\nwe find support for previous theoretical predictions about the contributions of\ncertain binding pockets to the binding energy. Additionally, we find that\nbinding pockets P 4 and P 5 of HLA-DP, which were not previously considered as\nprimary anchors, do make strong contributions to the binding energy. Together,\nthe results indicate that our model can serve as a useful complement to\nalternative approaches to predicting peptide-MHC interactions.\n",
"title": "Trans-allelic model for prediction of peptide:MHC-II interactions"
}
| null | null |
[
"Statistics"
] | null | true | null |
3950
| null |
Validated
| null | null |
null |
{
"abstract": " Speechreading is the task of inferring phonetic information from visually\nobserved articulatory facial movements, and is a notoriously difficult task for\nhumans to perform. In this paper we present an end-to-end model based on a\nconvolutional neural network (CNN) for generating an intelligible and\nnatural-sounding acoustic speech signal from silent video frames of a speaking\nperson. We train our model on speakers from the GRID and TCD-TIMIT datasets,\nand evaluate the quality and intelligibility of reconstructed speech using\ncommon objective measurements. We show that speech predictions from the\nproposed model attain scores which indicate significantly improved quality over\nexisting models. In addition, we show promising results towards reconstructing\nspeech from an unconstrained dictionary.\n",
"title": "Improved Speech Reconstruction from Silent Video"
}
| null | null | null | null | true | null |
3951
| null |
Default
| null | null |
null |
{
"abstract": " Fedotovite K$_2$Cu$_3$O(SO$_4$)$_3$ is a candidate of new quantum spin\nsystems, in which the edge-shared tetrahedral (EST) spin-clusters consisting of\nCu$^{2+}$ are connected by weak inter-cluster couplings to from one-dimensional\narray. Comprehensive experimental studies by magnetic susceptibility,\nmagnetization, heat capacity, and inelastic neutron scattering measurements\nreveal the presence of an effective $S$ = 1 Haldane state below $T \\cong 4$ K.\nRigorous theoretical studies provide an insight into the magnetic state of\nK$_2$Cu$_3$O(SO$_4$)$_3$: an EST cluster makes a triplet in the ground state\nand one-dimensional chain of the EST induces a cluster-based Haldane state. We\npredict that the cluster-based Haldene state emerges whenever the number of\ntetrahedra in the EST is $even$.\n",
"title": "Cluster-based Haldane state in edge-shared tetrahedral spin-cluster chain: Fedotovite K$_2$Cu$_3$O(SO$_4$)$_3$"
}
| null | null | null | null | true | null |
3952
| null |
Default
| null | null |
null |
{
"abstract": " We call a simple abelian variety over $\\mathbb{F}_p$ super-isolated if its\n($\\mathbb{F}_p$-rational) isogeny class contains no other varieties. The\nmotivation for considering these varieties comes from concerns about isogeny\nbased attacks on the discrete log problem. We heuristically estimate that the\nnumber of super-isolated elliptic curves over $\\mathbb{F}_p$ with prime order\nand $p \\leq N$, is roughly $\\tilde{\\Theta}(\\sqrt{N})$. In contrast, we prove\nthat there are only 2 super-isolated surfaces of cryptographic size and\nnear-prime order.\n",
"title": "Super-Isolated Elliptic Curves and Abelian Surfaces in Cryptography"
}
| null | null | null | null | true | null |
3953
| null |
Default
| null | null |
null |
{
"abstract": " Multivariate techniques based on engineered features have found wide adoption\nin the identification of jets resulting from hadronic top decays at the Large\nHadron Collider (LHC). Recent Deep Learning developments in this area include\nthe treatment of the calorimeter activation as an image or supplying a list of\njet constituent momenta to a fully connected network. This latter approach\nlends itself well to the use of Recurrent Neural Networks. In this work the\napplicability of architectures incorporating Long Short-Term Memory (LSTM)\nnetworks is explored. Several network architectures, methods of ordering of jet\nconstituents, and input pre-processing are studied. The best performing LSTM\nnetwork achieves a background rejection of 100 for 50% signal efficiency. This\nrepresents more than a factor of two improvement over a fully connected Deep\nNeural Network (DNN) trained on similar types of inputs.\n",
"title": "Long Short-Term Memory (LSTM) networks with jet constituents for boosted top tagging at the LHC"
}
| null | null | null | null | true | null |
3954
| null |
Default
| null | null |
null |
{
"abstract": " Tangent measure and blow-up methods, are powerful tools for understanding the\nrelationship between the infinitesimal structure of the boundary of a domain\nand the behavior of its harmonic measure. We introduce a method for studying\ntangent measures of elliptic measures in arbitrary domains associated with\n(possibly non-symmetric) elliptic operators in divergence form whose\ncoefficients have vanishing mean oscillation at the boundary. In this setting,\nwe show the following for domains $ \\Omega \\subset \\mathbb{R}^{n+1}$:\n1. We extend the results of Kenig, Preiss, and Toro [KPT09] by showing mutual\nabsolute continuity of interior and exterior elliptic measures for {\\it any}\ndomains implies the tangent measures are a.e. flat and the elliptic measures\nhave dimension $n$.\n2. We generalize the work of Kenig and Toro [KT06] and show that VMO\nequivalence of doubling interior and exterior elliptic measures for general\ndomains implies the tangent measures are always elliptic polynomials.\n3. In a uniform domain that satisfies the capacity density condition and\nwhose boundary is locally finite and has a.e. positive lower $n$-Hausdorff\ndensity, we show that if the elliptic measure is absolutely continuous with\nrespect to $n$-Hausdorff measure then the boundary is rectifiable. This\ngeneralizes the work of Akman, Badger, Hofmann, and Martell [ABHM17].\nFinally, we generalize one of the main results of [Bad11] by showing that if\n$\\omega$ is a Radon measure for which all tangent measures at a point are\nharmonic polynomials vanishing at the origin, then they are all homogeneous\nharmonic polynomials.\n",
"title": "Tangent measures of elliptic harmonic measure and applications"
}
| null | null | null | null | true | null |
3955
| null |
Default
| null | null |
null |
{
"abstract": " Programmers often write code which have similarity to existing code written\nsomewhere. A tool that could help programmers to search such similar code would\nbe immensely useful. Such a tool could help programmers to extend partially\nwritten code snippets to completely implement necessary functionality, help to\ndiscover extensions to the partial code which are commonly done by other\nprogrammers, help to cross-check against similar code written by other\nprogrammers, or help to add extra code which would avoid common mistakes and\nerrors. We propose Aroma, a tool and technique for code recommendation via\nstructural code search. Aroma indexes a huge code corpus including thousands of\nopen-source projects, takes a partial code snippet as input, searches the\nindexed method bodies which contain the partial code snippet, clusters and\nintersects the results of search to recommend a small set of succinct code\nsnippets which contain the query snippet and which appears as part of several\nprograms in the corpus. We evaluated Aroma on several randomly selected queries\ncreated from the corpus and as well as those derived from the code snippets\nobtained from Stack Overflow, a popular website for discussing code. We found\nthat Aroma was able to retrieve and recommend most relevant code snippets\nefficiently.\n",
"title": "Aroma: Code Recommendation via Structural Code Search"
}
| null | null | null | null | true | null |
3956
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we shall prove the equality \\[\n\\zeta(3,\\{2\\}^{n},1,2)=\\zeta(\\{2\\}^{n+3})+2\\zeta(3,3,\\{2\\}^{n}) \\] conjectured\nby Hoffman using certain identities among iterated integrals on\n$\\mathbb{P}^{1}\\setminus\\{0,1,\\infty,z\\}$.\n",
"title": "On Hoffman's conjectural identity"
}
| null | null | null | null | true | null |
3957
| null |
Default
| null | null |
null |
{
"abstract": " We demonstrate sub-picosecond wavelength conversion in the C-band via four\nwave mixing in a 45cm long high index doped silica spiral waveguide. We achieve\nan on/off conversion efficiency (signal to idler) of +16.5dB as well as a\nparametric gain of +15dB for a peak pump power of 38W over a wavelength range\nof 100nm. Furthermore, we demonstrated a minimum gain of +5dB over a wavelength\nrange as large as 200nm.\n",
"title": "Parametric gain and wavelength conversion via third order nonlinear optics a CMOS compatible waveguide"
}
| null | null |
[
"Physics"
] | null | true | null |
3958
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper, we develop a new accelerated stochastic gradient method for\nefficiently solving the convex regularized empirical risk minimization problem\nin mini-batch settings. The use of mini-batches is becoming a golden standard\nin the machine learning community, because mini-batch settings stabilize the\ngradient estimate and can easily make good use of parallel computing. The core\nof our proposed method is the incorporation of our new \"double acceleration\"\ntechnique and variance reduction technique. We theoretically analyze our\nproposed method and show that our method much improves the mini-batch\nefficiencies of previous accelerated stochastic methods, and essentially only\nneeds size $\\sqrt{n}$ mini-batches for achieving the optimal iteration\ncomplexities for both non-strongly and strongly convex objectives, where $n$ is\nthe training set size. Further, we show that even in non-mini-batch settings,\nour method achieves the best known convergence rate for both non-strongly and\nstrongly convex objectives.\n",
"title": "Doubly Accelerated Stochastic Variance Reduced Dual Averaging Method for Regularized Empirical Risk Minimization"
}
| null | null | null | null | true | null |
3959
| null |
Default
| null | null |
null |
{
"abstract": " MapReduce is a popular programming paradigm for developing large-scale,\ndata-intensive computation. Many frameworks that implement this paradigm have\nrecently been developed. To leverage these frameworks, however, developers must\nbecome familiar with their APIs and rewrite existing code. Casper is a new tool\nthat automatically translates sequential Java programs into the MapReduce\nparadigm. Casper identifies potential code fragments to rewrite and translates\nthem in two steps: (1) Casper uses program synthesis to search for a program\nsummary (i.e., a functional specification) of each code fragment. The summary\nis expressed using a high-level intermediate language resembling the MapReduce\nparadigm and verified to be semantically equivalent to the original using a\ntheorem prover. (2) Casper generates executable code from the summary, using\neither the Hadoop, Spark, or Flink API. We evaluated Casper by automatically\nconverting real-world, sequential Java benchmarks to MapReduce. The resulting\nbenchmarks perform up to 48.2x faster compared to the original.\n",
"title": "Automatically Leveraging MapReduce Frameworks for Data-Intensive Applications"
}
| null | null | null | null | true | null |
3960
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we propose a unified view of gradient-based algorithms for\nstochastic convex composite optimization. By extending the concept of estimate\nsequence introduced by Nesterov, we interpret a large class of stochastic\noptimization methods as procedures that iteratively minimize a surrogate of the\nobjective. This point of view covers stochastic gradient descent (SGD), the\nvariance-reduction approaches SAGA, SVRG, MISO, their proximal variants, and\nhas several advantages: (i) we provide a simple generic proof of convergence\nfor all of the aforementioned methods; (ii) we naturally obtain new algorithms\nwith the same guarantees; (iii) we derive generic strategies to make these\nalgorithms robust to stochastic noise, which is useful when data is corrupted\nby small random perturbations. Finally, we show that this viewpoint is useful\nto obtain accelerated algorithms.\n",
"title": "Estimate Sequences for Stochastic Composite Optimization: Variance Reduction, Acceleration, and Robustness to Noise"
}
| null | null | null | null | true | null |
3961
| null |
Default
| null | null |
null |
{
"abstract": " Over almost three decades the TAUP conference has seen a remarkable momentum\ngain in direct dark matter search. An important accelerator were first\nindications for a modulating signal rate in the DAMA/NaI experiment reported in\n1997. Today the presence of an annual modulation, which matches in period and\nphase the expectation for dark matter, is supported at > 9$\\sigma$ confidence.\nThe underlying nature of dark matter, however, is still considered an open and\nfundamental question of particle physics. No other direct dark matter search\ncould confirm the DAMA claim up to now; moreover, numerous null-results are in\nclear contradiction under so-called standard assumptions for the dark matter\nhalo and the interaction mechanism of dark with ordinary matter. As both bear a\ndependence on the target material, resolving this controversial situation will\nconvincingly only be possible with an experiment using sodium iodide (NaI) as\ntarget. COSINUS aims to even go a step further by combining NaI with a novel\ndetection approach. COSINUS aims to operate NaI as a cryogenic calorimeter\nreading scintillation light and phonon/heat signal. Two distinct advantages\narise from this approach, a substantially lower energy threshold for nuclear\nrecoils and particle identification on an event-by-event basis. These key\nbenefits will allow COSINUS to clarify a possible nuclear recoil origin of the\nDAMA signal with comparatively little exposure of O(100kg days) and, thereby,\nanswer a long-standing question of particle physics. Today COSINUS is in R&D\nphase; in this contribution we show results from the 2nd prototype, albeit the\nfirst one of the final foreseen detector design. The key finding of this\nmeasurement is that pure, undoped NaI is a truly excellent scintillator at low\ntemperatures: We measure 13.1% of the total deposited energy in the NaI crystal\nin the form of scintillation light (in the light detector).\n",
"title": "Results of the first NaI scintillating calorimeter prototypes by COSINUS"
}
| null | null | null | null | true | null |
3962
| null |
Default
| null | null |
null |
{
"abstract": " Diagnosis and risk stratification of cancer and many other diseases require\nthe detection of genomic breakpoints as a prerequisite of calling copy number\nalterations (CNA). This, however, is still challenging and requires\ntime-consuming manual curation. As deep-learning methods outperformed classical\nstate-of-the-art algorithms in various domains and have also been successfully\napplied to life science problems including medicine and biology, we here\npropose Deep SNP, a novel Deep Neural Network to learn from genomic data.\nSpecifically, we used a manually curated dataset from 12 genomic single\nnucleotide polymorphism array (SNPa) profiles as truth-set and aimed at\npredicting the presence or absence of genomic breakpoints, an indicator of\nstructural chromosomal variations, in windows of 40,000 probes. We compare our\nresults with well-known neural network models as well as Rawcopy though this\ntool is designed to predict breakpoints and in addition genomic segments with\nhigh sensitivity. We show, that Deep SNP is capable of successfully predicting\nthe presence or absence of a breakpoint in large genomic windows and\noutperforms state-of-the-art neural network models. Qualitative examples\nsuggest that integration of a localization unit may enable breakpoint detection\nand prediction of genomic segments, even if the breakpoint coordinates were not\nprovided for network training. These results warrant further evaluation of\nDeepSNP for breakpoint localization and subsequent calling of genomic segments.\n",
"title": "Deep SNP: An End-to-end Deep Neural Network with Attention-based Localization for Break-point Detection in SNP Array Genomic data"
}
| null | null | null | null | true | null |
3963
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, the notion of $(L,M)$-fuzzy convex structures is introduced.\nIt is a generalization of $L$-convex structures and $M$-fuzzifying convex\nstructures. In our definition of $(L,M)$-fuzzy convex structures, each\n$L$-fuzzy subset can be regarded as an $L$-convex set to some degree. The\nnotion of convexity preserving functions is also generalized to lattice-valued\ncase. Moreover, under the framework of $(L,M)$-fuzzy convex structures, the\nconcepts of quotient structures, substructures and products are presented and\ntheir fundamental properties are discussed. Finally, we create a functor\n$\\omega$ from $\\mathbf{MYCS}$ to $\\mathbf{LMCS}$ and show that there exists an\nadjunction between $\\mathbf{MYCS}$ and $\\mathbf{LMCS}$, where $\\mathbf{MYCS}$\nand $\\mathbf{LMCS}$ denote the category of $M$-fuzzifying convex structures,\nand the category of $(L,M)$-fuzzy convex structures, respectively.\n",
"title": "$(L,M)$-fuzzy convex structures"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3964
| null |
Validated
| null | null |
null |
{
"abstract": " This paper presents a passive compliance control for aerial manipulators to\nachieve stable environmental interactions. The main challenge is the absence of\nactuation along body-planar directions of the aerial vehicle which might be\nrequired during the interaction to preserve passivity. The controller proposed\nin this paper guarantees passivity of the manipulator through a proper choice\nof end-effector coordinates, and that of vehicle fuselage is guaranteed by\nexploiting time domain passivity technique. Simulation studies validate the\nproposed approach.\n",
"title": "Passive Compliance Control of Aerial Manipulators"
}
| null | null |
[
"Computer Science"
] | null | true | null |
3965
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper we consider the phase retrieval problem for Herglotz functions,\nthat is, solutions of the Helmholtz equation $\\Delta u+\\lambda^2u=0$ on domains\n$\\Omega\\subset\\mathbb{R}^d$, $d\\geq2$. In dimension $d=2$, if $u,v$ are two\nsuch solutions then $|u|=|v|$ implies that either $u=cv$ or $u=c\\bar v$ for\nsome $c\\in\\mathbb{C}$ with $|c|=1$. In dimension $d\\geq3$, the same conclusion\nholds under some restriction on $u$ and $v$: either they are real valued or\nzonal functions or have non vanishing mean.\n",
"title": "The phase retrieval problem for solutions of the Helmholtz equation"
}
| null | null | null | null | true | null |
3966
| null |
Default
| null | null |
null |
{
"abstract": " A family of subsets of $\\{1,\\ldots,n\\}$ is called {\\it intersecting} if any\ntwo of its sets intersect. A classical result in extremal combinatorics due to\nErdős, Ko, and Rado determines the maximum size of an intersecting family\nof $k$-subsets of $\\{1,\\ldots, n\\}$. In this paper we study the following\nproblem: how many intersecting families of $k$-subsets of $\\{1,\\ldots, n\\}$ are\nthere? Improving a result of Balogh, Das, Delcourt, Liu, and Sharifzadeh, we\ndetermine this quantity asymptotically for $n\\ge 2k+2+2\\sqrt{k\\log k}$ and\n$k\\to \\infty$. Moreover, under the same assumptions we also determine\nasymptotically the number of {\\it non-trivial} intersecting families, that is,\nintersecting families for which the intersection of all sets is empty. We\nobtain analogous results for pairs of cross-intersecting families.\n",
"title": "Counting intersecting and pairs of cross-intersecting families"
}
| null | null | null | null | true | null |
3967
| null |
Default
| null | null |
null |
{
"abstract": " Magnetic skyrmions are swirling spin textures with topologically protected\nnoncoplanarity. Recently, skyrmions with the topological number of unity have\nbeen extensively studied in both experiment and theory. We here show that a\nskyrmion crystal with an unusually high topological number of two is stabilized\nin itinerant magnets at zero magnetic field. The results are obtained for a\nminimal Kondo lattice model on a triangular lattice by an unrestricted\nlarge-scale numerical simulation and variational calculations. We find that the\ntopological number can be switched by a magnetic field as $2\\leftrightarrow\n1\\leftrightarrow 0$. The skyrmion crystals are formed by the superpositions of\nthree spin density waves induced by the Fermi surface effect, and hence, the\nsize of skyrmions can be controlled by the band structure and electron filling.\nWe also discuss the charge and spin textures of itinerant electrons in the\nskyrmion crystals which are directly obtained in our numerical simulations.\n",
"title": "Zero-field Skyrmions with a High Topological Number in Itinerant Magnets"
}
| null | null | null | null | true | null |
3968
| null |
Default
| null | null |
null |
{
"abstract": " Training deep neural network policies end-to-end for real-world applications\nso far requires big demonstration datasets in the real world or big sets\nconsisting of a large variety of realistic and closely related 3D CAD models.\nThese real or virtual data should, moreover, have very similar characteristics\nto the conditions expected at test time. These stringent requirements and the\ntime consuming data collection processes that they entail, are currently the\nmost important impediment that keeps deep reinforcement learning from being\ndeployed in real-world applications. Therefore, in this work we advocate an\nalternative approach, where instead of avoiding any domain shift by carefully\nselecting the training data, the goal is to learn a policy that can cope with\nit. To this end, we propose the DoShiCo challenge: to train a model in very\nbasic synthetic environments, far from realistic, in a way that it can be\napplied in more realistic environments as well as take the control decisions on\nreal-world data. In particular, we focus on the task of collision avoidance for\ndrones. We created a set of simulated environments that can be used as\nbenchmark and implemented a baseline method, exploiting depth prediction as an\nauxiliary task to help overcome the domain shift. Even though the policy is\ntrained in very basic environments, it can learn to fly without collisions in a\nvery different realistic simulated environment. Of course several benchmarks\nfor reinforcement learning already exist - but they never include a large\ndomain shift. On the other hand, several benchmarks in computer vision focus on\nthe domain shift, but they take the form of a static datasets instead of\nsimulated environments. In this work we claim that it is crucial to take the\ntwo challenges together in one benchmark.\n",
"title": "DoShiCo Challenge: Domain Shift in Control Prediction"
}
| null | null | null | null | true | null |
3969
| null |
Default
| null | null |
null |
{
"abstract": " Interpretability has become an important issue in the machine learning field,\nalong with the success of layered neural networks in various practical tasks.\nSince a trained layered neural network consists of a complex nonlinear\nrelationship between large number of parameters, we failed to understand how\nthey could achieve input-output mappings with a given data set. In this paper,\nwe propose the non-negative task decomposition method, which applies\nnon-negative matrix factorization to a trained layered neural network. This\nenables us to decompose the inference mechanism of a trained layered neural\nnetwork into multiple principal tasks of input-output mapping, and reveal the\nroles of hidden units in terms of their contribution to each principal task.\n",
"title": "Knowledge Discovery from Layered Neural Networks based on Non-negative Task Decomposition"
}
| null | null | null | null | true | null |
3970
| null |
Default
| null | null |
null |
{
"abstract": " We propose an optimal sequential methodology for obtaining confidence\nintervals for a binomial proportion $\\theta$. Assuming that an i.i.d. random\nsequence of Benoulli($\\theta$) trials is observed sequentially, we are\ninterested in designing a)~a stopping time $T$ that will decide when is the\nbest time to stop sampling the process, and b)~an optimum estimator\n$\\hat{\\theta}_{T}$ that will provide the optimum center of the interval\nestimate of $\\theta$. We follow a semi-Bayesian approach, where we assume that\nthere exists a prior distribution for $\\theta$, and our goal is to minimize the\naverage number of samples while we guarantee a minimal coverage probability\nlevel. The solution is obtained by applying standard optimal stopping theory\nand computing the optimum pair $(T,\\hat{\\theta}_{T})$ numerically. Regarding\nthe optimum stopping time component $T$, we demonstrate that it enjoys certain\nvery uncommon characteristics not encountered in solutions of other classical\noptimal stopping problems. Finally, we compare our method with the optimum\nfixed-sample-size procedure but also with existing alternative sequential\nschemes.\n",
"title": "Optimal Stopping for Interval Estimation in Bernoulli Trials"
}
| null | null |
[
"Statistics"
] | null | true | null |
3971
| null |
Validated
| null | null |
null |
{
"abstract": " A linear Boltzmann equation with nonautonomous collision operator is\nrigorously derived in the Boltzmann-Grad limit for the deterministic dynamics\nof a Rayleigh gas where a tagged particle is undergoing hard-sphere collisions\nwith heterogeneously distributed background particles, which do not interact\namong each other. The validity of the linear Boltzmann equation holds for\narbitrary long times under moderate assumptions on spatial continuity and\nhigher moments of the initial distributions of the tagged particle and the\nheterogeneous, non-equilibrium distribution of the background. The empiric\nparticle dynamics are compared to the Boltzmann dynamics using evolution\nsemigroups for Kolmogorov equations of associated probability measures on\ncollision histories.\n",
"title": "Derivation of a Non-autonomous Linear Boltzmann Equation from a Heterogeneous Rayleigh Gas"
}
| null | null | null | null | true | null |
3972
| null |
Default
| null | null |
null |
{
"abstract": " For any $n\\geq 3$ and $ q\\geq 3$, we prove that the {\\sc Equality} function\n$(=_n)$ on $n$ variables over a domain of size $q$ cannot be realized by\nmatchgates under holographic transformations. This is a consequence of our\ntheorem on the structure of blockwise symmetric matchgate signatures. %due to\nthe rank of the matrix form of the blockwise symmetric standard signatures,\n%where $(=_n)$ is an equality signature on domain $\\{0, 1, \\cdots, q-1\\}$. This\nhas the implication that the standard holographic algorithms based on\nmatchgates, a methodology known to be universal for \\#CSP over the Boolean\ndomain, cannot produce P-time algorithms for planar \\#CSP over any higher\ndomain $q\\geq 3$.\n",
"title": "On Blockwise Symmetric Matchgate Signatures and Higher Domain \\#CSP"
}
| null | null | null | null | true | null |
3973
| null |
Default
| null | null |
null |
{
"abstract": " The deconfined quantum critical point (QCP), separating the Néel and\nvalence bond solid phases in a 2D antiferromagnet, was proposed as an example\nof $2+1$D criticality fundamentally different from standard\nLandau-Ginzburg-Wilson-Fisher {criticality}. In this work we present multiple\nequivalent descriptions of deconfined QCPs, and use these to address the\npossibility of enlarged emergent symmetries in the low energy limit. The\neasy-plane deconfined QCP, besides its previously discussed self-duality, is\ndual to $N_f = 2$ fermionic quantum electrodynamics (QED), which has its own\nself-duality and hence may have an O(4)$\\times Z_2^T$ symmetry. We propose\nseveral dualities for the deconfined QCP with ${\\mathrm{SU}(2)}$ spin symmetry\nwhich together make natural the emergence of a previously suggested $SO(5)$\nsymmetry rotating the Néel and VBS orders. These emergent symmetries are\nimplemented anomalously. The associated infra-red theories can also be viewed\nas surface descriptions of 3+1D topological paramagnets, giving further insight\ninto the dualities. We describe a number of numerical tests of these dualities.\nWe also discuss the possibility of \"pseudocritical\" behavior for deconfined\ncritical points, and the meaning of the dualities and emergent symmetries in\nsuch a scenario.\n",
"title": "Deconfined quantum critical points: symmetries and dualities"
}
| null | null |
[
"Physics"
] | null | true | null |
3974
| null |
Validated
| null | null |
null |
{
"abstract": " Deep learning has become the state of the art approach in many machine\nlearning problems such as classification. It has recently been shown that deep\nlearning is highly vulnerable to adversarial perturbations. Taking the camera\nsystems of self-driving cars as an example, small adversarial perturbations can\ncause the system to make errors in important tasks, such as classifying traffic\nsigns or detecting pedestrians. Hence, in order to use deep learning without\nsafety concerns a proper defense strategy is required. We propose to use\nensemble methods as a defense strategy against adversarial perturbations. We\nfind that an attack leading one model to misclassify does not imply the same\nfor other networks performing the same task. This makes ensemble methods an\nattractive defense strategy against adversarial attacks. We empirically show\nfor the MNIST and the CIFAR-10 data sets that ensemble methods not only improve\nthe accuracy of neural networks on test data but also increase their robustness\nagainst adversarial perturbations.\n",
"title": "Ensemble Methods as a Defense to Adversarial Perturbations Against Deep Neural Networks"
}
| null | null |
[
"Statistics"
] | null | true | null |
3975
| null |
Validated
| null | null |
null |
{
"abstract": " In this letter we prove that the unrolled small quantum group, appearing in\nquantum topology, is a Hopf subalgebra of Lusztig's quantum group of divided\npowers. We do so by writing down non-obvious primitive elements with the right\nadjoint action. We also construct a new larger Hopf algebra that contains the\nfull unrolled quantum group. In fact this Hopf algebra contains both the\nenveloping of the Lie algebra and the ring of functions on the Lie group, and\nit should be interesting in its own right. We finally explain how this gives a\nrealization of the unrolled quantum group as operators on a conformal field\ntheory and match some calculations on this side. Our result extends to other\nNichols algebras of diagonal type, including super Lie algebras.\n",
"title": "The unrolled quantum group inside Lusztig's quantum group of divided powers"
}
| null | null | null | null | true | null |
3976
| null |
Default
| null | null |
null |
{
"abstract": " Increasing safety and automation in transportation systems has led to the\nproliferation of radar and IEEE 802.11 dedicated short range communication\n(DSRC) in vehicles. Current implementations of vehicular radar devices,\nhowever, are expensive, use a substantial amount of bandwidth, and are\nsusceptible to multiple security risks. Consider the feasibility of using an\nIEEE 802.11 orthogonal frequency division multiplexing (OFDM) communications\nwaveform to perform radar functions. In this paper, we present an approach that\ndetermines the mean-normalized channel energy from frequency domain channel\nestimates and models it as a direct sinusoidal function of target range,\nenabling closest target range estimation. In addition, we propose an\nalternative to vehicular forward collision detection by extending IEEE 802.11\ndedicated short-range communications (DSRC) and WiFi technology to radar,\nproviding a foundation for joint communications and radar framework.\nFurthermore, we perform an experimental demonstration using existing IEEE\n802.11 devices with minimal modification through algorithm processing on\nfrequency-domain channel estimates. The results of this paper show that our\nsolution delivers similar accuracy and reliability to mmWave radar devices with\nas little as 20 MHz of spectrum (doubling DSRC's 10 MHz allocation), indicating\nsignificant potential for industrial devices with joint vehicular\ncommunications and radar capabilities.\n",
"title": "Forward Collision Vehicular Radar with IEEE 802.11: Feasibility Demonstration through Measurements"
}
| null | null | null | null | true | null |
3977
| null |
Default
| null | null |
null |
{
"abstract": " With the advent of numerous online content providers, utilities and\napplications, each with their own specific version of privacy policies and its\nassociated overhead, it is becoming increasingly difficult for concerned users\nto manage and track the confidential information that they share with the\nproviders. Users consent to providers to gather and share their Personally\nIdentifiable Information (PII). We have developed a novel framework to\nautomatically track details about how a users' PII data is stored, used and\nshared by the provider. We have integrated our Data Privacy ontology with the\nproperties of blockchain, to develop an automated access control and audit\nmechanism that enforces users' data privacy policies when sharing their data\nacross third parties. We have also validated this framework by implementing a\nworking system LinkShare. In this paper, we describe our framework on detail\nalong with the LinkShare system. Our approach can be adopted by Big Data users\nto automatically apply their privacy policy on data operations and track the\nflow of that data across various stakeholders.\n",
"title": "Link Before You Share: Managing Privacy Policies through Blockchain"
}
| null | null |
[
"Computer Science"
] | null | true | null |
3978
| null |
Validated
| null | null |
null |
{
"abstract": " Accurately predicting and detecting interstitial lung disease (ILD) patterns\ngiven any computed tomography (CT) slice without any pre-processing\nprerequisites, such as manually delineated regions of interest (ROIs), is a\nclinically desirable, yet challenging goal. The majority of existing work\nrelies on manually-provided ILD ROIs to extract sampled 2D image patches from\nCT slices and, from there, performs patch-based ILD categorization. Acquiring\nmanual ROIs is labor intensive and serves as a bottleneck towards\nfully-automated CT imaging ILD screening over large-scale populations.\nFurthermore, despite the considerable high frequency of more than one ILD\npattern on a single CT slice, previous works are only designed to detect one\nILD pattern per slice or patch.\nTo tackle these two critical challenges, we present multi-label deep\nconvolutional neural networks (CNNs) for detecting ILDs from holistic CT slices\n(instead of ROIs or sub-images). Conventional single-labeled CNN models can be\naugmented to cope with the possible presence of multiple ILD pattern labels,\nvia 1) continuous-valued deep regression based robust norm loss functions or 2)\na categorical objective as the sum of element-wise binary logistic losses. Our\nmethods are evaluated and validated using a publicly available database of 658\npatient CT scans under five-fold cross-validation, achieving promising\nperformance on detecting four major ILD patterns: Ground Glass, Reticular,\nHoneycomb, and Emphysema. We also investigate the effectiveness of a CNN\nactivation-based deep-feature encoding scheme using Fisher vector encoding,\nwhich treats ILD detection as spatially-unordered deep texture classification.\n",
"title": "Holistic Interstitial Lung Disease Detection using Deep Convolutional Neural Networks: Multi-label Learning and Unordered Pooling"
}
| null | null |
[
"Computer Science"
] | null | true | null |
3979
| null |
Validated
| null | null |
null |
{
"abstract": " We study whether a depth two neural network can learn another depth two\nnetwork using gradient descent. Assuming a linear output node, we show that the\nquestion of whether gradient descent converges to the target function is\nequivalent to the following question in electrodynamics: Given $k$ fixed\nprotons in $\\mathbb{R}^d,$ and $k$ electrons, each moving due to the attractive\nforce from the protons and repulsive force from the remaining electrons,\nwhether at equilibrium all the electrons will be matched up with the protons,\nup to a permutation. Under the standard electrical force, this follows from the\nclassic Earnshaw's theorem. In our setting, the force is determined by the\nactivation function and the input distribution. Building on this equivalence,\nwe prove the existence of an activation function such that gradient descent\nlearns at least one of the hidden nodes in the target network. Iterating, we\nshow that gradient descent can be used to learn the entire network one node at\na time.\n",
"title": "Convergence Results for Neural Networks via Electrodynamics"
}
| null | null | null | null | true | null |
3980
| null |
Default
| null | null |
null |
{
"abstract": " Since the development of higher local class field theory, several explicit\nreciprocity laws have been constructed. In particular, there are formulas\ndescribing the higher-dimensional Hilbert symbol given, among others, by M.\nKurihara, A. Zinoviev and S. Vostokov. K. Kato also has explicit formulas for\nthe higher-dimensional Kummer pairing associated to certain (one-dimensional)\n$p$-divisible groups.\nIn this paper we construct an explicit reciprocity law describing the Kummer\npairing associated to any (one-dimensional) formal group. The formulas are a\ngeneralization to higher-dimensional local fields of Kolyvagin's reciprocity\nlaws. The formulas obtained describe the values of the pairing in terms of\nmultidimensional $p$-adic differentiation, the logarithm of the formal group,\nthe generalized trace and the norm on Milnor K-groups.\nIn the second part of this paper, we will apply the results obtained here to\ngive explicit formulas for the generalized Hilbert symbol and the Kummer\npairing associated to a Lubin-Tate formal group. The results obtained in the\nsecond paper constitute a generalization to higher local fields, of the\nformulas of Artin-Hasse, K. Iwasawa and A. Wiles.\n",
"title": "The norm residue symbol for higher local fields"
}
| null | null | null | null | true | null |
3981
| null |
Default
| null | null |
null |
{
"abstract": " The classic algorithm of Bodlaender and Kloks [J. Algorithms, 1996] solves\nthe following problem in linear fixed-parameter time: given a tree\ndecomposition of a graph of (possibly suboptimal) width $k$, compute an\noptimum-width tree decomposition of the graph. In this work, we prove that this\nproblem can also be solved in MSO in the following sense: for every positive\ninteger $k$, there is an MSO transduction from tree decompositions of width $k$\nto tree decompositions of optimum width. Together with our recent results [LICS\n2016], this implies that for every $k$ there exists an MSO transduction which\ninputs a graph of treewidth $k$, and nondeterministically outputs its tree\ndecomposition of optimum width. We also show that MSO transductions can be\nimplemented in linear fixed-parameter time, which enables us to derive the\nalgorithmic result of Bodlaender and Kloks as a corollary of our main result.\n",
"title": "Optimizing tree decompositions in MSO"
}
| null | null | null | null | true | null |
3982
| null |
Default
| null | null |
null |
{
"abstract": " Let $ (T_i)_i$ be a sequence of independent identically distributed (i.i.d.)\nrandom variables (r.v.) of interest distributed as $ T$ and $(X_i)_i$ be a\ncorresponding vector of covariates taking values on $ \\mathbb{R}^d$. In\ncensorship models the r.v. $T$ is subject to random censoring by another r.v.\n$C$. In this paper we built a new kernel estimator based on the so-called\nsynthetic data of the mean squared relative error for the regression function.\nWe establish the uniform almost sure convergence with rate over a compact set\nand its asymptotic normality. The asymptotic variance is explicitly given and\nas product we give a confidence bands. A simulation study has been conducted to\ncomfort our theoretical results\n",
"title": "Nonparametric relative error estimation of the regression function for censored data"
}
| null | null | null | null | true | null |
3983
| null |
Default
| null | null |
null |
{
"abstract": " This paper presents an intelligent home energy management system integrated\nwith dispatchable loads (e.g., clothes washers and dryers), distributed\nrenewable generators (e.g., roof-top solar panels), and distributed energy\nstorage devices (e.g., plug-in electric vehicles). The overall goal is to\nreduce the total operating costs and the carbon emissions for a future\nresidential house, while satisfying the end-users comfort levels. This paper\nmodels a wide variety of home appliances and formulates the economic operation\nproblem using mixed integer linear programming. Case studies are performed to\nvalidate and demonstrate the effectiveness of the proposed solution algorithm.\nSimulation results also show the positive impact of dispatchable loads,\ndistributed renewable generators, and distributed energy storage devices on a\nfuture residential house.\n",
"title": "Intelligent Home Energy Management System for Distributed Renewable Generators, Dispatchable Residential Loads and Distributed Energy Storage Devices"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3984
| null |
Validated
| null | null |
null |
{
"abstract": " This paper develops a general framework for learning interpretable data\nrepresentation via Long Short-Term Memory (LSTM) recurrent neural networks over\nhierarchal graph structures. Instead of learning LSTM models over the pre-fixed\nstructures, we propose to further learn the intermediate interpretable\nmulti-level graph structures in a progressive and stochastic way from data\nduring the LSTM network optimization. We thus call this model the\nstructure-evolving LSTM. In particular, starting with an initial element-level\ngraph representation where each node is a small data element, the\nstructure-evolving LSTM gradually evolves the multi-level graph representations\nby stochastically merging the graph nodes with high compatibilities along the\nstacked LSTM layers. In each LSTM layer, we estimate the compatibility of two\nconnected nodes from their corresponding LSTM gate outputs, which is used to\ngenerate a merging probability. The candidate graph structures are accordingly\ngenerated where the nodes are grouped into cliques with their merging\nprobabilities. We then produce the new graph structure with a\nMetropolis-Hasting algorithm, which alleviates the risk of getting stuck in\nlocal optimums by stochastic sampling with an acceptance probability. Once a\ngraph structure is accepted, a higher-level graph is then constructed by taking\nthe partitioned cliques as its nodes. During the evolving process,\nrepresentation becomes more abstracted in higher-levels where redundant\ninformation is filtered out, allowing more efficient propagation of long-range\ndata dependencies. We evaluate the effectiveness of structure-evolving LSTM in\nthe application of semantic object parsing and demonstrate its advantage over\nstate-of-the-art LSTM models on standard benchmarks.\n",
"title": "Interpretable Structure-Evolving LSTM"
}
| null | null | null | null | true | null |
3985
| null |
Default
| null | null |
null |
{
"abstract": " We investigate the use of optimization to compute bounds for extremal\nperformance measures. This approach takes a non-parametric viewpoint that aims\nto alleviate the issue of model misspecification possibly encountered by\nconventional methods in extreme event analysis. We make two contributions\ntowards solving these formulations, paying especial attention to the arising\ntail issues. First, we provide a technique in parallel to Choquet's theory, via\na combination of integration by parts and change of measures, to transform\nshape constrained problems (e.g., monotonicity of derivatives) into families of\nmoment problems. Second, we show how a moment problem cast over infinite\nsupport can be reformulated into a problem over compact support with an\nadditional slack variable. In the context of optimization over tail\ndistributions, the latter helps resolve the issue of non-convergence of\nsolutions when using algorithms such as generalized linear programming. We\nfurther demonstrate the applicability of this result to problems with\ninfinite-value constraints, which can arise in modeling heavy tails.\n",
"title": "On Optimization over Tail Distributions"
}
| null | null | null | null | true | null |
3986
| null |
Default
| null | null |
null |
{
"abstract": " We develop parametric classes of covariance functions on linear networks and\ntheir extension to graphs with Euclidean edges, i.e., graphs with edges viewed\nas line segments or more general sets with a coordinate system allowing us to\nconsider points on the graph which are vertices or points on an edge. Our\ncovariance functions are defined on the vertices and edge points of these\ngraphs and are isotropic in the sense that they depend only on the geodesic\ndistance or on a new metric called the resistance metric (which extends the\nclassical resistance metric developed in electrical network theory on the\nvertices of a graph to the continuum of edge points). We discuss the advantages\nof using the resistance metric in comparison with the geodesic metric as well\nas the restrictions these metrics impose on the investigated covariance\nfunctions. In particular, many of the commonly used isotropic covariance\nfunctions in the spatial statistics literature (the power exponential,\nMat{é}rn, generalized Cauchy, and Dagum classes) are shown to be valid with\nrespect to the resistance metric for any graph with Euclidean edges, whilst\nthey are only valid with respect to the geodesic metric in more special cases.\n",
"title": "Isotropic covariance functions on graphs and their edges"
}
| null | null | null | null | true | null |
3987
| null |
Default
| null | null |
null |
{
"abstract": " Cell injection is a technique in the domain of biological cell\nmicro-manipulation for the delivery of small volumes of samples into the\nsuspended or adherent cells. It has been widely applied in various areas, such\nas gene injection, in-vitro fertilization (IVF), intracytoplasmic sperm\ninjection (ISCI) and drug development. However, the existing manual and\nsemi-automated cell injection systems require lengthy training and suffer from\nhigh probability of contamination and low success rate. In the recently\nintroduced fully automated cell injection systems, the injection force plays a\nvital role in the success of the process since even a tiny excessive force can\ndestroy the membrane or tissue of the biological cell. Traditionally, the force\ncontrol algorithms are analyzed using simulation, which is inherently\nnon-exhaustive and incomplete in terms of detecting system failures. Moreover,\nthe uncertainties in the system are generally ignored in the analysis. To\novercome these limitations, we present a formal analysis methodology based on\nprobabilistic model checking to analyze a robotic cell injection system\nutilizing the impedance force control algorithm. The proposed methodology,\ndeveloped using the PRISM model checker, allowed to find a discrepancy in the\nalgorithm, which was not found by any of the previous analysis using the\ntraditional methods.\n",
"title": "Towards Probabilistic Formal Modeling of Robotic Cell Injection Systems"
}
| null | null | null | null | true | null |
3988
| null |
Default
| null | null |
null |
{
"abstract": " We present a visually grounded model of speech perception which projects\nspoken utterances and images to a joint semantic space. We use a multi-layer\nrecurrent highway network to model the temporal nature of spoken speech, and\nshow that it learns to extract both form and meaning-based linguistic knowledge\nfrom the input signal. We carry out an in-depth analysis of the representations\nused by different components of the trained model and show that encoding of\nsemantic aspects tends to become richer as we go up the hierarchy of layers,\nwhereas encoding of form-related aspects of the language input tends to\ninitially increase and then plateau or decrease.\n",
"title": "Representations of language in a model of visually grounded speech signal"
}
| null | null | null | null | true | null |
3989
| null |
Default
| null | null |
null |
{
"abstract": " This paper considers the challenging task of long-term video interpolation.\nUnlike most existing methods that only generate few intermediate frames between\nexisting adjacent ones, we attempt to speculate or imagine the procedure of an\nepisode and further generate multiple frames between two non-consecutive frames\nin videos. In this paper, we present a novel deep architecture called\nbidirectional predictive network (BiPN) that predicts intermediate frames from\ntwo opposite directions. The bidirectional architecture allows the model to\nlearn scene transformation with time as well as generate longer video\nsequences. Besides, our model can be extended to predict multiple possible\nprocedures by sampling different noise vectors. A joint loss composed of clues\nin image and feature spaces and adversarial loss is designed to train our\nmodel. We demonstrate the advantages of BiPN on two benchmarks Moving 2D Shapes\nand UCF101 and report competitive results to recent approaches.\n",
"title": "Long-Term Video Interpolation with Bidirectional Predictive Network"
}
| null | null | null | null | true | null |
3990
| null |
Default
| null | null |
null |
{
"abstract": " Cooperation is a difficult proposition in the face of Darwinian selection.\nThose that defect have an evolutionary advantage over cooperators who should\ntherefore die out. However, spatial structure enables cooperators to survive\nthrough the formation of homogeneous clusters, which is the hallmark of network\nreciprocity. Here we go beyond this traditional setup and study the\nspatiotemporal dynamics of cooperation in a population of populations. We use\nthe prisoner's dilemma game as the mathematical model and show that considering\nseveral populations simultaneously give rise to fascinating spatiotemporal\ndynamics and pattern formation. Even the simplest assumption that strategies\nbetween different populations are payoff-neutral with one another results in\nthe spontaneous emergence of cyclic dominance, where defectors of one\npopulation become prey of cooperators in the other population, and vice versa.\nMoreover, if social interactions within different populations are characterized\nby significantly different temptations to defect, we observe that defectors in\nthe population with the largest temptation counterintuitively vanish the\nfastest, while cooperators that hang on eventually take over the whole\navailable space. Our results reveal that considering the simultaneous presence\nof different populations significantly expands the complexity of evolutionary\ndynamics in structured populations, and it allow us to understand the stability\nof cooperation under adverse conditions that could never be bridged by network\nreciprocity alone.\n",
"title": "Evolutionary dynamics of cooperation in neutral populations"
}
| null | null | null | null | true | null |
3991
| null |
Default
| null | null |
null |
{
"abstract": " We consider the parabolic-elliptic model for the chemotaxis with fractional\n(anomalous) diffusion. Global-in-time solutions are constructed under (nearly)\noptimal assumptions on the size of radial initial data. Moreover, criteria for\nblowup of radial solutions in terms of suitable Morrey spaces norms are\nderived.\n",
"title": "Large global-in-time solutions to a nonlocal model of chemotaxis"
}
| null | null | null | null | true | null |
3992
| null |
Default
| null | null |
null |
{
"abstract": " Microservice Architectures (MA) have the potential to increase the agility of\nsoftware development. In an era where businesses require software applications\nto evolve to support software emerging requirements, particularly for Internet\nof Things (IoT) applications, we examine the issue of microservice granularity\nand explore its effect upon application latency. Two approaches to microservice\ndeployment are simulated; the first with microservices in a single container,\nand the second with microservices partitioned across separate containers. We\nobserved a neglibible increase in service latency for the multiple container\ndeployment over a single container.\n",
"title": "Microservices: Granularity vs. Performance"
}
| null | null |
[
"Computer Science"
] | null | true | null |
3993
| null |
Validated
| null | null |
null |
{
"abstract": " The strong-interaction limit of the Hohenberg-Kohn functional defines a\nmultimarginal optimal transport problem with Coulomb cost. From physical\narguments, the solution of this limit is expected to yield strictly-correlated\nparticle positions, related to each other by co-motion functions (or optimal\nmaps), but the existence of such a deterministic solution in the general\nthree-dimensional case is still an open question. A conjecture for the\nco-motion functions for radially symmetric densities was presented in\nPhys.~Rev.~A {\\bf 75}, 042511 (2007), and later used to build approximate\nexchange-correlation functionals for electrons confined in low-density quantum\ndots. Colombo and Stra [Math.~Models Methods Appl.~Sci., {\\bf 26} 1025 (2016)]\nhave recently shown that these conjectured maps are not always optimal. Here we\nrevisit the whole issue both from the formal and numerical point of view,\nfinding that even if the conjectured maps are not always optimal, they still\nyield an interaction energy (cost) that is numerically very close to the true\nminimum. We also prove that the functional built from the conjectured maps has\nthe expected functional derivative also when they are not optimal.\n",
"title": "The strictly-correlated electron functional for spherically symmetric systems revisited"
}
| null | null | null | null | true | null |
3994
| null |
Default
| null | null |
null |
{
"abstract": " The coupling of Reynolds and Rayleigh-Plesset equations has been used in\nseveral works to simulate lubricated devices considering cavitation. The\nnumerical strategies proposed so far are variants of a staggered strategy where\nReynolds equation is solved considering the bubble dynamics frozen, and then\nthe Rayleigh-Plesset equation is solved to update the bubble radius with the\npressure frozen. We show that this strategy has severe stability issues and a\nstable methodology is proposed. The proposed methodology performance is\nassessed on two physical settings. The first one concerns the propagation of a\ndecompression wave along a fracture considering the presence of cavitation\nnuclei. The second one is a typical journal bearing, in which the coupled model\nis compared with the Elrod-Adams model.\n",
"title": "A stable numerical strategy for Reynolds-Rayleigh-Plesset coupling"
}
| null | null | null | null | true | null |
3995
| null |
Default
| null | null |
null |
{
"abstract": " State-of-the-art algorithms for sparse subspace clustering perform spectral\nclustering on a similarity matrix typically obtained by representing each data\npoint as a sparse combination of other points using either basis pursuit (BP)\nor orthogonal matching pursuit (OMP). BP-based methods are often prohibitive in\npractice while the performance of OMP-based schemes are unsatisfactory,\nespecially in settings where data points are highly similar. In this paper, we\npropose a novel algorithm that exploits an accelerated variant of orthogonal\nleast-squares to efficiently find the underlying subspaces. We show that under\ncertain conditions the proposed algorithm returns a subspace-preserving\nsolution. Simulation results illustrate that the proposed method compares\nfavorably with BP-based method in terms of running time while being\nsignificantly more accurate than OMP-based schemes.\n",
"title": "Accelerated Sparse Subspace Clustering"
}
| null | null | null | null | true | null |
3996
| null |
Default
| null | null |
null |
{
"abstract": " Steady State Superconducting Tokamak (SST-1) at the Institute for Plasma\nResearch (IPR) is an operational device and is the first superconducting\nTokamak in India. Superconducting Magnets System (SCMS) in SST-1 comprises of\nsixteen Toroidal field (TF) magnets and nine Poloidal Field (PF) magnets\nmanufactured using NbTi/Cu based cable-in-conduit-conductor (CICC) concept.\nSST-1, superconducting TF magnets are operated in a Cryo-stable manner being\ncooled with two-phase (TP) flow helium. The typical operating pressure of the\nTP helium is 1.6 bar (a) at corresponding saturation temperature. The SCMS has\na typical cool-down time of about 14 days from 300 K down to 4.5 K using Helium\nplant of equivalent cooling capacity of 1350 W at 4.5 K. Using the onset of\nexperimental data from the HRL, we estimated the vapor quality for the input\nheat load on to the TF magnets system. In this paper, we report the\ncharacteristics of two-phase flow for given thermo-hydraulic conditions during\nlong steady state operation of the SST-1 TF magnets. Finally, the\nexperimentally obtained results have been compared with the well-known\ncorrelations of two-phase flow.\n",
"title": "Prediction of helium vapor quality in steady state Two-phase operation for SST-1 Toroidal field magnets"
}
| null | null | null | null | true | null |
3997
| null |
Default
| null | null |
null |
{
"abstract": " Pseudo-one dimensional (pseudo-1D) materials are a new-class of materials\nwhere atoms are arranged in chain like structures in two-dimensions (2D).\nExamples include recently discovered black phosphorus, ReS2 and ReSe2 from\ntransition metal dichalcogenides, TiS3 and ZrS3 from transition metal\ntrichalcogenides and most recently GaTe. The presence of structural anisotropy\nimpacts their physical properties and leads to direction dependent light-matter\ninteractions, dichroic optical responses, high mobility channels, and\nanisotropic thermal conduction. Despite the numerous reports on the vapor phase\ngrowth of isotropic TMDCs and post transition metal chalcogenides such as MoS2\nand GaSe, the synthesis of pseudo-1D materials is particularly difficult due to\nthe anisotropy in interfacial energy, which stabilizes dendritic growth rather\nthan single crystalline growth with well-defined orientation. The growth of\nmonoclinic GaTe has been demonstrated on flexible mica substrates with superior\nphotodetecting performance. In this work, we demonstrate that pseudo-1D\nmonoclinic GaTe layers can be synthesized on a variety of other substrates\nincluding GaAs (111), Si (111) and c-cut sapphire by physical vapor transport\ntechniques. High resolution transmission electron microscopy (HRTEM)\nmeasurements, together with angle resolved micro-PL and micro-Raman techniques,\nprovide for the very first time atomic scale resolution experiments on\npseudo-1D structures in monoclinic GaTe and anisotropic properties.\nInterestingly, GaTe nanomaterials grown on sapphire exhibit highly efficient\nand narrow localized emission peaks below the band gap energy, which are found\nto be related to select types of line and point defects as evidenced by PL and\nRaman mapping scans. It makes the samples grown on sapphire more prominent than\nthose grown on GaAs and Si, which demonstrate more regular properties.\n",
"title": "Synthesis of Highly Anisotropic Semiconducting GaTe Nanomaterials and Emerging Properties Enabled by Epitaxy"
}
| null | null | null | null | true | null |
3998
| null |
Default
| null | null |
null |
{
"abstract": " This paper outlines a methodological approach for designing adaptive agents\ndriving themselves near points of criticality. Using a synthetic approach we\nconstruct a conceptual model that, instead of specifying mechanistic\nrequirements to generate criticality, exploits the maintenance of an\norganizational structure capable of reproducing critical behavior. Our approach\nexploits the well-known principle of universality, which classifies critical\nphenomena inside a few universality classes of systems independently of their\nspecific mechanisms or topologies. In particular, we implement an artificial\nembodied agent controlled by a neural network maintaining a correlation\nstructure randomly sampled from a lattice Ising model at a critical point. We\nevaluate the agent in two classical reinforcement learning scenarios: the\nMountain Car benchmark and the Acrobot double pendulum, finding that in both\ncases the neural controller reaches a point of criticality, which coincides\nwith a transition point between two regimes of the agent's behaviour,\nmaximizing the mutual information between neurons and sensorimotor patterns.\nFinally, we discuss the possible applications of this synthetic approach to the\ncomprehension of deeper principles connected to the pervasive presence of\ncriticality in biological and cognitive systems.\n",
"title": "Criticality as It Could Be: organizational invariance as self-organized criticality in embodied agents"
}
| null | null | null | null | true | null |
3999
| null |
Default
| null | null |
null |
{
"abstract": " The Frank-Wolfe (FW) algorithm has been widely used in solving nuclear norm\nconstrained problems, since it does not require projections. However, FW often\nyields high rank intermediate iterates, which can be very expensive in time and\nspace costs for large problems. To address this issue, we propose a rank-drop\nmethod for nuclear norm constrained problems. The goal is to generate descent\nsteps that lead to rank decreases, maintaining low-rank solutions throughout\nthe algorithm. Moreover, the optimization problems are constrained to ensure\nthat the rank-drop step is also feasible and can be readily incorporated into a\nprojection-free minimization method, e.g., Frank-Wolfe. We demonstrate that by\nincorporating rank-drop steps into the Frank-Wolfe algorithm, the rank of the\nsolution is greatly reduced compared to the original Frank-Wolfe or its common\nvariants.\n",
"title": "Projection Free Rank-Drop Steps"
}
| null | null | null | null | true | null |
4000
| null |
Default
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.