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": " Subordinate diffusions are constructed by time changing diffusion processes\nwith an independent Lévy subordinator. This is a rich family of Markovian\njump processes which exhibit a variety of jump behavior and have found many\napplications. This paper studies parametric inference of discretely observed\nergodic subordinate diffusions. We solve the identifiability problem for these\nprocesses using spectral theory and propose a two-step estimation procedure\nbased on estimating functions. In the first step, we use an estimating function\nthat only involves diffusion parameters. In the second step, a martingale\nestimating function based on eigenvalues and eigenfunctions of the subordinate\ndiffusion is used to estimate the parameters of the Lévy subordinator and\nthe problem of how to choose the weighting matrix is solved. When the\neigenpairs do not have analytical expressions, we apply the constant\nperturbation method with high order corrections to calculate them numerically\nand the martingale estimating function can be computed efficiently. Consistency\nand asymptotic normality of our estimator are established considering the\neffect of numerical approximation. Through numerical examples, we show that our\nmethod is both computationally and statistically efficient. A subordinate\ndiffusion model for VIX (CBOE volatility index) is developed which provides\ngood fit to the data.\n",
"title": "Parametric Inference for Discretely Observed Subordinate Diffusions"
}
| null | null | null | null | true | null |
14601
| null |
Default
| null | null |
null |
{
"abstract": " Graph embedding is an effective method to represent graph data in a low\ndimensional space for graph analytics. Most existing embedding algorithms\ntypically focus on preserving the topological structure or minimizing the\nreconstruction errors of graph data, but they have mostly ignored the data\ndistribution of the latent codes from the graphs, which often results in\ninferior embedding in real-world graph data. In this paper, we propose a novel\nadversarial graph embedding framework for graph data. The framework encodes the\ntopological structure and node content in a graph to a compact representation,\non which a decoder is trained to reconstruct the graph structure. Furthermore,\nthe latent representation is enforced to match a prior distribution via an\nadversarial training scheme. To learn a robust embedding, two variants of\nadversarial approaches, adversarially regularized graph autoencoder (ARGA) and\nadversarially regularized variational graph autoencoder (ARVGA), are developed.\nExperimental studies on real-world graphs validate our design and demonstrate\nthat our algorithms outperform baselines by a wide margin in link prediction,\ngraph clustering, and graph visualization tasks.\n",
"title": "Adversarially Regularized Graph Autoencoder for Graph Embedding"
}
| null | null | null | null | true | null |
14602
| null |
Default
| null | null |
null |
{
"abstract": " We introduce a novel approach to perform first-order optimization with\northogonal and unitary constraints. This approach is based on a parametrization\nstemming from Lie group theory through the exponential map. The parametrization\ntransforms the constrained optimization problem into an unconstrained one over\na Euclidean space, for which common first-order optimization methods can be\nused. The theoretical results presented are general enough to cover the special\northogonal group, the unitary group and, in general, any connected compact Lie\ngroup. We discuss how this and other parametrizations can be computed\nefficiently through an implementation trick, making numerically complex\nparametrizations usable at a negligible runtime cost in neural networks. In\nparticular, we apply our results to RNNs with orthogonal recurrent weights,\nyielding a new architecture called expRNN. We demonstrate how our method\nconstitutes a more robust approach to optimization with orthogonal constraints,\nshowing faster, accurate, and more stable convergence in several tasks designed\nto test RNNs.\n",
"title": "Cheap Orthogonal Constraints in Neural Networks: A Simple Parametrization of the Orthogonal and Unitary Group"
}
| null | null | null | null | true | null |
14603
| null |
Default
| null | null |
null |
{
"abstract": " Without access to large compute clusters, building random forests on large\ndatasets is still a challenging problem. This is, in particular, the case if\nfully-grown trees are desired. We propose a simple yet effective framework that\nallows to efficiently construct ensembles of huge trees for hundreds of\nmillions or even billions of training instances using a cheap desktop computer\nwith commodity hardware. The basic idea is to consider a multi-level\nconstruction scheme, which builds top trees for small random subsets of the\navailable data and which subsequently distributes all training instances to the\ntop trees' leaves for further processing. While being conceptually simple, the\noverall efficiency crucially depends on the particular implementation of the\ndifferent phases. The practical merits of our approach are demonstrated using\ndense datasets with hundreds of millions of training instances.\n",
"title": "Training Big Random Forests with Little Resources"
}
| null | null | null | null | true | null |
14604
| null |
Default
| null | null |
null |
{
"abstract": " Semiconductor nanowires provide an ideal platform for various low-dimensional\nquantum devices. In particular, topological phases of matter hosting\nnon-Abelian quasi-particles can emerge when a semiconductor nanowire with\nstrong spin-orbit coupling is brought in contact with a superconductor. To\nfully exploit the potential of non-Abelian anyons for topological quantum\ncomputing, they need to be exchanged in a well-controlled braiding operation.\nEssential hardware for braiding is a network of single-crystalline nanowires\ncoupled to superconducting islands. Here, we demonstrate a technique for\ngeneric bottom-up synthesis of complex quantum devices with a special focus on\nnanowire networks having a predefined number of superconducting islands.\nStructural analysis confirms the high crystalline quality of the nanowire\njunctions, as well as an epitaxial superconductor-semiconductor interface.\nQuantum transport measurements of nanowire \"hashtags\" reveal Aharonov-Bohm and\nweak-antilocalization effects, indicating a phase coherent system with strong\nspin-orbit coupling. In addition, a proximity-induced hard superconducting gap\nis demonstrated in these hybrid superconductor-semiconductor nanowires,\nhighlighting the successful materials development necessary for a first\nbraiding experiment. Our approach opens new avenues for the realization of\nepitaxial 3-dimensional quantum device architectures.\n",
"title": "Epitaxy of Advanced Nanowire Quantum Devices"
}
| null | null | null | null | true | null |
14605
| null |
Default
| null | null |
null |
{
"abstract": " In the setting of finite type invariants for null-homologous knots in\nrational homology 3-spheres with respect to null Lagrangian-preserving\nsurgeries, there are two candidates to be universal invariants, defined\nrespectively by Kricker and Lescop. In a previous paper, the second author\ndefined maps between spaces of Jacobi diagrams. Injectivity for these maps\nwould imply that Kricker and Lescop invariants are indeed universal invariants;\nthis would prove in particular that these two invariants are equivalent. In the\npresent paper, we investigate the injectivity status of these maps for degree 2\ninvariants, in the case of knots whose Blanchfield modules are direct sums of\nisomorphic Blanchfield modules of Q-- dimension two. We prove that they are\nalways injective except in one case, for which we determine explicitly the\nkernel.\n",
"title": "Toward universality in degree 2 of the Kricker lift of the Kontsevich integral and the Lescop equivariant invariant"
}
| null | null | null | null | true | null |
14606
| null |
Default
| null | null |
null |
{
"abstract": " Automatic segmentation in MR brain images is important for quantitative\nanalysis in large-scale studies with images acquired at all ages.\nThis paper presents a method for the automatic segmentation of MR brain\nimages into a number of tissue classes using a convolutional neural network. To\nensure that the method obtains accurate segmentation details as well as spatial\nconsistency, the network uses multiple patch sizes and multiple convolution\nkernel sizes to acquire multi-scale information about each voxel. The method is\nnot dependent on explicit features, but learns to recognise the information\nthat is important for the classification based on training data. The method\nrequires a single anatomical MR image only.\nThe segmentation method is applied to five different data sets: coronal\nT2-weighted images of preterm infants acquired at 30 weeks postmenstrual age\n(PMA) and 40 weeks PMA, axial T2- weighted images of preterm infants acquired\nat 40 weeks PMA, axial T1-weighted images of ageing adults acquired at an\naverage age of 70 years, and T1-weighted images of young adults acquired at an\naverage age of 23 years. The method obtained the following average Dice\ncoefficients over all segmented tissue classes for each data set, respectively:\n0.87, 0.82, 0.84, 0.86 and 0.91.\nThe results demonstrate that the method obtains accurate segmentations in all\nfive sets, and hence demonstrates its robustness to differences in age and\nacquisition protocol.\n",
"title": "Automatic segmentation of MR brain images with a convolutional neural network"
}
| null | null | null | null | true | null |
14607
| null |
Default
| null | null |
null |
{
"abstract": " We present a brief review on integrability of multispecies zero range process\nin one dimension introduced recently. The topics range over stochastic $R$\nmatrices of quantum affine algebra $U_q (A^{(1)}_n)$, matrix product\nconstruction of stationary states for periodic systems, $q$-boson\nrepresentation of Zamolodchikov-Faddeev algebra, etc. We also introduce new\ncommuting Markov transfer matrices having a mixed boundary condition and prove\nthe factorization of a family of $R$ matrices associated with the tetrahedron\nequation and generalized quantum groups at a special point of the spectral\nparameter.\n",
"title": "Integrable Structure of Multispecies Zero Range Process"
}
| null | null |
[
"Physics",
"Mathematics"
] | null | true | null |
14608
| null |
Validated
| null | null |
null |
{
"abstract": " Clinical trial registries can be used to monitor the production of trial\nevidence and signal when systematic reviews become out of date. However, this\nuse has been limited to date due to the extensive manual review required to\nsearch for and screen relevant trial registrations. Our aim was to evaluate a\nnew method that could partially automate the identification of trial\nregistrations that may be relevant for systematic review updates. We identified\n179 systematic reviews of drug interventions for type 2 diabetes, which\nincluded 537 clinical trials that had registrations in ClinicalTrials.gov. We\ntested a matrix factorisation approach that uses a shared latent space to learn\nhow to rank relevant trial registrations for each systematic review, comparing\nthe performance to document similarity to rank relevant trial registrations.\nThe two approaches were tested on a holdout set of the newest trials from the\nset of type 2 diabetes systematic reviews and an unseen set of 141 clinical\ntrial registrations from 17 updated systematic reviews published in the\nCochrane Database of Systematic Reviews. The matrix factorisation approach\noutperformed the document similarity approach with a median rank of 59 and\nrecall@100 of 60.9%, compared to a median rank of 138 and recall@100 of 42.8%\nin the document similarity baseline. In the second set of systematic reviews\nand their updates, the highest performing approach used document similarity and\ngave a median rank of 67 (recall@100 of 62.9%). The proposed method was useful\nfor ranking trial registrations to reduce the manual workload associated with\nfinding relevant trials for systematic review updates. The results suggest that\nthe approach could be used as part of a semi-automated pipeline for monitoring\npotentially new evidence for inclusion in a review update.\n",
"title": "A shared latent space matrix factorisation method for recommending new trial evidence for systematic review updates"
}
| null | null | null | null | true | null |
14609
| null |
Default
| null | null |
null |
{
"abstract": " We adapt the well-known spectral decimation technique for computing spectra\nof Laplacians on certain symmetric self-similar sets to the case of magnetic\nSchrodinger operators and work through this method completely for the diamond\nlattice fractal. This connects results of physicists from the 1980's, who used\nsimilar techniques to compute spectra of sequences of magnetic operators on\ngraph approximations to fractals but did not verify existence of a limiting\nfractal operator, to recent work describing magnetic operators on fractals via\nfunctional analytic techniques.\n",
"title": "Spectra of Magnetic Operators on the Diamond Lattice Fractal"
}
| null | null | null | null | true | null |
14610
| null |
Default
| null | null |
null |
{
"abstract": " We develop a calculus for diagrams of knotted objects. We define Arrow\npresentations, which encode the crossing informations of a diagram into arrows\nin a way somewhat similar to Gauss diagrams, and more generally w-tree\npresentations, which can be seen as `higher order Gauss diagrams'. This Arrow\ncalculus is used to develop an analogue of Habiro's clasper theory for welded\nknotted objects, which contain classical link diagrams as a subset. This\nprovides a 'realization' of Polyak's algebra of arrow diagrams at the welded\nlevel, and leads to a characterization of finite type invariants of welded\nknots and long knots. As a corollary, we recover several topological results\ndue to K. Habiro and A. Shima and to T. Watanabe on knotted surfaces in\n4-space. We also classify welded string links up to homotopy, thus recovering a\nresult of the first author with B. Audoux, P. Bellingeri and E. Wagner.\n",
"title": "Arrow calculus for welded and classical links"
}
| null | null | null | null | true | null |
14611
| null |
Default
| null | null |
null |
{
"abstract": " Schmidt's game, and other similar intersection games have played an important\nrole in recent years in applications to number theory, dynamics, and\nDiophantine approximation theory. These games are real games, that is, games in\nwhich the players make moves from a complete separable metric space. The\ndeterminacy of these games trivially follows from the axiom of determinacy for\nreal games, $\\mathsf{AD}_\\mathbb{R}$, which is a much stronger axiom than that\nasserting all integer games are determined, $\\mathsf{AD}$. One of our main\nresults is a general theorem which under the hypothesis $\\mathsf{AD}$ implies\nthe determinacy of intersection games which have a property allowing strategies\nto be simplified. In particular, we show that Schmidt's $(\\alpha,\\beta,\\rho)$\ngame on $\\mathbb{R}$ is determined from $\\mathsf{AD}$ alone, but on\n$\\mathbb{R}^n$ for $n \\geq 3$ we show that $\\mathsf{AD}$ does not imply the\ndeterminacy of this game. We also prove several other results specifically\nrelated to the determinacy of Schmidt's game. These results highlight the\nobstacles in obtaining the determinacy of Schmidt's game from $\\mathsf{AD}$.\n",
"title": "Determinacy of Schmidt's Game and Other Intersection Games"
}
| null | null | null | null | true | null |
14612
| null |
Default
| null | null |
null |
{
"abstract": " Kernel $k$-means clustering can correctly identify and extract a far more\nvaried collection of cluster structures than the linear $k$-means clustering\nalgorithm. However, kernel $k$-means clustering is computationally expensive\nwhen the non-linear feature map is high-dimensional and there are many input\npoints. Kernel approximation, e.g., the Nyström method, has been applied in\nprevious works to approximately solve kernel learning problems when both of the\nabove conditions are present. This work analyzes the application of this\nparadigm to kernel $k$-means clustering, and shows that applying the linear\n$k$-means clustering algorithm to $\\frac{k}{\\epsilon} (1 + o(1))$ features\nconstructed using a so-called rank-restricted Nyström approximation results\nin cluster assignments that satisfy a $1 + \\epsilon$ approximation ratio in\nterms of the kernel $k$-means cost function, relative to the guarantee provided\nby the same algorithm without the use of the Nyström method. As part of the\nanalysis, this work establishes a novel $1 + \\epsilon$ relative-error trace\nnorm guarantee for low-rank approximation using the rank-restricted Nyström\napproximation. Empirical evaluations on the $8.1$ million instance MNIST8M\ndataset demonstrate the scalability and usefulness of kernel $k$-means\nclustering with Nyström approximation. This work argues that spectral\nclustering using Nyström approximation---a popular and computationally\nefficient, but theoretically unsound approach to non-linear clustering---should\nbe replaced with the efficient and theoretically sound combination of kernel\n$k$-means clustering with Nyström approximation. The superior performance of\nthe latter approach is empirically verified.\n",
"title": "Scalable Kernel K-Means Clustering with Nystrom Approximation: Relative-Error Bounds"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
14613
| null |
Validated
| null | null |
null |
{
"abstract": " The classification of time series data is a challenge common to all\ndata-driven fields. However, there is no agreement about which are the most\nefficient techniques to group unlabeled time-ordered data. This is because a\nsuccessful classification of time series patterns depends on the goal and the\ndomain of interest, i.e. it is application-dependent.\nIn this article, we study free-to-play game data. In this domain, clustering\nsimilar time series information is increasingly important due to the large\namount of data collected by current mobile and web applications. We evaluate\nwhich methods cluster accurately time series of mobile games, focusing on\nplayer behavior data. We identify and validate several aspects of the\nclustering: the similarity measures and the representation techniques to reduce\nthe high dimensionality of time series. As a robustness test, we compare\nvarious temporal datasets of player activity from two free-to-play video-games.\nWith these techniques we extract temporal patterns of player behavior\nrelevant for the evaluation of game events and game-business diagnosis. Our\nexperiments provide intuitive visualizations to validate the results of the\nclustering and to determine the optimal number of clusters. Additionally, we\nassess the common characteristics of the players belonging to the same group.\nThis study allows us to improve the understanding of player dynamics and churn\nbehavior.\n",
"title": "Discovering Playing Patterns: Time Series Clustering of Free-To-Play Game Data"
}
| null | null | null | null | true | null |
14614
| null |
Default
| null | null |
null |
{
"abstract": " This paper is mainly inspired by the conjecture about the existence of bound\nstates for magnetic Neumann Laplacians on planar wedges of any aperture\n$\\phi\\in (0,\\pi)$. So far, a proof was only obtained for apertures\n$\\phi\\lesssim 0.511\\pi$. The conviction in the validity of this conjecture for\napertures $\\phi\\gtrsim 0.511\\pi$ mainly relied on numerical computations. In\nthis paper we succeed to prove the existence of bound states for any aperture\n$\\phi \\lesssim 0.583\\pi$ using a variational argument with suitably chosen test\nfunctions. Employing some more involved test functions and combining a\nvariational argument with computer-assistance, we extend this interval up to\nany aperture $\\phi \\lesssim 0.595\\pi$. Moreover, we analyse the same question\nfor closely related problems concerning magnetic Robin Laplacians on wedges and\nfor magnetic Schrödinger operators in the plane with $\\delta$-interactions\nsupported on broken lines.\n",
"title": "On the bound states of magnetic Laplacians on wedges"
}
| null | null | null | null | true | null |
14615
| null |
Default
| null | null |
null |
{
"abstract": " This paper proposes distributed algorithms for multi-agent networks to\nachieve a solution in finite time to a linear equation $Ax=b$ where $A$ has\nfull row rank, and with the minimum $l_1$-norm in the underdetermined case\n(where $A$ has more columns than rows). The underlying network is assumed to be\nundirected and fixed, and an analytical proof is provided for the proposed\nalgorithm to drive all agents' individual states to converge to a common value,\nviz a solution of $Ax=b$, which is the minimum $l_1$-norm solution in the\nunderdetermined case. Numerical simulations are also provided as validation of\nthe proposed algorithms.\n",
"title": "Finite-Time Distributed Linear Equation Solver for Minimum $l_1$ Norm Solutions"
}
| null | null | null | null | true | null |
14616
| null |
Default
| null | null |
null |
{
"abstract": " Product distribution matching (PDM) is proposed to generate target\ndistributions over large alphabets by combining the output of several parallel\ndistribution matchers (DMs) with smaller output alphabets. The parallel\narchitecture of PDM enables low-complexity and high-throughput implementation.\nPDM is used as a shaping device for probabilistic amplitude shaping (PAS). For\n64-ASK and a spectral efficiency of 4.5 bits per channel use (bpcu), PDM is as\npower efficient as a single full-fledged DM. It is shown how PDM enables PAS\nfor parallel channels present in multi-carrier systems like digital subscriber\nline (DSL) and orthogonal frequency-division multiplexing (OFDM). The key\nfeature is that PDM shares the DMs for lower bit-levels among different\nsub-carriers, which improves the power efficiency significantly. A\nrepresentative parallel channel example shows that PAS with PDM is 0.93 dB more\npower efficient than conventional uniform signaling and PDM is 0.35 dB more\npower efficient than individual per channel DMs.\n",
"title": "High Throughput Probabilistic Shaping with Product Distribution Matching"
}
| null | null | null | null | true | null |
14617
| null |
Default
| null | null |
null |
{
"abstract": " Last decade witnesses significant methodological and theoretical advances in\nestimating large precision matrices. In particular, there are scientific\napplications such as longitudinal data, meteorology and spectroscopy in which\nthe ordering of the variables can be interpreted through a bandable structure\non the Cholesky factor of the precision matrix. However, the minimax theory has\nstill been largely unknown, as opposed to the well established minimax results\nover the corresponding bandable covariance matrices. In this paper, we focus on\ntwo commonly used types of parameter spaces, and develop the optimal rates of\nconvergence under both the operator norm and the Frobenius norm. A striking\nphenomenon is found: two types of parameter spaces are fundamentally different\nunder the operator norm but enjoy the same rate optimality under the Frobenius\nnorm, which is in sharp contrast to the equivalence of corresponding two types\nof bandable covariance matrices under both norms. This fundamental difference\nis established by carefully constructing the corresponding minimax lower\nbounds. Two new estimation procedures are developed: for the operator norm, our\noptimal procedure is based on a novel local cropping estimator targeting on all\nprinciple submatrices of the precision matrix while for the Frobenius norm, our\noptimal procedure relies on a delicate regression-based block-thresholding\nrule. We further establish rate optimality in the nonparanormal model.\nNumerical studies are carried out to confirm our theoretical findings.\n",
"title": "Minimax Estimation of Large Precision Matrices with Bandable Cholesky Factor"
}
| null | null | null | null | true | null |
14618
| null |
Default
| null | null |
null |
{
"abstract": " Machine learning-guided protein engineering is a new paradigm that enables\nthe optimization of complex protein functions. Machine-learning methods use\ndata to predict protein function without requiring a detailed model of the\nunderlying physics or biological pathways. They accelerate protein engineering\nby learning from information contained in all measured variants and using it to\nselect variants that are likely to be improved. In this review, we introduce\nthe steps required to collect protein data, train machine-learning models, and\nuse trained models to guide engineering. We make recommendations at each stage\nand look to future opportunities for machine learning to enable the discovery\nof new protein functions and uncover the relationship between protein sequence\nand function.\n",
"title": "Machine learning in protein engineering"
}
| null | null |
[
"Quantitative Biology"
] | null | true | null |
14619
| null |
Validated
| null | null |
null |
{
"abstract": " The inner structure of a material is called microstructure. It stores the\ngenesis of a material and determines all its physical and chemical properties.\nWhile microstructural characterization is widely spread and well known, the\nmicrostructural classification is mostly done manually by human experts, which\ngives rise to uncertainties due to subjectivity. Since the microstructure could\nbe a combination of different phases or constituents with complex substructures\nits automatic classification is very challenging and only a few prior studies\nexist. Prior works focused on designed and engineered features by experts and\nclassified microstructures separately from the feature extraction step.\nRecently, Deep Learning methods have shown strong performance in vision\napplications by learning the features from data together with the\nclassification step. In this work, we propose a Deep Learning method for\nmicrostructural classification in the examples of certain microstructural\nconstituents of low carbon steel. This novel method employs pixel-wise\nsegmentation via Fully Convolutional Neural Networks (FCNN) accompanied by a\nmax-voting scheme. Our system achieves 93.94% classification accuracy,\ndrastically outperforming the state-of-the-art method of 48.89% accuracy.\nBeyond the strong performance of our method, this line of research offers a\nmore robust and first of all objective way for the difficult task of steel\nquality appreciation.\n",
"title": "Advanced Steel Microstructural Classification by Deep Learning Methods"
}
| null | null | null | null | true | null |
14620
| null |
Default
| null | null |
null |
{
"abstract": " Utilizing the Hirsch index h and some of its variants for an exploratory\nfactor analysis we discuss whether one of the most important Hirsch-type\nindices, namely the g-index comprises information about not only the size of\nthe productive core but also the impact of the papers in the core. We also\nstudy the effect of logarithmic and square-root transformation of the data\nutilized in the factor analysis. To demonstrate our approach we use a real data\nexample analysing the citation records of 26 physicists compiled from the Web\nof Science.\n",
"title": "Categorizing Hirsch Index Variants"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
14621
| null |
Validated
| null | null |
null |
{
"abstract": " Although Darwinian models are rampant in the social sciences, social\nscientists do not face the problem that motivated Darwin's theory of natural\nselection: the problem of explaining how lineages evolve despite that any\ntraits they acquire are regularly discarded at the end of the lifetime of the\nindividuals that acquired them. While the rationale for framing culture as an\nevolutionary process is correct, it does not follow that culture is a Darwinian\nor selectionist process, or that population genetics and phylogenetics provide\nviable starting points for modeling cultural change. This paper lays out\nstep-by-step arguments as to why this approach is ill-conceived, focusing on\nthe lack of randomness and lack of a self-assembly code in cultural evolution,\nand summarizes an alternative approach.\n",
"title": "Why a Population Genetics Framework is Inappropriate for Cultural Evolution"
}
| null | null | null | null | true | null |
14622
| null |
Default
| null | null |
null |
{
"abstract": " Collaborations are an integral part of scientific research and publishing. In\nthe past, access to large-scale corpora has limited the ways in which questions\nabout collaborations could be investigated. However, with improvements in\ndata/metadata quality and access, it is possible to explore the idea of\nresearch collaboration in ways beyond the traditional definition of multiple\nauthorship. In this paper, we examine scientific works through three different\nlenses of collaboration: across multiple authors, multiple institutions, and\nmultiple departments. We believe this to be a first look at multiple\ndepartmental collaborations as we employ extensive data curation to\ndisambiguate authors' departmental affiliations for nearly 70,000 scientific\npapers. We then compare citation metrics across the different definitions of\ncollaboration and find that papers defined as being collaborative were more\nfrequently cited than their non-collaborative counterparts, regardless of the\ndefinition of collaboration used. We also share preliminary results from\nexamining the relationship between co-citation and co-authorship by analyzing\nthe extent to which similar fields (as determined by co-citation) are\ncollaborating on works (as determined by co-authorship). These preliminary\nresults reveal trends of compartmentalization with respect to\nintra-institutional collaboration and show promise in being expanded.\n",
"title": "Is together better? Examining scientific collaborations across multiple authors, institutions, and departments"
}
| null | null | null | null | true | null |
14623
| null |
Default
| null | null |
null |
{
"abstract": " For a second order operator on a compact manifold satisfying the strong\nHörmander condition, we give a bound for the spectral gap analogous to the\nLichnerowicz estimate for the Laplacian of a Riemannian manifold. We consider a\nwide class of such operators which includes horizontal lifts of the Laplacian\non Riemannian submersions with minimal leaves.\n",
"title": "A Lichnerowicz estimate for the spectral gap of the sub-Laplacian"
}
| null | null | null | null | true | null |
14624
| null |
Default
| null | null |
null |
{
"abstract": " We develop a theory based on the formalism of quasiclassical Green's\nfunctions to study the spin dynamics in superfluid $^3$He. First, we derive\nkinetic equations for the spin-dependent distribution function in the bulk\nsuperfluid reproducing the results obtained earlier without quasiclassical\napproximation. Then we consider a spin dynamics near the surface of fully\ngapped $^3$He-B phase taking into account spin relaxation due to the\ntransitions in the spectrum of localized fermionic states. The lifetime of\nlongitudinal and transverse spin waves is calculate taking into account the\nFermi-liquid corrections which lead to the crucial modification of fermionic\nspectrum and spin responses.\n",
"title": "Quasiclassical theory of spin dynamics in superfluid $^3$He: kinetic equations in the bulk and spin response of surface Majorana states"
}
| null | null | null | null | true | null |
14625
| null |
Default
| null | null |
null |
{
"abstract": " We present a dataset with models of 14 articulated objects commonly found in\nhuman environments and with RGB-D video sequences and wrenches recorded of\nhuman interactions with them. The 358 interaction sequences total 67 minutes of\nhuman manipulation under varying experimental conditions (type of interaction,\nlighting, perspective, and background). Each interaction with an object is\nannotated with the ground truth poses of its rigid parts and the kinematic\nstate obtained by a motion capture system. For a subset of 78 sequences (25\nminutes), we also measured the interaction wrenches. The object models contain\ntextured three-dimensional triangle meshes of each link and their motion\nconstraints. We provide Python scripts to download and visualize the data. The\ndata is available at this https URL and hosted\nat this https URL.\n",
"title": "The RBO Dataset of Articulated Objects and Interactions"
}
| null | null | null | null | true | null |
14626
| null |
Default
| null | null |
null |
{
"abstract": " We relate the minimax game of generative adversarial networks (GANs) to\nfinding the saddle points of the Lagrangian function for a convex optimization\nproblem, where the discriminator outputs and the distribution of generator\noutputs play the roles of primal variables and dual variables, respectively.\nThis formulation shows the connection between the standard GAN training process\nand the primal-dual subgradient methods for convex optimization. The inherent\nconnection does not only provide a theoretical convergence proof for training\nGANs in the function space, but also inspires a novel objective function for\ntraining. The modified objective function forces the distribution of generator\noutputs to be updated along the direction according to the primal-dual\nsubgradient methods. A toy example shows that the proposed method is able to\nresolve mode collapse, which in this case cannot be avoided by the standard GAN\nor Wasserstein GAN. Experiments on both Gaussian mixture synthetic data and\nreal-world image datasets demonstrate the performance of the proposed method on\ngenerating diverse samples.\n",
"title": "Training Generative Adversarial Networks via Primal-Dual Subgradient Methods: A Lagrangian Perspective on GAN"
}
| null | null |
[
"Statistics"
] | null | true | null |
14627
| null |
Validated
| null | null |
null |
{
"abstract": " It has recently been discovered that some, if not all, classical novae emit\nGeV gamma rays during outburst, but the mechanisms involved in the production\nof the gamma rays are still not well understood. We present here a\ncomprehensive multi-wavelength dataset---from radio to X-rays---for the most\ngamma-ray luminous classical nova to-date, V1324 Sco. Using this dataset, we\nshow that V1324 Sco is a canonical dusty Fe-II type nova, with a maximum ejecta\nvelocity of 2600 km s$^{-1}$ and an ejecta mass of few $\\times 10^{-5}$\nM$_{\\odot}$. There is also evidence for complex shock interactions, including a\ndouble-peaked radio light curve which shows high brightness temperatures at\nearly times. To explore why V1324~Sco was so gamma-ray luminous, we present a\nmodel of the nova ejecta featuring strong internal shocks, and find that higher\ngamma-ray luminosities result from higher ejecta velocities and/or mass-loss\nrates. Comparison of V1324~Sco with other gamma-ray detected novae does not\nshow clear signatures of either, and we conclude that a larger sample of\nsimilarly well-observed novae is needed to understand the origin and variation\nof gamma rays in novae.\n",
"title": "A Detailed Observational Analysis of V1324 Sco, the Most Gamma-Ray Luminous Classical Nova to Date"
}
| null | null |
[
"Physics"
] | null | true | null |
14628
| null |
Validated
| null | null |
null |
{
"abstract": " The size of a planet is an observable property directly connected to the\nphysics of its formation and evolution. We used precise radius measurements\nfrom the California-Kepler Survey (CKS) to study the size distribution of 2025\n$\\textit{Kepler}$ planets in fine detail. We detect a factor of $\\geq$2 deficit\nin the occurrence rate distribution at 1.5-2.0 R$_{\\oplus}$. This gap splits\nthe population of close-in ($P$ < 100 d) small planets into two size regimes:\nR$_P$ < 1.5 R$_{\\oplus}$ and R$_P$ = 2.0-3.0 R$_{\\oplus}$, with few planets in\nbetween. Planets in these two regimes have nearly the same intrinsic frequency\nbased on occurrence measurements that account for planet detection\nefficiencies. The paucity of planets between 1.5 and 2.0 R$_{\\oplus}$ supports\nthe emerging picture that close-in planets smaller than Neptune are composed of\nrocky cores measuring 1.5 R$_{\\oplus}$ or smaller with varying amounts of\nlow-density gas that determine their total sizes.\n",
"title": "The California-Kepler Survey. III. A Gap in the Radius Distribution of Small Planets"
}
| null | null | null | null | true | null |
14629
| null |
Default
| null | null |
null |
{
"abstract": " We prove that for any dimension function $h$ with $h \\prec x^d$ and for any\ncountable set of linear patterns, there exists a compact set $E$ with\n$\\mathcal{H}^h(E)>0$ avoiding all the given patterns. We also give several\napplications and recover results of Keleti, Maga, and Máthé.\n",
"title": "Large sets avoiding linear patterns"
}
| null | null | null | null | true | null |
14630
| null |
Default
| null | null |
null |
{
"abstract": " Pushdown systems (PDSs) and recursive state machines (RSMs), which are\nlinearly equivalent, are standard models for interprocedural analysis. Yet RSMs\nare more convenient as they (a) explicitly model function calls and returns,\nand (b) specify many natural parameters for algorithmic analysis, e.g., the\nnumber of entries and exits. We consider a general framework where RSM\ntransitions are labeled from a semiring and path properties are algebraic with\nsemiring operations, which can model, e.g., interprocedural reachability and\ndataflow analysis problems.\nOur main contributions are new algorithms for several fundamental problems.\nAs compared to a direct translation of RSMs to PDSs and the best-known existing\nbounds of PDSs, our analysis algorithm improves the complexity for\nfinite-height semirings (that subsumes reachability and standard dataflow\nproperties). We further consider the problem of extracting distance values from\nthe representation structures computed by our algorithm, and give efficient\nalgorithms that distinguish the complexity of a one-time preprocessing from the\ncomplexity of each individual query. Another advantage of our algorithm is that\nour improvements carry over to the concurrent setting, where we improve the\nbest-known complexity for the context-bounded analysis of concurrent RSMs.\nFinally, we provide a prototype implementation that gives a significant\nspeed-up on several benchmarks from the SLAM/SDV project.\n",
"title": "Faster Algorithms for Weighted Recursive State Machines"
}
| null | null | null | null | true | null |
14631
| null |
Default
| null | null |
null |
{
"abstract": " We discuss the average-field approximation for a trapped gas of\nnon-interacting anyons in the quasi-bosonic regime. In the homogeneous case,\ni.e., for a confinement to a bounded region, we prove that the energy in the\nregime of large statistics parameter, i.e., for \"less-bosonic\" anyons, is\nindependent of boundary conditions and of the shape of the domain. When a\nnon-trivial trapping potential is present, we derive a local density\napproximation in terms of a Thomas-Fermi-like model.\n",
"title": "Local Density Approximation for Almost-Bosonic Anyons"
}
| null | null | null | null | true | null |
14632
| null |
Default
| null | null |
null |
{
"abstract": " In chemical or physical reaction dynamics, it is essential to distinguish\nprecisely between reactants and products for all time. This task is especially\ndemanding in time-dependent or driven systems because therein the dividing\nsurface (DS) between these states often exhibits a nontrivial time-dependence.\nThe so-called transition state (TS) trajectory has been seen to define a DS\nwhich is free of recrossings in a large number of one-dimensional reactions\nacross time-dependent barriers, and, thus, allows one to determine exact\nreaction rates. A fundamental challenge to applying this method is the\nconstruction of the TS trajectory itself. The minimization of Lagrangian\ndescriptors (LDs) provides a general and powerful scheme to obtain that\ntrajectory even when perturbation theory fails. Both approaches encounter\npossible breakdowns when the overall potential is bounded, admitting the\npossibility of returns to the barrier long after trajectories have reached the\nproduct or reactant wells. Such global dynamics cannot be captured by\nperturbation theory. Meanwhile, in the LD-DS approach, it leads to the\nemergence of additional local minima which make it difficult to extract the\noptimal branch associated with the desired TS trajectory. In this work, we\nillustrate this behavior for a time-dependent double-well potential revealing a\nself-similar structure of the LD, and we demonstrate how the reflections and\nside-minima can be addressed by an appropriate modification of the LD\nassociated with the direct rate across the barrier.\n",
"title": "Chemical dynamics between wells across a time-dependent barrier: Self-similarity in the Lagrangian descriptor and reactive basins"
}
| null | null | null | null | true | null |
14633
| null |
Default
| null | null |
null |
{
"abstract": " The radial drift problem constitutes one of the most fundamental problems in\nplanet formation theory, as it predicts particles to drift into the star before\nthey are able to grow to planetesimal size. Dust-trapping vortices have been\nproposed as a possible solution to this problem, as they might be able to trap\nparticles over millions of years, allowing them to grow beyond the radial drift\nbarrier. Here, we present ALMA 0.04\"-resolution imaging of the pre-transitional\ndisk of V1247 Orionis that reveals an asymmetric ring as well as a\nsharply-confined crescent structure, resembling morphologies seen in\ntheoretical models of vortex formation. The asymmetric ring (at 0.17\"=54 au\nseparation from the star) and the crescent (at 0.38\"=120 au) seem smoothly\nconnected through a one-armed spiral arm structure that has been found\npreviously in scattered light. We propose a physical scenario with a planet\norbiting at $\\sim0.3$\"$\\approx$100 au, where the one-armed spiral arm detected\nin polarised light traces the accretion stream feeding the protoplanet. The\ndynamical influence of the planet clears the gap between the ring and the\ncrescent and triggers two vortices that trap mm-sized particles, namely the\ncrescent and the bright asymmetry seen in the ring. We conducted dedicated\nhydrodynamics simulations of a disk with an embedded planet, which results in\nsimilar spiral-arm morphologies as seen in our scattered light images. At the\nposition of the spiral wake and the crescent we also observe $^{12}$CO (3-2)\nand H$^{12}$CO$^{+}$ (4-3) excess line emission, likely tracing the increased\nscale-height in these disk regions.\n",
"title": "Dust-trapping vortices and a potentially planet-triggered spiral wake in the pre-transitional disk of V1247 Orionis"
}
| null | null |
[
"Physics"
] | null | true | null |
14634
| null |
Validated
| null | null |
null |
{
"abstract": " As a counterpart of the classical Yamabe problem, a fractional Yamabe flow\nhas been introduced by Jin and Xiong (2014) on the sphere. Here we pursue its\nstudy in the context of general compact smooth manifolds with positive\nfractional curvature. First, we prove that the flow is locally well posed in\nthe weak sense on any compact manifold. If the manifold is locally conformally\nflat with positive Yamabe invariant, we also prove that the flow is smooth and\nconverges to a constant scalar curvature metric. We provide different proofs\nusing extension properties introduced by Chang and González (2011) for the\nconformally covariant fractional order operators.\n",
"title": "Weak and smooth solutions for a fractional Yamabe flow: the case of general compact and locally conformally flat manifolds"
}
| null | null | null | null | true | null |
14635
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we discuss the stability properties of convolutional neural\nnetworks. Convolutional neural networks are widely used in machine learning. In\nclassification they are mainly used as feature extractors. Ideally, we expect\nsimilar features when the inputs are from the same class. That is, we hope to\nsee a small change in the feature vector with respect to a deformation on the\ninput signal. This can be established mathematically, and the key step is to\nderive the Lipschitz properties. Further, we establish that the stability\nresults can be extended for more general networks. We give a formula for\ncomputing the Lipschitz bound, and compare it with other methods to show it is\ncloser to the optimal value.\n",
"title": "Lipschitz Properties for Deep Convolutional Networks"
}
| null | null |
[
"Computer Science",
"Mathematics"
] | null | true | null |
14636
| null |
Validated
| null | null |
null |
{
"abstract": " Compound distributions allow construction of a rich set of distributions.\nTypically they involve an intractable integral. Here we use a quadrature\napproximation to that integral to define the quadrature compound family.\nSpecial care is taken that this approximation is suitable for computation of\ngradients with respect to distribution parameters. This technique is applied to\ndiscrete (Poisson LogNormal) and continuous distributions. In the continuous\ncase, quadrature compound family naturally makes use of parameterized\ntransformations of unparameterized distributions (a.k.a \"reparameterization\"),\nallowing for gradients of expectations to be estimated as the gradient of a\nsample mean. This is demonstrated in a novel distribution, the diffeomixture,\nwhich is is a reparameterizable approximation to a mixture distribution.\n",
"title": "Quadrature Compound: An approximating family of distributions"
}
| null | null | null | null | true | null |
14637
| null |
Default
| null | null |
null |
{
"abstract": " Identifying influential nodes in a network is a fundamental issue due to its\nwide applications, such as accelerating information diffusion or halting virus\nspreading. Many measures based on the network topology have emerged over the\nyears to identify influential nodes such as Betweenness, Closeness, and\nEigenvalue centrality. However, although most real-world networks are modular,\nfew measures exploit this property. Recent works have shown that it has a\nsignificant effect on the dynamics on networks. In a modular network, a node\nhas two types of influence: a local influence (on the nodes of its community)\nthrough its intra-community links and a global influence (on the nodes in other\ncommunities) through its inter-community links. Depending of the strength of\nthe community structure, these two components are more or less influential.\nBased on this idea, we propose to extend all the standard centrality measures\ndefined for networks with no community structure to modular networks. The\nso-called \"Modular centrality\" is a two dimensional vector. Its first component\nquantifies the local influence of a node in its community while the second\ncomponent quantifies its global influence on the other communities of the\nnetwork. In order to illustrate the effectiveness of the Modular centrality\nextensions, comparison with their scalar counterpart are performed in an\nepidemic process setting. Simulation results using the\nSusceptible-Infected-Recovered (SIR) model on synthetic networks with\ncontrolled community structure allows getting a clear idea about the relation\nbetween the strength of the community structure and the major type of influence\n(global/local). Furthermore, experiments on real-world networks demonstrate the\nmerit of this approach.\n",
"title": "Centrality in Modular Networks"
}
| null | null | null | null | true | null |
14638
| null |
Default
| null | null |
null |
{
"abstract": " For population systems modeled by age-structured hyperbolic partial\ndifferential equations (PDEs) that are bilinear in the input and evolve with a\npositive-valued infinite-dimensional state, global stabilization of constant\nyield set points was achieved in prior work. Seasonal demands in\nbiotechnological production processes give rise to time-varying yield\nreferences. For the proposed control objective aiming at a global attractivity\nof desired yield trajectories, multiple non-standard features have to be\nconsidered: a non-local boundary condition, a PDE state restricted to the\npositive orthant of the function space and arbitrary restrictive but physically\nmeaningful input constraints. Moreover, we provide Control Lyapunov Functionals\nensuring an exponentially fast attraction of adequate reference trajectories.\nTo achieve this goal, we make use of the relation between first-order\nhyperbolic PDEs and integral delay equations leading to a decoupling of the\ninput-dependent dynamics and the infinite-dimensional internal one.\nFurthermore, the dynamic control structure does not necessitate exact knowledge\nof the model parameters or online measurements of the age-profile. With a\nGalerkin-based numerical simulation scheme using the key ideas of the\nKarhunen-Loève-decomposition, we demonstrate the controller's performance.\n",
"title": "Yield Trajectory Tracking for Hyperbolic Age-Structured Population Systems"
}
| null | null |
[
"Computer Science",
"Mathematics"
] | null | true | null |
14639
| null |
Validated
| null | null |
null |
{
"abstract": " We provide an explicit presentation of an infinite hyperbolic Kazhdan group\nwith $4$ generators and $16$ relators of length at most $73$. That group acts\nproperly and cocompactly on a hyperbolic triangle building of type $(3,4,4)$.\nWe also point out a variation of the construction that yields examples of\nlattices in $\\tilde A_2$-buildings admitting non-Desarguesian residues of\narbitrary prime power order.\n",
"title": "A sixteen-relator presentation of an infinite hyperbolic Kazhdan group"
}
| null | null | null | null | true | null |
14640
| null |
Default
| null | null |
null |
{
"abstract": " Bayesian hierarchical models are increasingly popular for realistic modelling\nand analysis of complex data. This trend is accompanied by the need for\nflexible, general, and computationally efficient methods for model criticism\nand conflict detection. Usually, a Bayesian hierarchical model incorporates a\ngrouping of the individual data points, for example individuals in repeated\nmeasurement data. In such cases, the following question arises: Are any of the\ngroups \"outliers\", or in conflict with the remaining groups? Existing general\napproaches aiming to answer such questions tend to be extremely computationally\ndemanding when model fitting is based on MCMC. We show how group-level model\ncriticism and conflict detection can be done quickly and accurately through\nintegrated nested Laplace approximations (INLA). The new method is implemented\nas a part of the open source R-INLA package for Bayesian computing\n(this http URL).\n",
"title": "Fast and accurate Bayesian model criticism and conflict diagnostics using R-INLA"
}
| null | null |
[
"Mathematics",
"Statistics"
] | null | true | null |
14641
| null |
Validated
| null | null |
null |
{
"abstract": " The comparison of observed brain activity with the statistics generated by\nartificial intelligence systems is useful to probe brain functional\norganization under ecological conditions. Here we study fMRI activity in ten\nsubjects watching color natural movies and compute deep representations of\nthese movies with an architecture that relies on optical flow and image\ncontent. The association of activity in visual areas with the different layers\nof the deep architecture displays complexity-related contrasts across visual\nareas and reveals a striking foveal/peripheral dichotomy.\n",
"title": "Optimizing deep video representation to match brain activity"
}
| null | null | null | null | true | null |
14642
| null |
Default
| null | null |
null |
{
"abstract": " There has been a recent surge of interest in dualities relating theories of\nChern-Simons gauge fields coupled to either bosons or fermions within the\ncondensed matter community, particularly in the context of topological\ninsulators and the half-filled Landau level. Here, we study the application of\none such duality to the long-standing problem of quantum Hall inter-plateaux\ntransitions. The key motivating experimental observations are the anomalously\nlarge value of the correlation length exponent $\\nu \\approx 2.3$ and that $\\nu$\nis observed to be super-universal, i.e., the same in the vicinity of distinct\ncritical points. Duality motivates effective descriptions for a fractional\nquantum Hall plateau transition involving a Chern-Simons field with $U(N_c)$\ngauge group coupled to $N_f = 1$ fermion. We study one class of theories in a\ncontrolled limit where $N_f \\gg N_c$ and calculate $\\nu$ to leading non-trivial\norder in the absence of disorder. Although these theories do not yield an\nanomalously large exponent $\\nu$ within the large $N_f \\gg N_c$ expansion, they\ndo offer a new parameter space of theories that is apparently different from\nprior works involving abelian Chern-Simons gauge fields.\n",
"title": "Non-Abelian Fermionization and Fractional Quantum Hall Transitions"
}
| null | null | null | null | true | null |
14643
| null |
Default
| null | null |
null |
{
"abstract": " We consider a Kepler problem in dimension two or three, with a time-dependent\n$T$-periodic perturbation. We prove that for any prescribed positive integer\n$N$, there exist at least $N$ periodic solutions (with period $T$) as long as\nthe perturbation is small enough. Here the solutions are understood in a\ngeneral sense as they can have collisions. The concept of generalized solutions\nis defined intrinsically and it coincides with the notion obtained in Celestial\nMechanics via the theory of regularization of collisions.\n",
"title": "Periodic solutions and regularization of a Kepler problem with time-dependent perturbation"
}
| null | null | null | null | true | null |
14644
| null |
Default
| null | null |
null |
{
"abstract": " Slow light propagation in structured materials is a highly promising approach\nfor realizing on-chip integrated photonic devices based on enhanced optical\nnonlinearities. One of the most successful research avenues consists in\nengineering the band dispersion of light-guiding photonic crystal (PC)\nstructures. The primary goal of such devices is to achieve slow-light operation\nover the largest possible bandwidth, with large group index, minimal index\ndispersion, and constant transmission spectrum. Here, we report on the\nexperimental demonstration of to date record high GBP in silicon-based\ncoupled-cavity waveguides (CCWs) operating at telecom wavelengths. Our results\nrely on novel CCW designs, optimized using a genetic algorithm, and refined\nnanofabrication processes.\n",
"title": "Ultra-wide-band slow light in photonic crystal coupled-cavity waveguides"
}
| null | null | null | null | true | null |
14645
| null |
Default
| null | null |
null |
{
"abstract": " High-resolution imaging has delivered new prospects for detecting the\nmaterial composition and structure of cultural treasures. Despite the various\ntechniques for analysis, a significant diagnostic gap remained in the range of\navailable research capabilities for works on paper. Old master drawings were\nmostly composed in a multi-step manner with various materials. This resulted in\nthe overlapping of different layers which made the subjacent strata difficult\nto differentiate. The separation of stratified layers using imaging methods\ncould provide insights into the artistic work processes and help answer\nquestions about the object, its attribution, or in identifying forgeries. The\npattern recognition procedure was tested with mock replicas to achieve the\nseparation and the capability of displaying concealed red chalk under ink. In\ncontrast to RGB-sensor based imaging, the multi- or hyperspectral technology\nallows accurate layer separation by recording the characteristic signatures of\nthe material's reflectance. The risk of damage to the artworks as a result of\nthe examination can be reduced by using combinations of defined spectra for\nlightning and image capturing. By guaranteeing the maximum level of\nreadability, our results suggest that the technique can be applied to a broader\nrange of objects and assist in diagnostic research into cultural treasures in\nthe future.\n",
"title": "Sketch Layer Separation in Multi-Spectral Historical Document Images"
}
| null | null | null | null | true | null |
14646
| null |
Default
| null | null |
null |
{
"abstract": " In a recent paper [A. Alberucci, C. Jisha, N. Smyth, and G. Assanto, Phys.\nRev. A 91, 013841 (2015)], Alberucci et al. have studied the propagation of\nbright spatial solitary waves in highly nonlocal media. We find that the main\nresults in that and related papers, concerning soliton shape and dynamics,\nbased on the accessible soliton (AS) approximation, are incorrect; the correct\nresults have already been published by others. These and other inconsistencies\nin the paper follow from the problems in applying the AS approximation in\nearlier papers by the group that propagated to the later papers. The accessible\nsoliton theory cannot describe accurately the features and dynamics of solitons\nin highly nonlocal media.\n",
"title": "Comment on \"Spatial optical solitons in highly nonlocal media\" and related papers"
}
| null | null |
[
"Physics"
] | null | true | null |
14647
| null |
Validated
| null | null |
null |
{
"abstract": " Place recognition is a challenging problem in mobile robotics, especially in\nunstructured environments or under viewpoint and illumination changes. Most\nLiDAR-based methods rely on geometrical features to overcome such challenges,\nas generally scene geometry is invariant to these changes, but tend to affect\ncamera-based solutions significantly. Compared to cameras, however, LiDARs lack\nthe strong and descriptive appearance information that imaging can provide.\nTo combine the benefits of geometry and appearance, we propose coupling the\nconventional geometric information from the LiDAR with its calibrated intensity\nreturn. This strategy extracts extremely useful information in the form of a\nnew descriptor design, coined ISHOT, outperforming popular state-of-art\ngeometric-only descriptors by significant margin in our local descriptor\nevaluation. To complete the framework, we furthermore develop a probabilistic\nkeypoint voting place recognition algorithm, leveraging the new descriptor and\nyielding sublinear place recognition performance. The efficacy of our approach\nis validated in challenging global localization experiments in large-scale\nbuilt-up and unstructured environments.\n",
"title": "Local Descriptor for Robust Place Recognition using LiDAR Intensity"
}
| null | null |
[
"Computer Science"
] | null | true | null |
14648
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper we assess the predictive power of the self-consistent hybrid\nfunctional scPBE0 in calculating the band gap of oxide semiconductors. The\ncomputational procedure is based on the self-consistent evaluation of the\nmixing parameter $\\alpha$ by means of an iterative calculation of the static\ndielectric constant using the perturbation expansion after discretization\n(PEAD) method and making use of the relation $\\alpha = 1/\\epsilon_{\\infty}$.\nOur materials dataset is formed by 30 compounds covering a wide range of band\ngaps and dielectric properties, and includes materials with a wide spectrum of\napplication as thermoelectrics, photocatalysis, photovoltaics, transparent\nconducting oxides, and refractory materials. Our results show that the scPBE0\nfunctional provides better band gaps than the non self-consistent hybrids PBE0\nand HSE06, but scPBE0 does not show significant improvement on the description\nof the static dielectric constants. Overall, the scPBE0 data exhibit a mean\nabsolute percentage error of 14 \\% (band gaps) and 10 \\% ($\\epsilon_\\infty$).\nFor materials with weak dielectric screening and large excitonic biding\nenergies scPBE0, unlike PBE0 and HSE06, overestimates the band gaps, but the\nvalue of the gap become very close to the experimental value when excitonic\neffects are included (e.g. for SiO$_2$). However, special caution must be given\nto the compounds with small band gaps due to the tendency of scPBE0 to\noverestimate the dielectric constant in proximity of the metallic limit.\n",
"title": "Assessing the performance of self-consistent hybrid functional for band gap calculation in oxide semiconductors"
}
| null | null | null | null | true | null |
14649
| null |
Default
| null | null |
null |
{
"abstract": " We present a smooth distributed nonlinear control law for local\nsynchronization of identical driftless kinematic agents on a Cartesian product\nof matrix Lie groups with a connected communication graph. If the agents are\ninitialized sufficiently close to one another, then synchronization is achieved\nexponentially fast. We first analyze the special case of commutative Lie groups\nand show that in exponential coordinates, the closed-loop dynamics are linear.\nWe characterize all equilibria of the network and, in the case of an\nunweighted, complete graph, characterize the settling time and conditions for\ndeadbeat performance. Using the Baker-Campbell-Hausdorff theorem, we show that,\nin a neighbourhood of the identity element, all results generalize to arbitrary\nmatrix Lie groups.\n",
"title": "Local Synchronization of Sampled-Data Systems on Lie Groups"
}
| null | null | null | null | true | null |
14650
| null |
Default
| null | null |
null |
{
"abstract": " The Milky Way bulge shows a box/peanut or X-shaped bulge (hereafter BP/X)\nwhen viewed in infrared or microwave bands. We examine orbits in an N-body\nmodel of a barred disk galaxy that is scaled to match the kinematics of the\nMilky Way (MW) bulge. We generate maps of projected stellar surface density,\nunsharp masked images, 3D excess-mass distributions (showing mass outside\nellipsoids), line-of-sight number count distributions, and 2D line-of-sight\nkinematics for the simulation as well as co-added orbit families, in order to\nidentify the orbits primarily responsible for the BP/X shape. We estimate that\nbetween 19-23\\% of the mass of the bar is associated with the BP/X shape and\nthat most bar orbits contribute to this shape which is clearly seen in\nprojected surface density maps and 3D excess mass for non-resonant box orbits,\n\"banana\" orbits, \"fish/pretzel\" orbits and \"brezel\" orbits. {We find that\nnearly all bar orbit families contribute some mass to the 3D BP/X-shape. All\nco-added orbit families show a bifurcation in stellar number count distribution\nwith heliocentric distance that resembles the bifurcation observed in red clump\nstars in the MW. However, only the box orbit family shows an increasing\nseparation of peaks with increasing galactic latitude $|b|$, similar to that\nobserved.} Our analysis shows that no single orbit family fully explains all\nthe observed features associated with the MW's BP/X shaped bulge, but\ncollectively the non-resonant boxes and various resonant boxlet orbits\ncontribute at different distances from the center to produce this feature. We\npropose that since box orbits have three incommensurable orbital fundamental\nfrequencies, their 3-dimensional shapes are highly flexible and, like Lissajous\nfigures, this family of orbits is most easily able to adapt to evolution in the\nshape of the underlying potential.\n",
"title": "On the orbits that generate the X-shape in the Milky Way bulge"
}
| null | null | null | null | true | null |
14651
| null |
Default
| null | null |
null |
{
"abstract": " This paper analyzes the impact of peer effects on electricity consumption of\na network of rational, utility-maximizing users. Users derive utility from\nconsuming electricity as well as consuming less energy than their neighbors.\nHowever, a disutility is incurred for consuming more than their neighbors. To\nmaximize the profit of the load-serving entity that provides electricity to\nsuch users, we develop a two-stage game-theoretic model, where the entity sets\nthe prices in the first stage. In the second stage, consumers decide on their\ndemand in response to the observed price set in the first stage so as to\nmaximize their utility. To this end, we derive theoretical statements under\nwhich such peer effects reduce aggregate user consumption. Further, we obtain\nexpressions for the resulting electricity consumption and profit of the load\nserving entity for the case of perfect price discrimination and a single price\nunder complete information, and approximations under incomplete information.\nSimulations suggest that exposing only a selected subset of all users to peer\neffects maximizes the entity's profit.\n",
"title": "How Peer Effects Influence Energy Consumption"
}
| null | null | null | null | true | null |
14652
| null |
Default
| null | null |
null |
{
"abstract": " The Hitomi X-ray satellite has provided the first direct measurements of the\nplasma velocity dispersion in a galaxy cluster. It finds a relatively\n\"quiescent\" gas with a line-of-sight velocity dispersion ~ 160 km/s, at 30 kpc\nto 60 kpc from the cluster center. This is surprising given the presence of\njets and X-ray cavities that indicates on-going activity and feedback from the\nactive galactic nucleus (AGN) at the cluster center. Using a set of mock Hitomi\nobservations generated from a suite of state-of-the-art cosmological cluster\nsimulations, and an isolated but higher resolution simulation of gas physics in\nthe cluster core, including the effects of cooling and AGN feedback, we examine\nthe likelihood of Hitomi detecting a cluster with the observed velocities. As\nlong as the Perseus has not experienced a major merger in the last few\ngigayears, and AGN feedback is operating in a \"gentle\" mode, we reproduce the\nlevel of gas motions observed by Hitomi. The frequent mechanical AGN feedback\ngenerates net line-of-sight velocity dispersions ~100-200 km/s, bracketing the\nvalues measured in the Perseus core. The large-scale velocity shear observed\nacross the core, on the other hand, is generated mainly by cosmic accretion\nsuch as mergers. We discuss the implications of these results for AGN feedback\nphysics and cluster cosmology and progress that needs to be made in both\nsimulations and observations, including a Hitomi re-flight and\ncalorimeter-based instruments with higher spatial resolution.\n",
"title": "Physical Origins of Gas Motions in Galaxy Cluster Cores: Interpreting Hitomi Observations of the Perseus Cluster"
}
| null | null | null | null | true | null |
14653
| null |
Default
| null | null |
null |
{
"abstract": " Auto-encoding generative adversarial networks (GANs) combine the standard GAN\nalgorithm, which discriminates between real and model-generated data, with a\nreconstruction loss given by an auto-encoder. Such models aim to prevent mode\ncollapse in the learned generative model by ensuring that it is grounded in all\nthe available training data. In this paper, we develop a principle upon which\nauto-encoders can be combined with generative adversarial networks by\nexploiting the hierarchical structure of the generative model. The underlying\nprinciple shows that variational inference can be used a basic tool for\nlearning, but with the in- tractable likelihood replaced by a synthetic\nlikelihood, and the unknown posterior distribution replaced by an implicit\ndistribution; both synthetic likelihoods and implicit posterior distributions\ncan be learned using discriminators. This allows us to develop a natural fusion\nof variational auto-encoders and generative adversarial networks, combining the\nbest of both these methods. We describe a unified objective for optimization,\ndiscuss the constraints needed to guide learning, connect to the wide range of\nexisting work, and use a battery of tests to systematically and quantitatively\nassess the performance of our method.\n",
"title": "Variational Approaches for Auto-Encoding Generative Adversarial Networks"
}
| null | null | null | null | true | null |
14654
| null |
Default
| null | null |
null |
{
"abstract": " Edwin Powel Hubble is regarded as one of the most important astronomers of\n20th century. In despite of his great contributions to the field of astronomy,\nhe never received the Nobel Prize because astronomy was not considered as the\nfield of the Nobel Prize in Physics at that era. There is an anecdote about the\nrelation between Hubble and the Nobel Prize. According to this anecdote, the\nNobel Committee decided to award the Nobel Prize in Physics in 1953 to Hubble\nas the first Nobel laureate as an astronomer (Christianson 1995). However,\nHubble was died just before its announcement, and the Nobel prize is not\nawarded posthumously. Documents of the Nobel selection committee are open after\n50 years, thus this anecdote can be verified. I confirmed that the Nobel\nselection committee endorsed Frederik Zernike as the Nobel laureate in Physics\nin 1953 on September 15th, 1953, which is 13 days before the Hubble's death in\nSeptember 28th, 1953. I also confirmed that Hubble and Henry Norris Russell\nwere nominated but they are not endorsed because the Committee concluded their\nastronomical works were not appropriate for the Nobel Prize in Physics.\n",
"title": "Verification of the anecdote about Edwin Hubble and the Nobel Prize"
}
| null | null | null | null | true | null |
14655
| null |
Default
| null | null |
null |
{
"abstract": " Fully convolutional neural networks (FCN) have been shown to achieve\nstate-of-the-art performance on the task of classifying time series sequences.\nWe propose the augmentation of fully convolutional networks with long short\nterm memory recurrent neural network (LSTM RNN) sub-modules for time series\nclassification. Our proposed models significantly enhance the performance of\nfully convolutional networks with a nominal increase in model size and require\nminimal preprocessing of the dataset. The proposed Long Short Term Memory Fully\nConvolutional Network (LSTM-FCN) achieves state-of-the-art performance compared\nto others. We also explore the usage of attention mechanism to improve time\nseries classification with the Attention Long Short Term Memory Fully\nConvolutional Network (ALSTM-FCN). Utilization of the attention mechanism\nallows one to visualize the decision process of the LSTM cell. Furthermore, we\npropose fine-tuning as a method to enhance the performance of trained models.\nAn overall analysis of the performance of our model is provided and compared to\nother techniques.\n",
"title": "LSTM Fully Convolutional Networks for Time Series Classification"
}
| null | null | null | null | true | null |
14656
| null |
Default
| null | null |
null |
{
"abstract": " We extend some of the results proved for scalar equations in [3,4], to the\ncase of systems of integrable conservation laws. In particular, for such\nsystems we prove that the eigenvalues of a matrix obtained from the quasilinear\npart of the system are invariants under Miura transformations and we show how\nthese invariants are related to dispersion relations. Furthermore, focusing on\none-parameter families of dispersionless systems of integrable conservation\nlaws associated to the Coxeter groups of rank $2$ found in [1], we study the\ncorresponding integrable deformations up to order $2$ in the deformation\nparameter $\\epsilon$. Each family contains both bi-Hamiltonian and\nnon-Hamiltonian systems of conservation laws and therefore we use it to probe\nto which extent the properties of the dispersionless limit impact the nature\nand the existence of integrable deformations. It turns out that a part two\nvalues of the parameter all deformations of order one in $\\epsilon$ are\nMiura-trivial, while all those of order two in $\\epsilon$ are essentially\nparameterized by two arbitrary functions of single variables (the Riemann\ninvariants) both in the bi-Hamiltonian and in the non-Hamiltonian case. In the\ntwo remaining cases, due to the existence of non-trivial first order\ndeformations, there is an additional functional parameter.\n",
"title": "Flat $F$-manifolds, Miura invariants and integrable systems of conservation laws"
}
| null | null |
[
"Physics"
] | null | true | null |
14657
| null |
Validated
| null | null |
null |
{
"abstract": " This paper introduces a new approach to cost-effective, high-throughput\nhardware designs for Low Density Parity Check (LDPC) decoders. The proposed\napproach, called Non-Surjective Finite Alphabet Iterative Decoders (NS-FAIDs),\nexploits the robustness of message-passing LDPC decoders to inaccuracies in the\ncalculation of exchanged messages, and it is shown to provide a unified\nframework for several designs previously proposed in the literature. NS-FAIDs\nare optimized by density evolution for regular and irregular LDPC codes, and\nare shown to provide different trade-offs between hardware complexity and\ndecoding performance. Two hardware architectures targeting high-throughput\napplications are also proposed, integrating both Min-Sum (MS) and NS-FAID\ndecoding kernels. ASIC post synthesis implementation results on 65nm CMOS\ntechnology show that NS-FAIDs yield significant improvements in the throughput\nto area ratio, by up to 58.75% with respect to the MS decoder, with even better\nor only slightly degraded error correction performance.\n",
"title": "Analysis and Design of Cost-Effective, High-Throughput LDPC Decoders"
}
| null | null | null | null | true | null |
14658
| null |
Default
| null | null |
null |
{
"abstract": " The TensorFlow Distributions library implements a vision of probability\ntheory adapted to the modern deep-learning paradigm of end-to-end\ndifferentiable computation. Building on two basic abstractions, it offers\nflexible building blocks for probabilistic computation. Distributions provide\nfast, numerically stable methods for generating samples and computing\nstatistics, e.g., log density. Bijectors provide composable volume-tracking\ntransformations with automatic caching. Together these enable modular\nconstruction of high dimensional distributions and transformations not possible\nwith previous libraries (e.g., pixelCNNs, autoregressive flows, and reversible\nresidual networks). They are the workhorse behind deep probabilistic\nprogramming systems like Edward and empower fast black-box inference in\nprobabilistic models built on deep-network components. TensorFlow Distributions\nhas proven an important part of the TensorFlow toolkit within Google and in the\nbroader deep learning community.\n",
"title": "TensorFlow Distributions"
}
| null | null | null | null | true | null |
14659
| null |
Default
| null | null |
null |
{
"abstract": " We present a one-parameter family of mathematical models describing the\ndynamics of polarons in linear periodic structures such as polypeptides. By\ntuning the parameter, we are able to recover the Davydov and the Scott models.\nWe describe the physical significance of this parameter. In the continuum\nlimit, we derive analytical solutions which represent stationary polarons. On a\ndiscrete lattice, we compute stationary polaron solutions numerically. We\ninvestigate polaron propagation induced by several external forcing mechanisms.\nWe show that an electric field consisting of a constant and a periodic\ncomponent can induce polaron motion with minimal energy loss. We also show that\nthermal fluctuations can facilitate the onset of polaron motion. Finally, we\ndiscuss the bio-physical implications of our results.\n",
"title": "A generalised Davydov-Scott model for polarons in linear peptide chains"
}
| null | null | null | null | true | null |
14660
| null |
Default
| null | null |
null |
{
"abstract": " We mine the Tycho-{\\it Gaia} astrometric solution (TGAS) catalog for wide\nstellar binaries by matching positions, proper motions, and astrometric\nparallaxes. We separate genuine binaries from unassociated stellar pairs\nthrough a Bayesian formulation that includes correlated uncertainties in the\nproper motions and parallaxes. Rather than relying on assumptions about the\nstructure of the Galaxy, we calculate Bayesian priors and likelihoods based on\nthe nature of Keplerian orbits and the TGAS catalog itself. We calibrate our\nmethod using radial velocity measurements and obtain 6196 high-confidence\ncandidate wide binaries with projected separations $s\\lesssim1$ pc. The\nnormalization of this distribution suggests that at least 0.6\\% of TGAS stars\nhave an associated, distant TGAS companion in a wide binary. We demonstrate\nthat {\\it Gaia}'s astrometry is precise enough that it can detect projected\norbital velocities in wide binaries with orbital periods as large as 10$^6$ yr.\nFor pairs with $s\\ \\lesssim\\ 4\\times10^4$~AU, characterization of random\nalignments indicate our contamination to be $\\approx$5\\%. For $s \\lesssim\n5\\times10^3$~AU, our distribution is consistent with Öpik's Law. At larger\nseparations, the distribution is steeper and consistent with a power-law\n$P(s)\\propto s^{-1.6}$; there is no evidence in our data of any bimodality in\nthis distribution for $s \\lesssim$ 1 pc. Using radial velocities, we\ndemonstrate that at large separations, i.e., of order $s \\sim$ 1 pc and beyond,\nany potential sample of genuine wide binaries in TGAS cannot be easily\ndistinguished from ionized former wide binaries, moving groups, or\ncontamination from randomly aligned stars.\n",
"title": "Wide Binaries in Tycho-{\\it Gaia}: Search Method and the Distribution of Orbital Separations"
}
| null | null | null | null | true | null |
14661
| null |
Default
| null | null |
null |
{
"abstract": " When each data point is a large graph, graph statistics such as densities of\ncertain subgraphs (motifs) can be used as feature vectors for machine learning.\nWhile intuitive, motif counts are expensive to compute and difficult to work\nwith theoretically. Via graphon theory, we give an explicit quantitative bound\nfor the ability of motif homomorphisms to distinguish large networks under both\ngenerative and sampling noise. Furthermore, we give similar bounds for the\ngraph spectrum and connect it to homomorphism densities of cycles. This results\nin an easily computable classifier on graph data with theoretical performance\nguarantee. Our method yields competitive results on classification tasks for\nthe autoimmune disease Lupus Erythematosus.\n",
"title": "Classification on Large Networks: A Quantitative Bound via Motifs and Graphons"
}
| null | null | null | null | true | null |
14662
| null |
Default
| null | null |
null |
{
"abstract": " Accurately predicting machine failures in advance can decrease maintenance\ncost and help allocate maintenance resources more efficiently. Logistic\nregression was applied to predict machine state 24 hours in the future given\nthe current machine state.\n",
"title": "Predicting Future Machine Failure from Machine State Using Logistic Regression"
}
| null | null | null | null | true | null |
14663
| null |
Default
| null | null |
null |
{
"abstract": " Collective Adaptive Systems (CAS) consist of a large number of interacting\nobjects. The design of such systems requires scalable analysis tools and\nmethods, which have necessarily to rely on some form of approximation of the\nsystem's actual behaviour. Promising techniques are those based on mean-field\napproximation. The FlyFast model-checker uses an on-the-fly algorithm for\nbounded PCTL model-checking of selected individual(s) in the context of very\nlarge populations whose global behaviour is approximated using deterministic\nlimit mean-field techniques. Recently, a front-end for FlyFast has been\nproposed which provides a modelling language, PiFF in the sequel, for the\nPredicate-based Interaction for FlyFast. In this paper we present details of\nPiFF design and an approach to state-space reduction based on probabilistic\nbisimulation for inhomogeneous DTMCs.\n",
"title": "Design and Optimisation of the FlyFast Front-end for Attribute-based Coordination"
}
| null | null | null | null | true | null |
14664
| null |
Default
| null | null |
null |
{
"abstract": " The use of functional brain imaging for research and diagnosis has benefitted\ngreatly from the recent advancements in neuroimaging technologies, as well as\nthe explosive growth in size and availability of fMRI data. While it has been\nshown in literature that using multiple and large scale fMRI datasets can\nimprove reproducibility and lead to new discoveries, the computational and\ninformatics systems supporting the analysis and visualization of such fMRI big\ndata are extremely limited and largely under-discussed. We propose to address\nthese shortcomings in this work, based on previous success in using dictionary\nlearning method for functional network decomposition studies on fMRI data. We\npresented a distributed dictionary learning framework based on rank-1 matrix\ndecomposition with sparseness constraint (D-r1DL framework). The framework was\nimplemented using the Spark distributed computing engine and deployed on three\ndifferent processing units: an in-house server, in-house high performance\nclusters, and the Amazon Elastic Compute Cloud (EC2) service. The whole\nanalysis pipeline was integrated with our neuroinformatics system for data\nmanagement, user input/output, and real-time visualization. Performance and\naccuracy of D-r1DL on both individual and group-wise fMRI Human Connectome\nProject (HCP) dataset shows that the proposed framework is highly scalable. The\nresulting group-wise functional network decompositions are highly accurate, and\nthe fast processing time confirm this claim. In addition, D-r1DL can provide\nreal-time user feedback and results visualization which are vital for\nlarge-scale data analysis.\n",
"title": "Distributed rank-1 dictionary learning: Towards fast and scalable solutions for fMRI big data analytics"
}
| null | null | null | null | true | null |
14665
| null |
Default
| null | null |
null |
{
"abstract": " A presentation at the SciNeGHE conference of the past achievements, of the\npresent activities and of the perspectives for the future of the HARPO project,\nthe development of a time projection chamber as a high-performance gamma-ray\ntelescope and linear polarimeter in the e+e- pair creation regime.\n",
"title": "HARPO: 1.7 - 74 MeV gamma-ray beam validation of a high angular resolutio n, high linear polarisation dilution, gas time projection chamber telescope and polarimeter"
}
| null | null | null | null | true | null |
14666
| null |
Default
| null | null |
null |
{
"abstract": " Knaster continua and solenoids are well-known examples of indecomposable\ncontinua whose composants (maximal arcwise-connected subsets) are one-to-one\nimages of lines. We show that essentially all non-trivial one-to-one composant\nimages of (half-)lines are indecomposable. And if $f$ is a one-to-one mapping\nof $[0,\\infty)$ or $(-\\infty,\\infty)$, then there is an indecomposable\ncontinuum of which $X:=$ran$(f)$ is a composant if and only if $f$ maps all\nfinal or initial segments densely and every non-closed sequence of arcs in $X$\nhas a convergent subsequence in the hyperspace $K(X)\\cup \\{X\\}$. We also prove\nthe existence of composant-preserving embeddings in Euclidean $3$-space.\nAccompanying the proofs are illustrations and examples.\n",
"title": "One-to-one composant mappings of $[0,\\infty)$ and $(-\\infty,\\infty)$"
}
| null | null |
[
"Mathematics"
] | null | true | null |
14667
| null |
Validated
| null | null |
null |
{
"abstract": " We introduce a class of theories called metastable, including the theory of\nalgebraically closed valued fields (ACVF) as a motivating example. The key\nlocal notion is that of definable types dominated by their stable part. A\ntheory is metastable (over a sort $\\Gamma$) if every type over a sufficiently\nrich base structure can be viewed as part of a $\\Gamma$-parametrized family of\nstably dominated types. We initiate a study of definable groups in metastable\ntheories of finite rank. Groups with a stably dominated generic type are shown\nto have a canonical stable quotient. Abelian groups are shown to be\ndecomposable into a part coming from $\\Gamma$, and a definable direct limit\nsystem of groups with stably dominated generic. In the case of ACVF, among\ndefinable subgroups of affine algebraic groups, we characterize the groups with\nstably dominated generics in terms of group schemes over the valuation ring.\nFinally, we classify all fields definable in ACVF.\n",
"title": "Valued fields, Metastable groups"
}
| null | null | null | null | true | null |
14668
| null |
Default
| null | null |
null |
{
"abstract": " We study the mutual alignment of radio sources within two surveys, FIRST and\nTGSS. This is done by producing two position angle catalogues containing the\npreferential directions of respectively $30\\,059$ and $11\\,674$ extended\nsources distributed over more than $7\\,000$ and $17\\,000$ square degrees. The\nidentification of the sources in the FIRST sample was performed in advance by\nvolunteers of the Radio Galaxy Zoo project, while for the TGSS sample it is the\nresult of an automated process presented here. After taking into account\nsystematic effects, marginal evidence of a local alignment on scales smaller\nthan $2.5°$ is found in the FIRST sample. The probability of this happening\nby chance is found to be less than $2$ per cent. Further study suggests that on\nscales up to $1.5°$ the alignment is maximal. For one third of the sources,\nthe Radio Galaxy Zoo volunteers identified an optical counterpart. Assuming a\nflat $\\Lambda$CDM cosmology with $\\Omega_m = 0.31, \\Omega_\\Lambda = 0.69$, we\nconvert the maximum angular scale on which alignment is seen into a physical\nscale in the range $[19, 38]$ Mpc $h_{70}^{-1}$. This result supports recent\nevidence reported by Taylor and Jagannathan of radio jet alignment in the $1.4$\ndeg$^2$ ELAIS N1 field observed with the Giant Metrewave Radio Telescope. The\nTGSS sample is found to be too sparsely populated to manifest a similar signal.\n",
"title": "Radio Galaxy Zoo: Cosmological Alignment of Radio Sources"
}
| null | null |
[
"Physics"
] | null | true | null |
14669
| null |
Validated
| null | null |
null |
{
"abstract": " Graph based semi-supervised learning (GSSL) has intuitive representation and\ncan be improved by exploiting the matrix calculation. However, it has to\nperform iterative optimization to achieve a preset objective, which usually\nleads to low efficiency. Another inconvenience lying in GSSL is that when new\ndata come, the graph construction and the optimization have to be conducted all\nover again. We propose a sound assumption, arguing that: the neighboring data\npoints are not in peer-to-peer relation, but in a partial-ordered relation\ninduced by the local density and distance between the data; and the label of a\ncenter can be regarded as the contribution of its followers. Starting from the\nassumption, we develop a highly efficient non-iterative label propagation\nalgorithm based on a novel data structure named as optimal leading forest\n(LaPOLeaF). The major weaknesses of the traditional GSSL are addressed by this\nstudy. We further scale LaPOLeaF to accommodate big data by utilizing block\ndistance matrix technique, parallel computing, and Locality-Sensitive Hashing\n(LSH). Experiments on large datasets have shown the promising results of the\nproposed methods.\n",
"title": "Non-iterative Label Propagation in Optimal Leading Forest"
}
| null | null |
[
"Computer Science"
] | null | true | null |
14670
| null |
Validated
| null | null |
null |
{
"abstract": " Peer-to-peer (P2P) botnets have become one of the major threats in network\nsecurity for serving as the infrastructure that responsible for various of\ncyber-crimes. Though a few existing work claimed to detect traditional botnets\neffectively, the problem of detecting P2P botnets involves more challenges. In\nthis paper, we present PeerHunter, a community behavior analysis based method,\nwhich is capable of detecting botnets that communicate via a P2P structure.\nPeerHunter starts from a P2P hosts detection component. Then, it uses mutual\ncontacts as the main feature to cluster bots into communities. Finally, it uses\ncommunity behavior analysis to detect potential botnet communities and further\nidentify bot candidates. Through extensive experiments with real and simulated\nnetwork traces, PeerHunter can achieve very high detection rate and low false\npositives.\n",
"title": "PeerHunter: Detecting Peer-to-Peer Botnets through Community Behavior Analysis"
}
| null | null | null | null | true | null |
14671
| null |
Default
| null | null |
null |
{
"abstract": " Galaxy-scale outflows are nowadays observed in many active galactic nuclei\n(AGNs); however, their characterisation in terms of (multi-) phase nature,\namount of flowing material, effects on the host galaxy, is still unsettled. In\nparticular, ionized gas mass outflow rate and related energetics are still\naffected by many sources of uncertainties. In this respect, outflowing gas\nplasma conditions, being largely unknown, play a crucial role.\nTaking advantage of the spectroscopic analysis results we obtained studying\nthe X-ray/SDSS sample of 563 AGNs at z $<0.8$ presented in our companion paper,\nwe analyse stacked spectra and sub-samples of sources with high signal-to-noise\ntemperature- and density-sensitive emission lines to derive the plasma\nproperties of the outflowing ionized gas component. For these sources, we also\nstudy in detail various diagnostic diagrams to infer information about\noutflowing gas ionization mechanisms. We derive, for the first time, median\nvalues for electron temperature and density of outflowing gas from medium-size\nsamples ($\\sim 30$ targets) and stacked spectra of AGNs. Evidences of shock\nexcitation are found for outflowing gas.\nWe measure electron temperatures of the order of $\\sim 1.7\\times10^4$ K and\ndensities of $\\sim 1200$ cm$^{-3}$ for faint and moderately luminous AGNs\n(intrinsic X-ray luminosity $40.5<log(L_X)<44$ in the 2-10 keV band). We\ncaution that the usually assumed electron density ($N_e=100$ cm$^{-3}$) in\nejected material might result in relevant overestimates of flow mass rates and\nenergetics and, as a consequence, of the effects of AGN-driven outflows on the\nhost galaxy.\n",
"title": "An X-ray/SDSS sample (II): outflowing gas plasma properties"
}
| null | null | null | null | true | null |
14672
| null |
Default
| null | null |
null |
{
"abstract": " Deep neural networks (DNNs) are one of the most prominent technologies of our\ntime, as they achieve state-of-the-art performance in many machine learning\ntasks, including but not limited to image classification, text mining, and\nspeech processing. However, recent research on DNNs has indicated\never-increasing concern on the robustness to adversarial examples, especially\nfor security-critical tasks such as traffic sign identification for autonomous\ndriving. Studies have unveiled the vulnerability of a well-trained DNN by\ndemonstrating the ability of generating barely noticeable (to both human and\nmachines) adversarial images that lead to misclassification. Furthermore,\nresearchers have shown that these adversarial images are highly transferable by\nsimply training and attacking a substitute model built upon the target model,\nknown as a black-box attack to DNNs.\nSimilar to the setting of training substitute models, in this paper we\npropose an effective black-box attack that also only has access to the input\n(images) and the output (confidence scores) of a targeted DNN. However,\ndifferent from leveraging attack transferability from substitute models, we\npropose zeroth order optimization (ZOO) based attacks to directly estimate the\ngradients of the targeted DNN for generating adversarial examples. We use\nzeroth order stochastic coordinate descent along with dimension reduction,\nhierarchical attack and importance sampling techniques to efficiently attack\nblack-box models. By exploiting zeroth order optimization, improved attacks to\nthe targeted DNN can be accomplished, sparing the need for training substitute\nmodels and avoiding the loss in attack transferability. Experimental results on\nMNIST, CIFAR10 and ImageNet show that the proposed ZOO attack is as effective\nas the state-of-the-art white-box attack and significantly outperforms existing\nblack-box attacks via substitute models.\n",
"title": "ZOO: Zeroth Order Optimization based Black-box Attacks to Deep Neural Networks without Training Substitute Models"
}
| null | null | null | null | true | null |
14673
| null |
Default
| null | null |
null |
{
"abstract": " We prove, by a computer aided proof, the existence of noise induced order in\nthe model of chaotic chemical reactions where it was first discovered\nnumerically by Matsumoto and Tsuda in 1983. We prove that in this random\ndynamical system the increase in amplitude of the noise causes the Lyapunov\nexponent to decrease from positive to negative, stabilizing the system. The\nmethod used is based on a certified approximation of the stationary measure in\nthe $L^{1}$ norm. This is done by an efficient algorithm which is general\nenough to be adapted to any piecewise differentiable dynamical system on the\ninterval with additive noise. We also prove that the stationary measure of the\nsystem and its Lyapunov exponent have a Lipschitz stability under several kinds\nof perturbation of the noise and of the system itself. The Lipschitz constants\nof this stability result are also estimated explicitly.\n",
"title": "Existence of Noise Induced Order, a Computer Aided Proof"
}
| null | null | null | null | true | null |
14674
| null |
Default
| null | null |
null |
{
"abstract": " Recent work in fairness in machine learning has proposed adjusting for\nfairness by equalizing accuracy metrics across groups and has also studied how\ndatasets affected by historical prejudices may lead to unfair decision\npolicies. We connect these lines of work and study the residual unfairness that\narises when a fairness-adjusted predictor is not actually fair on the target\npopulation due to systematic censoring of training data by existing biased\npolicies. This scenario is particularly common in the same applications where\nfairness is a concern. We characterize theoretically the impact of such\ncensoring on standard fairness metrics for binary classifiers and provide\ncriteria for when residual unfairness may or may not appear. We prove that,\nunder certain conditions, fairness-adjusted classifiers will in fact induce\nresidual unfairness that perpetuates the same injustices, against the same\ngroups, that biased the data to begin with, thus showing that even\nstate-of-the-art fair machine learning can have a \"bias in, bias out\" property.\nWhen certain benchmark data is available, we show how sample reweighting can\nestimate and adjust fairness metrics while accounting for censoring. We use\nthis to study the case of Stop, Question, and Frisk (SQF) and demonstrate that\nattempting to adjust for fairness perpetuates the same injustices that the\npolicy is infamous for.\n",
"title": "Residual Unfairness in Fair Machine Learning from Prejudiced Data"
}
| null | null | null | null | true | null |
14675
| null |
Default
| null | null |
null |
{
"abstract": " Predicting the future location of vehicles is essential for safety-critical\napplications such as advanced driver assistance systems (ADAS) and autonomous\ndriving. This paper introduces a novel approach to simultaneously predict both\nthe location and scale of target vehicles in the first-person (egocentric) view\nof an ego-vehicle. We present a multi-stream recurrent neural network (RNN)\nencoder-decoder model that separately captures both object location and scale\nand pixel-level observations for future vehicle localization. We show that\nincorporating dense optical flow improves prediction results significantly\nsince it captures information about motion as well as appearance change. We\nalso find that explicitly modeling future motion of the ego-vehicle improves\nthe prediction accuracy, which could be especially beneficial in intelligent\nand automated vehicles that have motion planning capability. To evaluate the\nperformance of our approach, we present a new dataset of first-person videos\ncollected from a variety of scenarios at road intersections, which are\nparticularly challenging moments for prediction because vehicle trajectories\nare diverse and dynamic.\n",
"title": "Egocentric Vision-based Future Vehicle Localization for Intelligent Driving Assistance Systems"
}
| null | null | null | null | true | null |
14676
| null |
Default
| null | null |
null |
{
"abstract": " The metriplectic formalism couples Poisson brackets of the Hamiltonian\ndescription with metric brackets for describing systems with both Hamiltonian\nand dissipative components. The construction builds in asymptotic convergence\nto a preselected equilibrium state. Phenomena such as friction, electric\nresistivity, thermal conductivity and collisions in kinetic theories are well\nrepresented in this framework. In this paper we present an application of the\nmetriplectic formalism of interest for the theory of control: a suitable torque\nis applied to a free rigid body, which is expressed through a metriplectic\nextension of its \"natural\" Poisson algebra. On practical grounds, the effect is\nto drive the body to align its angular velocity to rotation about a stable\nprincipal axis of inertia, while conserving its kinetic energy in the process.\nOn theoretical grounds, this example shows how the non-Hamiltonian part of a\nmetriplectic system may include convergence to a limit cycle, the first example\nof a non-zero dimensional attractor in this formalism. The method suggests a\nway to extend metriplectic dynamics to systems with general attractors, e.g.\nchaotic ones, with the hope of representing bio-physical, geophysical and\necological models.\n",
"title": "Metriplectic formalism: friction and much more"
}
| null | null |
[
"Physics"
] | null | true | null |
14677
| null |
Validated
| null | null |
null |
{
"abstract": " We introduce a hierarchical architecture for video understanding that\nexploits the structure of real world actions by capturing targets at different\nlevels of granularity. We design the model such that it first learns simpler\ncoarse-grained tasks, and then moves on to learn more fine-grained targets. The\nmodel is trained with a joint loss on different granularity levels. We\ndemonstrate empirical results on the recent release of Something-Something\ndataset, which provides a hierarchy of targets, namely coarse-grained action\ngroups, fine-grained action categories, and captions. Experiments suggest that\nmodels that exploit targets at different levels of granularity achieve better\nperformance on all levels.\n",
"title": "Hierarchical Video Understanding"
}
| null | null | null | null | true | null |
14678
| null |
Default
| null | null |
null |
{
"abstract": " Finding similar user pairs is a fundamental task in social networks, with\nnumerous applications in ranking and personalization tasks such as link\nprediction and tie strength detection. A common manifestation of user\nsimilarity is based upon network structure: each user is represented by a\nvector that represents the user's network connections, where pairwise cosine\nsimilarity among these vectors defines user similarity. The predominant task\nfor user similarity applications is to discover all similar pairs that have a\npairwise cosine similarity value larger than a given threshold $\\tau$. In\ncontrast to previous work where $\\tau$ is assumed to be quite close to 1, we\nfocus on recommendation applications where $\\tau$ is small, but still\nmeaningful. The all pairs cosine similarity problem is computationally\nchallenging on networks with billions of edges, and especially so for settings\nwith small $\\tau$. To the best of our knowledge, there is no practical solution\nfor computing all user pairs with, say $\\tau = 0.2$ on large social networks,\neven using the power of distributed algorithms.\nOur work directly addresses this challenge by introducing a new algorithm ---\nWHIMP --- that solves this problem efficiently in the MapReduce model. The key\ninsight in WHIMP is to combine the \"wedge-sampling\" approach of Cohen-Lewis for\napproximate matrix multiplication with the SimHash random projection techniques\nof Charikar. We provide a theoretical analysis of WHIMP, proving that it has\nnear optimal communication costs while maintaining computation cost comparable\nwith the state of the art. We also empirically demonstrate WHIMP's scalability\nby computing all highly similar pairs on four massive data sets, and show that\nit accurately finds high similarity pairs. In particular, we note that WHIMP\nsuccessfully processes the entire Twitter network, which has tens of billions\nof edges.\n",
"title": "When Hashes Met Wedges: A Distributed Algorithm for Finding High Similarity Vectors"
}
| null | null | null | null | true | null |
14679
| null |
Default
| null | null |
null |
{
"abstract": " Square arrays of sub-micrometer columnar defects in thin\nYBa$_{2}$Cu$_{3}$O$_{7-\\delta}$ (YBCO) films with spacings down to 300 nm have\nbeen fabricated by a He ion beam projection technique. Pronounced peaks in the\ncritical current and corresponding minima in the resistance demonstrate the\ncommensurate arrangement of flux quanta with the artificial pinning landscape,\ndespite the strong intrinsic pinning in epitaxial YBCO films. Whereas these\nvortex matching signatures are exactly at predicted values in field-cooled\nexperiments, they are displaced in zero-field cooled, magnetic-field ramped\nexperiments, conserving the equidistance of the matching peaks and minima.\nThese observations reveal an unconventional critical state in a cuprate\nsuperconductor with an artificial, periodic pinning array. The long-term\nstability of such out-of-equilibrium vortex arrangements paves the way for\nelectronic applications employing fluxons.\n",
"title": "Hysteretic vortex matching effects in high-$T_c$ superconductors with nanoscale periodic pinning landscapes fabricated by He ion beam projection technique"
}
| null | null | null | null | true | null |
14680
| null |
Default
| null | null |
null |
{
"abstract": " Gathering information about forest variables is an expensive and arduous\nactivity. As such, directly collecting the data required to produce\nhigh-resolution maps over large spatial domains is infeasible. Next generation\ncollection initiatives of remotely sensed Light Detection and Ranging (LiDAR)\ndata are specifically aimed at producing complete-coverage maps over large\nspatial domains. Given that LiDAR data and forest characteristics are often\nstrongly correlated, it is possible to make use of the former to model,\npredict, and map forest variables over regions of interest. This entails\ndealing with the high-dimensional ($\\sim$$10^2$) spatially dependent LiDAR\noutcomes over a large number of locations (~10^5-10^6). With this in mind, we\ndevelop the Spatial Factor Nearest Neighbor Gaussian Process (SF-NNGP) model,\nand embed it in a two-stage approach that connects the spatial structure found\nin LiDAR signals with forest variables. We provide a simulation experiment that\ndemonstrates inferential and predictive performance of the SF-NNGP, and use the\ntwo-stage modeling strategy to generate complete-coverage maps of forest\nvariables with associated uncertainty over a large region of boreal forests in\ninterior Alaska.\n",
"title": "Spatial Factor Models for High-Dimensional and Large Spatial Data: An Application in Forest Variable Mapping"
}
| null | null | null | null | true | null |
14681
| null |
Default
| null | null |
null |
{
"abstract": " Recently, many studies have demonstrated deep neural network (DNN)\nclassifiers can be fooled by the adversarial example, which is crafted via\nintroducing some perturbations into an original sample. Accordingly, some\npowerful defense techniques were proposed. However, existing defense techniques\noften require modifying the target model or depend on the prior knowledge of\nattacks. In this paper, we propose a straightforward method for detecting\nadversarial image examples, which can be directly deployed into unmodified\noff-the-shelf DNN models. We consider the perturbation to images as a kind of\nnoise and introduce two classic image processing techniques, scalar\nquantization and smoothing spatial filter, to reduce its effect. The image\nentropy is employed as a metric to implement an adaptive noise reduction for\ndifferent kinds of images. Consequently, the adversarial example can be\neffectively detected by comparing the classification results of a given sample\nand its denoised version, without referring to any prior knowledge of attacks.\nMore than 20,000 adversarial examples against some state-of-the-art DNN models\nare used to evaluate the proposed method, which are crafted with different\nattack techniques. The experiments show that our detection method can achieve a\nhigh overall F1 score of 96.39% and certainly raises the bar for defense-aware\nattacks.\n",
"title": "Detecting Adversarial Image Examples in Deep Networks with Adaptive Noise Reduction"
}
| null | null | null | null | true | null |
14682
| null |
Default
| null | null |
null |
{
"abstract": " Sentiment analysis aims to uncover emotions conveyed through information. In\nits simplest form, it is performed on a polarity basis, where the goal is to\nclassify information with positive or negative emotion. Recent research has\nexplored more nuanced ways to capture emotions that go beyond polarity. For\nthese methods to work, they require a critical resource: a lexicon that is\nappropriate for the task at hand, in terms of the range of emotions it captures\ndiversity. In the past, sentiment analysis lexicons have been created by\nexperts, such as linguists and behavioural scientists, with strict rules.\nLexicon evaluation was also performed by experts or gold standards. In our\npaper, we propose a crowdsourcing method for lexicon acquisition, which is\nscalable, cost-effective, and doesn't require experts or gold standards. We\nalso compare crowd and expert evaluations of the lexicon, to assess the overall\nlexicon quality, and the evaluation capabilities of the crowd.\n",
"title": "Crowdsourcing for Beyond Polarity Sentiment Analysis A Pure Emotion Lexicon"
}
| null | null | null | null | true | null |
14683
| null |
Default
| null | null |
null |
{
"abstract": " We address the statistical and optimization impacts of the classical sketch\nand Hessian sketch used to approximately solve the Matrix Ridge Regression\n(MRR) problem. Prior research has quantified the effects of classical sketch on\nthe strictly simpler least squares regression (LSR) problem. We establish that\nclassical sketch has a similar effect upon the optimization properties of MRR\nas it does on those of LSR: namely, it recovers nearly optimal solutions. By\ncontrast, Hessian sketch does not have this guarantee, instead, the\napproximation error is governed by a subtle interplay between the \"mass\" in the\nresponses and the optimal objective value.\nFor both types of approximation, the regularization in the sketched MRR\nproblem results in significantly different statistical properties from those of\nthe sketched LSR problem. In particular, there is a bias-variance trade-off in\nsketched MRR that is not present in sketched LSR. We provide upper and lower\nbounds on the bias and variance of sketched MRR, these bounds show that\nclassical sketch significantly increases the variance, while Hessian sketch\nsignificantly increases the bias. Empirically, sketched MRR solutions can have\nrisks that are higher by an order-of-magnitude than those of the optimal MRR\nsolutions.\nWe establish theoretically and empirically that model averaging greatly\ndecreases the gap between the risks of the true and sketched solutions to the\nMRR problem. Thus, in parallel or distributed settings, sketching combined with\nmodel averaging is a powerful technique that quickly obtains near-optimal\nsolutions to the MRR problem while greatly mitigating the increased statistical\nrisk incurred by sketching.\n",
"title": "Sketched Ridge Regression: Optimization Perspective, Statistical Perspective, and Model Averaging"
}
| null | null | null | null | true | null |
14684
| null |
Default
| null | null |
null |
{
"abstract": " We report on the observation of magnon thermal conductivity $\\kappa_m\\sim$ 70\nW/mK near 5 K in the helimagnetic insulator Cu$_2$OSeO$_3$, exceeding that\nmeasured in any other ferromagnet by almost two orders of magnitude. Ballistic,\nboundary-limited transport for both magnons and phonons is established below 1\nK, and Poiseuille flow of magnons is proposed to explain a magnon mean-free\npath substantially exceeding the specimen width for the least defective\nspecimens in the range 2 K $<T<$ 10 K. These observations establish\nCu$_2$OSeO$_3$ as a model system for studying long-wavelength magnon dynamics.\n",
"title": "Ballistic magnon heat conduction and possible Poiseuille flow in the helimagnetic insulator Cu$_2$OSeO$_3$"
}
| null | null | null | null | true | null |
14685
| null |
Default
| null | null |
null |
{
"abstract": " Extensions and generalizations of Alzer's inequality; which is of Wirtinger\ntype are proved. As applications, sharp trapezoid type inequality and sharp\nbound for the geometric mean are deduced.\n",
"title": "On Alzer's inequality"
}
| null | null | null | null | true | null |
14686
| null |
Default
| null | null |
null |
{
"abstract": " We present Oncilla robot, a novel mobile, quadruped legged locomotion\nmachine. This large-cat sized, 5.1 robot is one of a kind of a recent,\nbioinspired legged robot class designed with the capability of model-free\nlocomotion control. Animal legged locomotion in rough terrain is clearly shaped\nby sensor feedback systems. Results with Oncilla robot show that agile and\nversatile locomotion is possible without sensory signals to some extend, and\ntracking becomes robust when feedback control is added (Ajaoolleian 2015). By\nincorporating mechanical and control blueprints inspired from animals, and by\nobserving the resulting robot locomotion characteristics, we aim to understand\nthe contribution of individual components. Legged robots have a wide mechanical\nand control design parameter space, and a unique potential as research tools to\ninvestigate principles of biomechanics and legged locomotion control. But the\nhardware and controller design can be a steep initial hurdle for academic\nresearch. To facilitate the easy start and development of legged robots,\nOncilla-robot's blueprints are available through open-source. [...]\n",
"title": "Oncilla robot: a versatile open-source quadruped research robot with compliant pantograph legs"
}
| null | null | null | null | true | null |
14687
| null |
Default
| null | null |
null |
{
"abstract": " Generating graphs that are similar to real ones is an open problem, while the\nsimilarity notion is quite elusive and hard to formalize. In this paper, we\nfocus on sparse digraphs and propose SDG, an algorithm that aims at generating\ngraphs similar to real ones. Since real graphs are evolving and this evolution\nis important to study in order to understand the underlying dynamical system,\nwe tackle the problem of generating series of graphs. We propose SEDGE, an\nalgorithm meant to generate series of graphs similar to a real series. SEDGE is\nan extension of SDG. We consider graphs that are representations of software\nprograms and show experimentally that our approach outperforms other existing\napproaches. Experiments show the performance of both algorithms.\n",
"title": "A generative model for sparse, evolving digraphs"
}
| null | null | null | null | true | null |
14688
| null |
Default
| null | null |
null |
{
"abstract": " Quadratic regression goes beyond the linear model by simultaneously including\nmain effects and interactions between the covariates. The problem of\ninteraction estimation in high dimensional quadratic regression has received\nextensive attention in the past decade. In this article we introduce a novel\nmethod which allows us to estimate the main effects and interactions\nseparately. Unlike existing methods for ultrahigh dimensional quadratic\nregressions, our proposal does not require the widely used heredity assumption.\nIn addition, our proposed estimates have explicit formulas and obey the\ninvariance principle at the population level. We estimate the interactions of\nmatrix form under penalized convex loss function. The resulting estimates are\nshown to be consistent even when the covariate dimension is an exponential\norder of the sample size. We develop an efficient ADMM algorithm to implement\nthe penalized estimation. This ADMM algorithm fully explores the cheap\ncomputational cost of matrix multiplication and is much more efficient than\nexisting penalized methods such as all pairs LASSO. We demonstrate the\npromising performance of our proposal through extensive numerical studies.\n",
"title": "Penalized Interaction Estimation for Ultrahigh Dimensional Quadratic Regression"
}
| null | null | null | null | true | null |
14689
| null |
Default
| null | null |
null |
{
"abstract": " This paper addresses the problems of quantum spectral curves and 4D limit for\nthe melting crystal model of 5D SUSY $U(1)$ Yang-Mills theory on\n$\\mathbb{R}^4\\times S^1$. The partition function $Z(\\mathbf{t})$ deformed by an\ninfinite number of external potentials is a tau function of the KP hierarchy\nwith respect to the coupling constants $\\mathbf{t} = (t_1,t_2,\\ldots)$. A\nsingle-variate specialization $Z(x)$ of $Z(\\mathbf{t})$ satisfies a\n$q$-difference equation representing the quantum spectral curve of the melting\ncrystal model. In the limit as the radius $R$ of $S^1$ in $\\mathbb{R}^4\\times\nS^1$ tends to $0$, it turns into a difference equation for a 4D counterpart\n$Z_{\\mathrm{4D}}(X)$ of $Z(x)$. This difference equation reproduces the quantum\nspectral curve of Gromov-Witten theory of $\\mathbb{CP}^1$. $Z_{\\mathrm{4D}}(X)$\nis obtained from $Z(x)$ by letting $R \\to 0$ under an $R$-dependent\ntransformation $x = x(X,R)$ of $x$ to $X$. A similar prescription of 4D limit\ncan be formulated for $Z(\\mathbf{t})$ with an $R$-dependent transformation\n$\\mathbf{t} = \\mathbf{t}(\\mathbf{T},R)$ of $\\mathbf{t}$ to $\\mathbf{T} =\n(T_1,T_2,\\ldots)$. This yields a 4D counterpart $Z_{\\mathrm{4D}}(\\mathbf{T})$\nof $Z(\\mathbf{t})$. $Z_{\\mathrm{4D}}(\\mathbf{T})$ agrees with a generating\nfunction of all-genus Gromov-Witten invariants of $\\mathbb{CP}^1$. Fay-type\nbilinear equations for $Z_{\\mathrm{4D}}(\\mathbf{T})$ can be derived from\nsimilar equations satisfied by $Z(\\mathbf{t})$. The bilinear equations imply\nthat $Z_{\\mathrm{4D}}(\\mathbf{T})$, too, is a tau function of the KP hierarchy.\nThese results are further extended to deformations $Z(\\mathbf{t},s)$ and\n$Z_{\\mathrm{4D}}(\\mathbf{T},s)$ by a discrete variable $s \\in \\mathbb{Z}$,\nwhich are shown to be tau functions of the 1D Toda hierarchy.\n",
"title": "4D limit of melting crystal model and its integrable structure"
}
| null | null | null | null | true | null |
14690
| null |
Default
| null | null |
null |
{
"abstract": " Let M be ternary, homogeneous and simple. We prove that if M is finitely\nconstrained, then it is supersimple with finite SU-rank and dependence is\n$k$-trivial for some $k < \\omega$ and for finite sets of real elements. Now\nsuppose that, in addition, M is supersimple with SU-rank 1. If M is finitely\nconstrained then algebraic closure in M is trivial. We also find connections\nbetween the nature of the constraints of M, the nature of the amalgamations\nallowed by the age of M, and the nature of definable equivalence relations. A\nkey method of proof is to \"extract\" constraints (of M) from instances of\ndividing and from definable equivalence relations. Finally, we give new\nexamples, including an uncountable family, of ternary homogeneous supersimple\nstructures of SU-rank 1.\n",
"title": "On constraints and dividing in ternary homogeneous structures"
}
| null | null | null | null | true | null |
14691
| null |
Default
| null | null |
null |
{
"abstract": " Amyloid beta peptides (A\\b{eta}), implicated in Alzheimers disease (AD),\ninteract with the cellular membrane and induce amyloid toxicity. The\ncomposition of cellular membranes changes in aging and AD. We designed multi\ncomponent lipid models to mimic healthy and diseased states of the neuronal\nmembrane. Using atomic force microscopy (AFM), Kelvin probe force microscopy\n(KPFM) and black lipid membrane (BLM) techniques, we demonstrated that these\nmodel membranes differ in their nanoscale structure and physical properties,\nand interact differently with A\\b{eta}. Based on our data, we propose a new\nhypothesis that changes in lipid membrane due to aging and AD may trigger\namyloid toxicity through electrostatic mechanisms, similar to the accepted\nmechanism of antimicrobial peptide action. Understanding the role of the\nmembrane changes as a key activating amyloid toxicity may aid in the\ndevelopment of a new avenue for the prevention and treatment of AD.\n",
"title": "Changes in lipid membranes may trigger amyloid toxicity in Alzheimer's disease"
}
| null | null | null | null | true | null |
14692
| null |
Default
| null | null |
null |
{
"abstract": " Let $\\mathbb{K}$ be the algebraic closure of a finite field $\\mathbb{F}_q$ of\nodd characteristic $p$. For a positive integer $m$ prime to $p$, let\n$F=\\mathbb{K}(x,y)$ be the transcendency degree $1$ function field defined by\n$y^q+y=x^m+x^{-m}$. Let $t=x^{m(q-1)}$ and $H=\\mathbb{K}(t)$. The extension\n$F|H$ is a non-Galois extension. Let $K$ be the Galois closure of $F$ with\nrespect to $H$. By a result of Stichtenoth, $K$ has genus $g(K)=(qm-1)(q-1)$,\n$p$-rank (Hasse-Witt invariant) $\\gamma(K)=(q-1)^2$ and a\n$\\mathbb{K}$-automorphism group of order at least $2q^2m(q-1)$. In this paper\nwe prove that this subgroup is the full $\\mathbb{K}$-automorphism group of $K$;\nmore precisely $Aut_{\\mathbb {K}}(K)=Q\\rtimes D$ where $Q$ is an elementary\nabelian $p$-group of order $q^2$ and $D$ has a index $2$ cyclic subgroup of\norder $m(q-1)$. In particular, $\\sqrt{m}|Aut_{\\mathbb{K}}(K)|> g(K)^{3/2}$, and\nif $K$ is ordinary (i.e. $g(K)=\\gamma(K)$) then\n$|Aut_{\\mathbb{K}}(K)|>g^{3/2}$. On the other hand, if $G$ is a solvable\nsubgroup of the $\\mathbb{K}$-automorphism group of an ordinary, transcendency\ndegree $1$ function field $L$ of genus $g(L)\\geq 2$ defined over $\\mathbb{K}$,\nthen by a result due to Korchmáros and Montanucci, $|Aut_{\\mathbb{K}}(K)|\\le\n34 (g(L)+1)^{3/2}<68\\sqrt{2}g(L)^{3/2}$. This shows that $K$ hits this bound up\nto the constant $68\\sqrt{2}$.\nSince $Aut_{\\mathbb{K}}(K)$ has several subgroups, the fixed subfield $F^N$\nof such a subgroup $N$ may happen to have many automorphisms provided that the\nnormalizer of $N$ in $Aut_{\\mathbb{K}}(K)$ is large enough. This possibility is\nworked out for subgroups of $Q$.\n",
"title": "Transcendency Degree One Function Fields Over a Finite Field with Many Automorphisms"
}
| null | null | null | null | true | null |
14693
| null |
Default
| null | null |
null |
{
"abstract": " The study of continuous phase transitions triggered by spontaneous symmetry\nbreaking has brought revolutionary ideas to physics. Recently, through the\ndiscovery of symmetry protected topological phases, it is realized that\ncontinuous quantum phase transition can also occur between states with the same\nsymmetry but different topology. Here we study a specific class of such phase\ntransitions in 1+1 dimensions -- the phase transition between bosonic\ntopological phases protected by $Z_n\\times Z_n$. We find in all cases the\ncritical point possesses two gap opening relevant operators: one leads to a\nLandau-forbidden symmetry breaking phase transition and the other to the\ntopological phase transition. We also obtained a constraint on the central\ncharge for general phase transitions between symmetry protected bosonic\ntopological phases in 1+1D.\n",
"title": "The phase transitions between $Z_n\\times Z_n$ bosonic topological phases in 1+1 D, and a constraint on the central charge for the critical points between bosonic symmetry protected topological phases"
}
| null | null |
[
"Physics"
] | null | true | null |
14694
| null |
Validated
| null | null |
null |
{
"abstract": " Many state-of-the-art reinforcement learning (RL) algorithms typically assume\nthat the environment is an ergodic Markov Decision Process (MDP). In contrast,\nthe field of universal reinforcement learning (URL) is concerned with\nalgorithms that make as few assumptions as possible about the environment. The\nuniversal Bayesian agent AIXI and a family of related URL algorithms have been\ndeveloped in this setting. While numerous theoretical optimality results have\nbeen proven for these agents, there has been no empirical investigation of\ntheir behavior to date. We present a short and accessible survey of these URL\nalgorithms under a unified notation and framework, along with results of some\nexperiments that qualitatively illustrate some properties of the resulting\npolicies, and their relative performance on partially-observable gridworld\nenvironments. We also present an open-source reference implementation of the\nalgorithms which we hope will facilitate further understanding of, and\nexperimentation with, these ideas.\n",
"title": "Universal Reinforcement Learning Algorithms: Survey and Experiments"
}
| null | null | null | null | true | null |
14695
| null |
Default
| null | null |
null |
{
"abstract": " Understanding cell identity is an important task in many biomedical areas.\nExpression patterns of specific marker genes have been used to characterize\nsome limited cell types, but exclusive markers are not available for many cell\ntypes. A second approach is to use machine learning to discriminate cell types\nbased on the whole gene expression profiles (GEPs). The accuracies of simple\nclassification algorithms such as linear discriminators or support vector\nmachines are limited due to the complexity of biological systems. We used deep\nneural networks to analyze 1040 GEPs from 16 different human tissues and cell\ntypes. After comparing different architectures, we identified a specific\nstructure of deep autoencoders that can encode a GEP into a vector of 30\nnumeric values, which we call the cell identity code (CIC). The original GEP\ncan be reproduced from the CIC with an accuracy comparable to technical\nreplicates of the same experiment. Although we use an unsupervised approach to\ntrain the autoencoder, we show different values of the CIC are connected to\ndifferent biological aspects of the cell, such as different pathways or\nbiological processes. This network can use CIC to reproduce the GEP of the cell\ntypes it has never seen during the training. It also can resist some noise in\nthe measurement of the GEP. Furthermore, we introduce classifier autoencoder,\nan architecture that can accurately identify cell type based on the GEP or the\nCIC.\n",
"title": "Cell Identity Codes: Understanding Cell Identity from Gene Expression Profiles using Deep Neural Networks"
}
| null | null | null | null | true | null |
14696
| null |
Default
| null | null |
null |
{
"abstract": " It is shown that continuously changing the effective number of interacting\nparticles in p-spin-glass-like model allows to describe the transition from the\nfull replica symmetry breaking glass solution to stable first replica symmetry\nbreaking glass solution in the case of non-reflective symmetry diagonal\noperators used instead of Ising spins. As an example, axial quadrupole moments\nin place of Ising spins are considered and the boundary value $p_{c_{1}}\\cong\n2.5$ is found.\n",
"title": "Full replica symmetry breaking in p-spin-glass-like systems"
}
| null | null | null | null | true | null |
14697
| null |
Default
| null | null |
null |
{
"abstract": " The density-matrix-renormalization-group (DMRG) method and the Hartree-Fock\n(HF) approximation with the charge-density-wave (CDW) instability are used to\nstudy a formation and condensation of excitonic bound states in the generalized\nFalicov-Kimball model. In particular, we examine effects of various factors,\nlike the $f$-electron hopping, the local and nonlocal hybridization, as well as\nthe increasing dimension of the system on the excitonic momentum distribution\n$N(q)$ and especially on the number of zero momentum excitons $N_0=N(q=0)$ in\nthe condensate. It is found that the negative values of the $f$-electron\nhopping integrals $t_f$ support the formation of zero-momentum condensate,\nwhile the positive values of $t_f$ have the fully opposite effect. The opposite\neffects on the formation of condensate exhibit also the local and nonlocal\nhybridization. The first one strongly supports the formation of condensate,\nwhile the second one destroys it completely. Moreover, it was shown that the\nzero-momentum condensate remains robust with increasing dimension of the\nsystem.\n",
"title": "Formation and condensation of excitonic bound states in the generalized Falicov-Kimball model"
}
| null | null | null | null | true | null |
14698
| null |
Default
| null | null |
null |
{
"abstract": " At the interface between two distinct materials desirable properties, such as\nsuperconductivity, can be greatly enhanced, or entirely new functionalities may\nemerge. Similar to in artificially engineered heterostructures, clean\nfunctional interfaces alternatively exist in electronically textured bulk\nmaterials. Electronic textures emerge spontaneously due to competing\natomic-scale interactions, the control of which, would enable a top-down\napproach for designing tunable intrinsic heterostructures. This is particularly\nattractive for correlated electron materials, where spontaneous\nheterostructures strongly affect the interplay between charge and spin degrees\nof freedom. Here we report high-resolution neutron spectroscopy on the\nprototypical strongly-correlated metal CeRhIn5, revealing competition between\nmagnetic frustration and easy-axis anisotropy -- a well-established mechanism\nfor generating spontaneous superstructures. Because the observed easy-axis\nanisotropy is field-induced and anomalously large, it can be controlled\nefficiently with small magnetic fields. The resulting field-controlled magnetic\nsuperstructure is closely tied to the formation of superconducting and\nelectronic nematic textures in CeRhIn5, suggesting that in-situ tunable\nheterostructures can be realized in correlated electron materials.\n",
"title": "Tunable Emergent Heterostructures in a Prototypical Correlated Metal"
}
| null | null | null | null | true | null |
14699
| null |
Default
| null | null |
null |
{
"abstract": " Deep learning models are very effective in source separation when there are\nlarge amounts of labeled data available. However it is not always possible to\nhave carefully labeled datasets. In this paper, we propose a weak supervision\nmethod that only uses class information rather than source signals for learning\nto separate short utterance mixtures. We associate a variational autoencoder\n(VAE) with each class within a non-negative model. We demonstrate that deep\nconvolutional VAEs provide a prior model to identify complex signals in a sound\nmixture without having access to any source signal. We show that the separation\nresults are on par with source signal supervision.\n",
"title": "Weak Label Supervision for Monaural Source Separation Using Non-negative Denoising Variational Autoencoders"
}
| null | null | null | null | true | null |
14700
| null |
Default
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.