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": " Let $K(B_{\\ell_p^n},B_{\\ell_q^n}) $ be the $n$-dimensional $(p,q)$-Bohr\nradius for holomorphic functions on $\\mathbb C^n$. That is,\n$K(B_{\\ell_p^n},B_{\\ell_q^n}) $ denotes the greatest constant $r\\geq 0$ such\nthat for every entire function $f(z)=\\sum_{\\alpha} c_{\\alpha} z^{\\alpha}$ in\n$n$-complex variables, we have the following (mixed) Bohr-type inequality\n$$\\sup_{z \\in r \\cdot B_{\\ell_q^n}} \\sum_{\\alpha} | c_{\\alpha} z^{\\alpha} |\n\\leq \\sup_{z \\in B_{\\ell_p^n}} | f(z) |,$$ where $B_{\\ell_r^n}$ denotes the\nclosed unit ball of the $n$-dimensional sequence space $\\ell_r^n$.\nFor every $1 \\leq p, q \\leq \\infty$, we exhibit the exact asymptotic growth\nof the $(p,q)$-Bohr radius as $n$ (the number of variables) goes to infinity.\n",
"title": "Mixed Bohr radius in several variables"
}
| null | null | null | null | true | null |
14701
| null |
Default
| null | null |
null |
{
"abstract": " With rapid progress and significant successes in a wide spectrum of\napplications, deep learning is being applied in many safety-critical\nenvironments. However, deep neural networks have been recently found vulnerable\nto well-designed input samples, called adversarial examples. Adversarial\nexamples are imperceptible to human but can easily fool deep neural networks in\nthe testing/deploying stage. The vulnerability to adversarial examples becomes\none of the major risks for applying deep neural networks in safety-critical\nenvironments. Therefore, attacks and defenses on adversarial examples draw\ngreat attention. In this paper, we review recent findings on adversarial\nexamples for deep neural networks, summarize the methods for generating\nadversarial examples, and propose a taxonomy of these methods. Under the\ntaxonomy, applications for adversarial examples are investigated. We further\nelaborate on countermeasures for adversarial examples and explore the\nchallenges and the potential solutions.\n",
"title": "Adversarial Examples: Attacks and Defenses for Deep Learning"
}
| null | null | null | null | true | null |
14702
| null |
Default
| null | null |
null |
{
"abstract": " Neural architecture search (NAS) has been proposed to automatically tune deep\nneural networks, but existing search algorithms usually suffer from expensive\ncomputational cost. Network morphism, which keeps the functionality of a neural\nnetwork while changing its neural architecture, could be helpful for NAS by\nenabling a more efficient training during the search. In this paper, we propose\na novel framework enabling Bayesian optimization to guide the network morphism\nfor efficient neural architecture search by introducing a neural network kernel\nand a tree-structured acquisition function optimization algorithm, which more\nefficiently explores the search space. Intensive experiments have been done to\ndemonstrate the superior performance of the developed framework over the\nstate-of-the-art methods. Moreover, we build an open-source AutoML system on\nour method, namely Auto-Keras. The system runs in parallel on CPU and GPU, with\nan adaptive search strategy for different GPU memory limits.\n",
"title": "Auto-Keras: Efficient Neural Architecture Search with Network Morphism"
}
| null | null |
[
"Statistics"
] | null | true | null |
14703
| null |
Validated
| null | null |
null |
{
"abstract": " Network embeddings have become very popular in learning effective feature\nrepresentations of networks. Motivated by the recent successes of embeddings in\nnatural language processing, researchers have tried to find network embeddings\nin order to exploit machine learning algorithms for mining tasks like node\nclassification and edge prediction. However, most of the work focuses on\nfinding distributed representations of nodes, which are inherently ill-suited\nto tasks such as community detection which are intuitively dependent on\nsubgraphs.\nHere, we propose sub2vec, an unsupervised scalable algorithm to learn feature\nrepresentations of arbitrary subgraphs. We provide means to characterize\nsimilarties between subgraphs and provide theoretical analysis of sub2vec and\ndemonstrate that it preserves the so-called local proximity. We also highlight\nthe usability of sub2vec by leveraging it for network mining tasks, like\ncommunity detection. We show that sub2vec gets significant gains over\nstate-of-the-art methods and node-embedding methods. In particular, sub2vec\noffers an approach to generate a richer vocabulary of features of subgraphs to\nsupport representation and reasoning.\n",
"title": "Distributed Representation of Subgraphs"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
14704
| null |
Validated
| null | null |
null |
{
"abstract": " There are two major questions that neuroimaging studies attempt to answer:\nFirst, how are sensory stimuli represented in the brain (which we term the\nstimulus-based setting)? And, second, how does the brain generate cognition\n(termed the response-based setting)? There has been a lively debate in the\nneuroimaging community whether encoding and decoding models can provide\ninsights into these questions. In this commentary, we construct two simple and\nanalytically tractable examples to demonstrate that while an encoding model\nanalysis helps with the former, neither model is appropriate to satisfactorily\nanswer the latter question. Consequently, we argue that if we want to\nunderstand how the brain generates cognition, we need to move beyond the\nencoding versus decoding dichotomy and instead discuss and develop tools that\nare specifically tailored to our endeavour.\n",
"title": "The right tool for the right question --- beyond the encoding versus decoding dichotomy"
}
| null | null |
[
"Statistics"
] | null | true | null |
14705
| null |
Validated
| null | null |
null |
{
"abstract": " Light Axionic Dark Matter, motivated by string theory, is increasingly\nfavored for the \"no-WIMP era\". Galaxy formation is suppressed below a Jeans\nscale, of $\\simeq 10^8 M_\\odot$ by setting the axion mass to, $m_B \\sim\n10^{-22}$eV, and the large dark cores of dwarf galaxies are explained as\nsolitons on the de-Broglie scale. This is persuasive, but detection of the\ninherent scalar field oscillation at the Compton frequency, $\\omega_B= (2.5{\\rm\n\\, months})^{-1}(m_B/10^{-22}eV)$, would be definitive. By evolving the coupled\nSchrödinger-Poisson equation for a Bose-Einstein condensate, we predict the\ndark matter is fully modulated by de-Broglie interference, with a dense soliton\ncore of size $\\simeq 150pc$, at the Galactic center. The oscillating field\npressure induces General Relativistic time dilation in proportion to the local\ndark matter density and pulsars within this dense core have detectably large\ntiming residuals, of $\\simeq 400nsec/(m_B/10^{-22}eV)$. This is encouraging as\nmany new pulsars should be discovered near the Galactic center with planned\nradio surveys. More generally, over the whole Galaxy, differences in dark\nmatter density between pairs of pulsars imprints a pairwise Galactocentric\nsignature that can be distinguished from an isotropic gravitational wave\nbackground.\n",
"title": "Recognising Axionic Dark Matter by Compton and de-Broglie Scale Modulation of Pulsar Timing"
}
| null | null |
[
"Physics"
] | null | true | null |
14706
| null |
Validated
| null | null |
null |
{
"abstract": " We use the information present in a bipartite network to detect cores of\ncommunities of each set of the bipartite system. Cores of communities are found\nby investigating statistically validated projected networks obtained using\ninformation present in the bipartite network. Cores of communities are highly\ninformative and robust with respect to the presence of errors or missing\nentries in the bipartite network. We assess the statistical robustness of cores\nby investigating an artificial benchmark network, the co-authorship network,\nand the actor-movie network. The accuracy and precision of the partition\nobtained with respect to the reference partition are measured in terms of the\nadjusted Rand index and of the adjusted Wallace index respectively. The\ndetection of cores is highly precise although the accuracy of the methodology\ncan be limited in some cases.\n",
"title": "Core of communities in bipartite networks"
}
| null | null | null | null | true | null |
14707
| null |
Default
| null | null |
null |
{
"abstract": " In textual information extraction and other sequence labeling tasks it is now\ncommon to use recurrent neural networks (such as LSTM) to form rich embedded\nrepresentations of long-term input co-occurrence patterns. Representation of\noutput co-occurrence patterns is typically limited to a hand-designed graphical\nmodel, such as a linear-chain CRF representing short-term Markov dependencies\namong successive labels. This paper presents a method that learns embedded\nrepresentations of latent output structure in sequence data. Our model takes\nthe form of a finite-state machine with a large number of latent states per\nlabel (a latent variable CRF), where the state-transition matrix is\nfactorized---effectively forming an embedded representation of\nstate-transitions capable of enforcing long-term label dependencies, while\nsupporting exact Viterbi inference over output labels. We demonstrate accuracy\nimprovements and interpretable latent structure in a synthetic but complex task\nbased on CoNLL named entity recognition.\n",
"title": "Low-Rank Hidden State Embeddings for Viterbi Sequence Labeling"
}
| null | null |
[
"Computer Science"
] | null | true | null |
14708
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper, we design and analyze a new zeroth-order online algorithm,\nnamely, the zeroth-order online alternating direction method of multipliers\n(ZOO-ADMM), which enjoys dual advantages of being gradient-free operation and\nemploying the ADMM to accommodate complex structured regularizers. Compared to\nthe first-order gradient-based online algorithm, we show that ZOO-ADMM requires\n$\\sqrt{m}$ times more iterations, leading to a convergence rate of\n$O(\\sqrt{m}/\\sqrt{T})$, where $m$ is the number of optimization variables, and\n$T$ is the number of iterations. To accelerate ZOO-ADMM, we propose two\nminibatch strategies: gradient sample averaging and observation averaging,\nresulting in an improved convergence rate of $O(\\sqrt{1+q^{-1}m}/\\sqrt{T})$,\nwhere $q$ is the minibatch size. In addition to convergence analysis, we also\ndemonstrate ZOO-ADMM to applications in signal processing, statistics, and\nmachine learning.\n",
"title": "Zeroth-Order Online Alternating Direction Method of Multipliers: Convergence Analysis and Applications"
}
| null | null | null | null | true | null |
14709
| null |
Default
| null | null |
null |
{
"abstract": " We consider the stochastic damped Navier-Stokes equations in $\\mathbb R^d$\n($d=2,3$), assuming as in our previous work [4] that the covariance of the\nnoise is not too regular, so Itô calculus cannot be applied in the space of\nfinite energy vector fields. We prove the existence of an invariant measure\nwhen $d=2$ and of a stationary solution when $d=3$.\n",
"title": "Stationary solutions for stochastic damped Navier-Stokes equations in $\\mathbb R^d$"
}
| null | null | null | null | true | null |
14710
| null |
Default
| null | null |
null |
{
"abstract": " A compact circle-packing $P$ of the Euclidean plane is a set of circles which\nbound mutually disjoint open discs with the property that, for every circle\n$S\\in P$, there exists a maximal indexed set $\\{A_{0},\\ldots,A_{n-1}\\}\\subseteq\nP$ so that, for every $i\\in\\{0,\\ldots,n-1\\}$, the circle $A_{i}$ is tangent to\nboth circles $S$ and $A_{i+1\\mod n}$ .\nWe show that there exist at most $11462$ pairs $(r,s)$ with $0<s<r<1$ for\nwhich there exist a compact circle-packing of the plane consisting of circles\nwith radii $s$, $r$ and $1$.\nWe discuss computing the exact values of such $0<s<r<1$ as roots of\npolynomials and exhibit a selection of compact circle-packings consisting of\ncircles of three radii. We also discuss the apparent infeasibility of computing\nall these values on contemporary consumer hardware.\n",
"title": "On compact packings of the plane with circles of three radii"
}
| null | null | null | null | true | null |
14711
| null |
Default
| null | null |
null |
{
"abstract": " Nowadays, a hot challenge for supermarket chains is to offer personalized\nservices for their customers. Next basket prediction, i.e., supplying the\ncustomer a shopping list for the next purchase according to her current needs,\nis one of these services. Current approaches are not capable to capture at the\nsame time the different factors influencing the customer's decision process:\nco-occurrency, sequentuality, periodicity and recurrency of the purchased\nitems. To this aim, we define a pattern Temporal Annotated Recurring Sequence\n(TARS) able to capture simultaneously and adaptively all these factors. We\ndefine the method to extract TARS and develop a predictor for next basket named\nTBP (TARS Based Predictor) that, on top of TARS, is able to to understand the\nlevel of the customer's stocks and recommend the set of most necessary items.\nBy adopting the TBP the supermarket chains could crop tailored suggestions for\neach individual customer which in turn could effectively speed up their\nshopping sessions. A deep experimentation shows that TARS are able to explain\nthe customer purchase behavior, and that TBP outperforms the state-of-the-art\ncompetitors.\n",
"title": "Next Basket Prediction using Recurring Sequential Patterns"
}
| null | null | null | null | true | null |
14712
| null |
Default
| null | null |
null |
{
"abstract": " In this paper the concept of Multirate Partial Differential Equations (MPDEs)\nis applied to obtain an efficient solution for nonlinear low-frequency\nelectrical circuits with pulsed excitation. The MPDEs are solved by a Galerkin\napproach and a conventional time discretization. Nonlinearities are efficiently\naccounted for by neglecting the high-frequency components (ripples) of the\nstate variables and using only their envelope for the evaluation. It is shown\nthat the impact of this approximation on the solution becomes increasingly\nnegligible for rising frequency and leads to significant performance gains.\n",
"title": "Solving nonlinear circuits with pulsed excitation by multirate partial differential equations"
}
| null | null | null | null | true | null |
14713
| null |
Default
| null | null |
null |
{
"abstract": " This paper considers the optimal modification of the likelihood ratio test\n(LRT) for the equality of two high-dimensional covariance matrices. The\nclassical LRT is not well defined when the dimensions are larger than or equal\nto one of the sample sizes. In this paper, an optimally modified test that\nworks well in cases where the dimensions may be larger than the sample sizes is\nproposed. In addition, the test is established under the weakest conditions on\nthe moments and the dimensions of the samples. We also present weakly\nconsistent estimators of the fourth moments, which are necessary for the\nproposed test, when they are not equal to 3. From the simulation results and\nreal data analysis, we find that the performances of the proposed statistics\nare robust against affine transformations.\n",
"title": "Optimal modification of the LRT for the equality of two high-dimensional covariance matrices"
}
| null | null |
[
"Mathematics",
"Statistics"
] | null | true | null |
14714
| null |
Validated
| null | null |
null |
{
"abstract": " We consider the binary classification problem when data are large and subject\nto unknown but bounded uncertainties. We address the problem by formulating the\nnonlinear support vector machine training problem with robust optimization. To\ndo so, we analyze and propose two bounding schemes for uncertainties associated\nto random approximate features in low dimensional spaces. The proposed\ntechniques are based on Random Fourier Features and the Nyström methods. The\nresulting formulations can be solved with efficient stochastic approximation\ntechniques such as stochastic (sub)-gradient, stochastic proximal gradient\ntechniques or their variants.\n",
"title": "Feature uncertainty bounding schemes for large robust nonlinear SVM classifiers"
}
| null | null | null | null | true | null |
14715
| null |
Default
| null | null |
null |
{
"abstract": " We examine the effect of carrier localization due to random alloy\nfluctuations on the radiative and Auger recombination rates in InGaN quantum\nwells as a function of alloy composition, crystal orientation, carrier density,\nand temperature. Our results show that alloy fluctuations reduce individual\ntransition matrix elements by the separate localization of electrons and holes,\nbut this effect is overcompensated by the additional transitions enabled by\ntranslational symmetry breaking and the resulting lack of momentum\nconservation. Hence, we find that localization increases both radiative and\nAuger recombination rates, but that Auger recombination rates increase by one\norder of magnitude more than radiative rates. Furthermore, we demonstrate that\nlocalization has an overall detrimental effect on the efficiency-droop and\ngreen-gap problems of InGaN LEDs.\n",
"title": "Impact of carrier localization on recombination in InGaN quantum wells and the efficiency of nitride light-emitting diodes: insights from theory and numerical simulations"
}
| null | null | null | null | true | null |
14716
| null |
Default
| null | null |
null |
{
"abstract": " Information in neural networks is represented as weighted connections, or\nsynapses, between neurons. This poses a problem as the primary computational\nbottleneck for neural networks is the vector-matrix multiply when inputs are\nmultiplied by the neural network weights. Conventional processing architectures\nare not well suited for simulating neural networks, often requiring large\namounts of energy and time. Additionally, synapses in biological neural\nnetworks are not binary connections, but exhibit a nonlinear response function\nas neurotransmitters are emitted and diffuse between neurons. Inspired by\nneuroscience principles, we present a digital neuromorphic architecture, the\nSpiking Temporal Processing Unit (STPU), capable of modeling arbitrary complex\nsynaptic response functions without requiring additional hardware components.\nWe consider the paradigm of spiking neurons with temporally coded information\nas opposed to non-spiking rate coded neurons used in most neural networks. In\nthis paradigm we examine liquid state machines applied to speech recognition\nand show how a liquid state machine with temporal dynamics maps onto the\nSTPU-demonstrating the flexibility and efficiency of the STPU for instantiating\nneural algorithms.\n",
"title": "A Digital Neuromorphic Architecture Efficiently Facilitating Complex Synaptic Response Functions Applied to Liquid State Machines"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
14717
| null |
Validated
| null | null |
null |
{
"abstract": " Given a connected real Lie group and a contractible homogeneous proper\n$G$--space $X$ furnished with a $G$--invariant volume form, a real valued\nvolume can be assigned to any representation $\\rho\\colon \\pi_1(M)\\to G$ for any\noriented closed smooth manifold $M$ of the same dimension as $X$. Suppose that\n$G$ contains a closed and cocompact semisimple subgroup, it is shown in this\npaper that the set of volumes is finite for any given $M$. From a perspective\nof model geometries, examples are investigated and applications with mapping\ndegrees are discussed.\n",
"title": "Volume of representations and mapping degree"
}
| null | null | null | null | true | null |
14718
| null |
Default
| null | null |
null |
{
"abstract": " The Large Synoptic Survey Telescope (LSST) will generate light curves for\napproximately 1 billion stars. Our previous work has demonstrated that, by the\nend of the LSST 10 year mission, large numbers of transiting exoplanetary\nsystems could be recovered using the LSST \"deep drilling\" cadence. Here we\nextend our previous work to examine how the recoverability of transiting\nplanets over a range of orbital periods and radii evolves per year of LSST\noperation. As specific example systems we consider hot Jupiters orbiting\nsolar-type stars and hot Neptunes orbiting K-Dwarfs at distances from Earth of\nseveral kpc, as well as super-Earths orbiting nearby low-mass M-dwarfs. The\ndetection of transiting planets increases steadily with the accumulation of\ndata over time, generally becoming large (greater than 10 percent) after 4 - 6\nyears of operation. However, we also find that short-period (less than 2 day)\nhot Jupiters orbiting G-dwarfs and hot Neptunes orbiting K-dwarfs can already\nbe discovered within the first 1 - 2 years of LSST operation.\n",
"title": "Transiting Planets with LSST III: Detection Rate per Year of Operation"
}
| null | null | null | null | true | null |
14719
| null |
Default
| null | null |
null |
{
"abstract": " A Monte Carlo method based on the GEANT4 toolkit has been developed to\ncorrect the full-energy peak (FEP) efficiencies of a high purity germanium\n(HPGe) detector equipped with a low background shielding system, and moreover\nevaluated using summing peaks in a numerical way. It is found that the FEP\nefficiencies of $^{60}$Co, $^{133}$Ba and $^{152}$Eu can be improved up to 18\\%\nby taking the calculated true summing \\mbox{coincidence} factors (TSCFs)\ncorrection into account. Counts of summing coincidence $\\gamma$ peaks in the\nspectrum of $^{152}$Eu can be well reproduced using the corrected efficiency\ncurve within an accuracy of 3\\%.\n",
"title": "Summing coincidence in rare event gamma-ray measurements under an ultra-low background environment"
}
| null | null | null | null | true | null |
14720
| null |
Default
| null | null |
null |
{
"abstract": " A general theory of presentations for d-frames does not yet exist. We review\nthe difficulties and give sufficient conditions for when they can be overcome.\nAs an application we prove that the category of d-frames is closed under\ncoproducts.\n",
"title": "Free constructions and coproducts of d-frames"
}
| null | null | null | null | true | null |
14721
| null |
Default
| null | null |
null |
{
"abstract": " Automata expressiveness is an essential feature in understanding which of the\nformalisms available should be chosen for modelling a particular problem.\nProbabilistic and stochastic automata are suitable for modelling systems\nexhibiting probabilistic behavior and their expressiveness has been studied\nrelative to non-probabilistic transition systems and Markov chains. In this\npaper, we consider previous formalisms of Timed, Probabilistic and Stochastic\nTimed Automata, we present our new model of Timed Automata with Polynomial\nDelay, we introduce a measure of expressiveness for automata we call trace\nexpressiveness and we characterize the expressiveness of these models relative\nto each other under this new measure.\n",
"title": "Trace Expressiveness of Timed and Probabilistic Automata"
}
| null | null | null | null | true | null |
14722
| null |
Default
| null | null |
null |
{
"abstract": " Nowadays, protecting trust in social sciences also means engaging in open\ncommunity dialogue, which helps to safeguard robustness and improve efficiency\nof research methods. The combination of open data, open review and open\ndialogue may sound simple but implementation in the real world will not be\nstraightforward. However, in view of Begley and Ellis's (2012) statement that,\n\"the scientific process demands the highest standards of quality, ethics and\nrigour,\" they are worth implementing. More importantly, they are feasible to\nwork on and likely will help to restore plausibility to social sciences\nresearch. Therefore, I feel it likely that the triplet of open data, open\nreview and open dialogue will gradually emerge to become policy requirements\nregardless of the research funding source.\n",
"title": "Open data, open review and open dialogue in making social sciences plausible"
}
| null | null | null | null | true | null |
14723
| null |
Default
| null | null |
null |
{
"abstract": " This paper is a continuation of the second author's previous work. We\ninvestigate the isoperimetric problem in the 2-dimensional Finsler space form\n$(F_B, B^2(1))$ with $k=0$ by using the Holmes-Thompson area and prove that the\ncircle centered the origin achieves the local maximum area of the isoperimetric\nproblem.\n",
"title": "The isoperimetric problem in the 2-dimensional Finsler space forms with k = 0. II"
}
| null | null | null | null | true | null |
14724
| null |
Default
| null | null |
null |
{
"abstract": " Modeling buildings' heat dynamics is a complex process which depends on\nvarious factors including weather, building thermal capacity, insulation\npreservation, and residents' behavior. Gray-box models offer a causal inference\nof those dynamics expressed in few parameters specific to built environments.\nThese parameters can provide compelling insights into the characteristics of\nbuilding artifacts and have various applications such as forecasting HVAC\nusage, indoor temperature control monitoring of built environments, etc. In\nthis paper, we present a systematic study of modeling buildings' thermal\ncharacteristics and thus derive the parameters of built conditions with a\nBayesian approach. We build a Bayesian state-space model that can adapt and\nincorporate buildings' thermal equations and propose a generalized solution\nthat can easily adapt prior knowledge regarding the parameters. We show that a\nfaster approximate approach using variational inference for parameter\nestimation can provide similar parameters as that of a more time-consuming\nMarkov Chain Monte Carlo (MCMC) approach. We perform extensive evaluations on\ntwo datasets to understand the generative process and show that the Bayesian\napproach is more interpretable. We further study the effects of prior selection\nfor the model parameters and transfer learning, where we learn parameters from\none season and use them to fit the model in the other. We perform extensive\nevaluations on controlled and real data traces to enumerate buildings'\nparameter within a 95% credible interval.\n",
"title": "Estimating Buildings' Parameters over Time Including Prior Knowledge"
}
| null | null | null | null | true | null |
14725
| null |
Default
| null | null |
null |
{
"abstract": " We present a decentralized and scalable approach for deployment of a robot\nswarm. Our approach tackles scenarios in which the swarm must reach multiple\nspatially distributed targets, and enforce the constraint that the robot\nnetwork cannot be split. The basic idea behind our work is to construct a\nlogical tree topology over the physical network formed by the robots. The\nlogical tree acts as a backbone used by robots to enforce connectivity\nconstraints. We study and compare two algorithms to form the logical tree:\noutwards and inwards. These algorithms differ in the order in which the robots\njoin the tree: the outwards algorithm starts at the tree root and grows towards\nthe targets, while the inwards algorithm proceeds in the opposite manner. Both\nalgorithms perform periodic reconfiguration, to prevent suboptimal topologies\nfrom halting the growth of the tree. Our contributions are (i) The formulation\nof the two algorithms; (ii) A comparison of the algorithms in extensive\nphysics-based simulations; (iii) A validation of our findings through\nreal-robot experiments.\n",
"title": "Decentralized Connectivity-Preserving Deployment of Large-Scale Robot Swarms"
}
| null | null | null | null | true | null |
14726
| null |
Default
| null | null |
null |
{
"abstract": " We propose an alternative evaluation of quantum entanglement by measuring the\nmaximum violation of the Bell's inequality without performing a partial trace\noperation. This proposal is demonstrated by bridging the maximum violation of\nthe Bell's inequality and the concurrence of a pure state in an $n$-qubit\nsystem, in which one subsystem only contains one qubit and the state is a\nlinear combination of two product states. We apply this relation to the ground\nstates of four qubits in the Wen-Plaquette model and show that they are\nmaximally entangled. A topological entanglement entropy of the Wen-Plaquette\nmodel could be obtained by relating the upper bound of the maximum violation of\nthe Bell's inequality to the concurrences of a pure state with respect to\ndifferent bipartitions.\n",
"title": "Bell's Inequality and Entanglement in Qubits"
}
| null | null |
[
"Physics"
] | null | true | null |
14727
| null |
Validated
| null | null |
null |
{
"abstract": " Conjugate gradient (CG) methods are a class of important methods for solving\nlinear equations and nonlinear optimization problems. In this paper, we propose\na new stochastic CG algorithm with variance reduction and we prove its linear\nconvergence with the Fletcher and Reeves method for strongly convex and smooth\nfunctions. We experimentally demonstrate that the CG with variance reduction\nalgorithm converges faster than its counterparts for four learning models,\nwhich may be convex, nonconvex or nonsmooth. In addition, its area under the\ncurve performance on six large-scale data sets is comparable to that of the\nLIBLINEAR solver for the L2-regularized L2-loss but with a significant\nimprovement in computational efficiency\n",
"title": "Stochastic Conjugate Gradient Algorithm with Variance Reduction"
}
| null | null | null | null | true | null |
14728
| null |
Default
| null | null |
null |
{
"abstract": " Kalman filters are routinely used for many data fusion applications including\nnavigation, tracking, and simultaneous localization and mapping problems.\nHowever, significant time and effort is frequently required to tune various\nKalman filter model parameters, e.g. process noise covariance, pre-whitening\nfilter models for non-white noise, etc. Conventional optimization techniques\nfor tuning can get stuck in poor local minima and can be expensive to implement\nwith real sensor data. To address these issues, a new \"black box\" Bayesian\noptimization strategy is developed for automatically tuning Kalman filters. In\nthis approach, performance is characterized by one of two stochastic objective\nfunctions: normalized estimation error squared (NEES) when ground truth state\nmodels are available, or the normalized innovation error squared (NIS) when\nonly sensor data is available. By intelligently sampling the parameter space to\nboth learn and exploit a nonparametric Gaussian process surrogate function for\nthe NEES/NIS costs, Bayesian optimization can efficiently identify multiple\nlocal minima and provide uncertainty quantification on its results.\n",
"title": "Weak in the NEES?: Auto-tuning Kalman Filters with Bayesian Optimization"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
14729
| null |
Validated
| null | null |
null |
{
"abstract": " Fully reconfigurable metasurfaces would enable new classes of optical devices\nthat provide unprecedented control of electromagnetic beamforms. The principal\nchallenge for achieving reconfigurability is the need to generate large\ntunability of subwavelength, low-Q metasurface resonators. Here, we demonstrate\nlarge refractive index tuning can be efficiently facilitated at mid-infrared\nwavelengths using novel temperature-dependent control over free-carrier\nrefraction. In doped InSb we demonstrate nearly two-fold increase in the\nelectron effective mass leading to a positive refractive index shift\n({\\Delta}n>1.5) far greater than conventional thermo-optic effects. In undoped\nfilms we demonstrate more than 10-fold change in the thermal free-carrier\nconcentration producing a near-unity negative refractive index shift.\nExploiting both effects within a single resonator system, intrinsic InSb wires\non a heavily doped (epsilon near zero) InSb substrate, we demonstrate\ndynamically tunable Mie resonances. The observed larger than line-width\nresonance shifts ({\\Delta}{\\lambda}>1.5{\\mu}m) suggest new avenues for highly\ntunable and reconfigurable mid-infrared semiconductor metasurfaces.\n",
"title": "Ultra-wide plasmonic tuning of semiconductor metasurface resonators on epsilon near zero media"
}
| null | null | null | null | true | null |
14730
| null |
Default
| null | null |
null |
{
"abstract": " We present an approach to automatic detection of Alzheimer's type dementia\nbased on characteristics of spontaneous spoken language dialogue consisting of\ninterviews recorded in natural settings. The proposed method employs additive\nlogistic regression (a machine learning boosting method) on content-free\nfeatures extracted from dialogical interaction to build a predictive model. The\nmodel training data consisted of 21 dialogues between patients with Alzheimer's\nand interviewers, and 17 dialogues between patients with other health\nconditions and interviewers. Features analysed included speech rate,\nturn-taking patterns and other speech parameters. Despite relying solely on\ncontent-free features, our method obtains overall accuracy of 86.5\\%, a result\ncomparable to those of state-of-the-art methods that employ more complex\nlexical, syntactic and semantic features. While further investigation is\nneeded, the fact that we were able to obtain promising results using only\nfeatures that can be easily extracted from spontaneous dialogues suggests the\npossibility of designing non-invasive and low-cost mental health monitoring\ntools for use at scale.\n",
"title": "A Method for Analysis of Patient Speech in Dialogue for Dementia Detection"
}
| null | null |
[
"Computer Science"
] | null | true | null |
14731
| null |
Validated
| null | null |
null |
{
"abstract": " As illustrated in recent years (Superstorm Sandy, the Northeast Ice Storm of\n1998, etc.), extreme weather events pose an enormous threat to the electric\npower transmission systems and the associated socio-economic systems that\ndepend on reliable delivery of electric power. Besides inevitable malfunction\nof power grid components, deliberate malicious attacks can cause high risks to\nthe service. These threats motivate the need for approaches and methods that\nimprove the resilience of power systems. In this paper, we develop a model and\ntractable methods for optimizing the upgrade of transmission systems through a\ncombination of hardening existing components, adding redundant lines, switches,\ngenerators, and FACTS and phase-shifting devices. While many of these\ncontrollable components are included in traditional design (expansion planning)\nproblems, we uniquely assess their benefits from a resiliency point of view.\nMore importantly, perhaps, we evaluate the suitability of using\nstate-of-the-art AC power flow relaxations versus the common DC approximation\nin resilience improvement studies. The resiliency model and algorithms are\ntested on a modified version of the RTS-96 (single area) system.\n",
"title": "Resilient Transmission Grid Design: AC Relaxation vs. DC approximation"
}
| null | null | null | null | true | null |
14732
| null |
Default
| null | null |
null |
{
"abstract": " We present a new technique for learning visual-semantic embeddings for\ncross-modal retrieval. Inspired by hard negative mining, the use of hard\nnegatives in structured prediction, and ranking loss functions, we introduce a\nsimple change to common loss functions used for multi-modal embeddings. That,\ncombined with fine-tuning and use of augmented data, yields significant gains\nin retrieval performance. We showcase our approach, VSE++, on MS-COCO and\nFlickr30K datasets, using ablation studies and comparisons with existing\nmethods. On MS-COCO our approach outperforms state-of-the-art methods by 8.8%\nin caption retrieval and 11.3% in image retrieval (at R@1).\n",
"title": "VSE++: Improving Visual-Semantic Embeddings with Hard Negatives"
}
| null | null | null | null | true | null |
14733
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we investigate the cooling-off effect (opposite to the magnet\neffect) from two aspects. Firstly, from the viewpoint of dynamics, we study the\nexistence of the cooling-off effect by following the dynamical evolution of\nsome financial variables over a period of time before the stock price hits its\nlimit. Secondly, from the probability perspective, we investigate, with the\nlogit model, the existence of the cooling-off effect through analyzing the\nhigh-frequency data of all A-share common stocks traded on the Shanghai Stock\nExchange and the Shenzhen Stock Exchange from 2000 to 2011 and inspecting the\ntrading period from the opening phase prior to the moment that the stock price\nhits its limits. A comparison is made of the properties between up-limit hits\nand down-limit hits, and the possible difference will also be compared between\nbullish and bearish market state by dividing the whole period into three\nalternating bullish periods and three bearish periods. We find that the\ncooling-off effect emerges for both up-limit hits and down-limit hits, and the\ncooling-off effect of the down-limit hits is stronger than that of the up-limit\nhits. The difference of the cooling-off effect between bullish period and\nbearish period is quite modest. Moreover, we examine the sub-optimal orders\neffect, and infer that the professional individual investors and institutional\ninvestors play a positive role in the cooling-off effects. All these findings\nindicate that the price limit trading rule exerts a positive effect on\nmaintaining the stability of the Chinese stock markets.\n",
"title": "The cooling-off effect of price limits in the Chinese stock markets"
}
| null | null | null | null | true | null |
14734
| null |
Default
| null | null |
null |
{
"abstract": " We unveil a novel and unexpected manifestation of Anderson localization of\nmatter wave packets that carry a finite average velocity: after an initial\nballistic motion, the packet center-of-mass experiences a retroreflection and\nslowly returns to its initial position. We describe this effect both\nnumerically and analytically in dimension 1, and show that it is destroyed by\nweak particle interactions which act as a decoherence process. The\nretroreflection is also present in higher dimensions, provided the dynamics is\nAnderson localized.\n",
"title": "When Anderson localization makes quantum particles move backward"
}
| null | null | null | null | true | null |
14735
| null |
Default
| null | null |
null |
{
"abstract": " Experiments and simulations have established that dynamics in a class of\nliving and abiotic systems that are far from equilibrium exhibit super\ndiffusive behavior at long times, which in some cases (for example evolving\ntumor) is preceded by slow glass-like dynamics. By using the evolution of a\ncollection of tumor cells, driven by mechanical forces and subject to cell\nbirth and apoptosis, as a case study we show theoretically that on short time\nscales the mean square displacement is sub-diffusive due to jamming, whereas at\nlong times it is super diffusive. The results obtained using stochastic\nquantization method, which is needed because of the absence of\nfluctuation-dissipation theorem (FDT), show that the super-diffusive behavior\nis universal and impervious to the nature of cell-cell interactions.\nSurprisingly, the theory also quantitatively accounts for the non-trivial\ndynamics observed in simulations of a model soap foam characterized by creation\nand destruction of spherical bubbles, which suggests that the two\nnon-equilibrium systems belong to the same universality class. The theoretical\nprediction for the super diffusion exponent is in excellent agreement with\nsimulations for collective motion of tumor cells and dynamics associated with\nsoap bubbles.\n",
"title": "On the origin of super-diffusive behavior in a class of non-equilibrium systems"
}
| null | null | null | null | true | null |
14736
| null |
Default
| null | null |
null |
{
"abstract": " The transition mechanism of jump processes between two different subsets in\nstate space reveals important dynamical information of the processes and\ntherefore has attracted considerable attention in the past years. In this\npaper, we study the first passage path ensemble of both discrete-time and\ncontinuous-time jump processes on a finite state space. The main approach is to\ndivide each first passage path into nonreactive and reactive segments and to\nstudy them separately. The analysis can be applied to jump processes which are\nnon-ergodic, as well as continuous-time jump processes where the waiting time\ndistributions are non-exponential. In the particular case that the jump\nprocesses are both Markovian and ergodic, our analysis elucidates the relations\nbetween the study of the first passage paths and the study of the transition\npaths in transition path theory. We provide algorithms to numerically compute\nstatistics of the first passage path ensemble. The computational complexity of\nthese algorithms scales with the complexity of solving a linear system, for\nwhich efficient methods are available. Several examples demonstrate the wide\napplicability of the derived results across research areas.\n",
"title": "Statistical analysis of the first passage path ensemble of jump processes"
}
| null | null | null | null | true | null |
14737
| null |
Default
| null | null |
null |
{
"abstract": " We use globular cluster kinematics data, primarily from the SLUGGS survey, to\nmeasure the dark matter fraction ($f_{\\rm DM}$) and the average dark matter\ndensity ($\\left< \\rho_{\\rm DM} \\right>$) within the inner 5 effective radii\n($R_{\\rm e}$) for 32 nearby early--type galaxies (ETGs) with stellar mass log\n$(M_*/\\rm M_\\odot)$ ranging from $10.1$ to $11.8$. We compare our results with\na simple galaxy model based on scaling relations as well as with cosmological\nhydrodynamical simulations where the dark matter profile has been modified\nthrough various physical processes.\nWe find a high $f_{\\rm DM}$ ($\\geq0.6$) within 5~$R_{\\rm e}$ in most of our\nsample, which we interpret as a signature of a late mass assembly history that\nis largely devoid of gas-rich major mergers. However, around log $(M_*/M_\\odot)\n\\sim 11$, there is a wide range of $f_{\\rm DM}$ which may be challenging to\nexplain with any single cosmological model. We find tentative evidence that\nlenticulars (S0s), unlike ellipticals, have mass distributions that are similar\nto spiral galaxies, with decreasing $f_{\\rm DM}$ within 5~$R_{\\rm e}$ as galaxy\nluminosity increases. However, we do not find any difference between the\n$\\left< \\rho_{\\rm DM} \\right>$ of S0s and ellipticals in our sample, despite\nthe differences in their stellar populations. We have also used $\\left<\n\\rho_{\\rm DM} \\right>$ to infer the epoch of halo assembly ($z{\\sim}2-4$). By\ncomparing the age of their central stars with the inferred epoch of halo\nformation, we are able to gain more insight into their mass assembly histories.\nOur results suggest a fundamental difference in the dominant late-phase mass\nassembly channel between lenticulars and elliptical galaxies.\n",
"title": "The SLUGGS Survey: Dark matter fractions at large radii and assembly epochs of early-type galaxies from globular cluster kinematics"
}
| null | null | null | null | true | null |
14738
| null |
Default
| null | null |
null |
{
"abstract": " We study the dynamics of the Bogoliubov wave packet in superconductors and\ncalculate the supercurrent carried by the wave packet. We discover an anomalous\ncontribution to the supercurrent, related to the quantum metric of the Bloch\nwave function. This anomalous contribution is most important for flat or\nquasiflat bands, as exemplified by the attractive Hubbard models on the Creutz\nladder and sawtooth lattice. Our theoretical framework is general and can be\nused to study a wide variety of phenomena, such as spin transport and exciton\ntransport.\n",
"title": "Wave packet dynamics of Bogoliubov quasiparticles: quantum metric effects"
}
| null | null | null | null | true | null |
14739
| null |
Default
| null | null |
null |
{
"abstract": " The heaviest of the transuranic elements known as superheavy elements (SHE)\nare produced in nuclear reactions in a few specialized laboratories located in\nGermany, the U.S., Russia, and Japan. The history of this branch of physical\nscience provides several case studies of interest to the philosophy and\nsociology of modern science. The story of SHE illuminates the crucial notion of\nwhat constitutes a chemical element, what the criteria are for discovering a\nnew element, and how an element is assigned a name. The story also cast light\non the sometimes uneasy relationship between physics and chemistry. It is far\nfrom obvious that elements with Z > 110 exist in the same sense as oxygen or\nsodium exists. The answers are not given by nature but by international\ncommissions responsible for the criteria and evaluation of discovery claims.\nThe works of these commissions and of SHE research in general have often been\ncontroversial.\n",
"title": "The search for superheavy elements: Historical and philosophical perspectives"
}
| null | null |
[
"Physics"
] | null | true | null |
14740
| null |
Validated
| null | null |
null |
{
"abstract": " Speckle reduction is a longstanding topic in synthetic aperture radar (SAR)\nimaging. Since most current and planned SAR imaging satellites operate in\npolarimetric, interferometric or tomographic modes, SAR images are\nmulti-channel and speckle reduction techniques must jointly process all\nchannels to recover polarimetric and interferometric information. The\ndistinctive nature of SAR signal (complex-valued, corrupted by multiplicative\nfluctuations) calls for the development of specialized methods for speckle\nreduction. Image denoising is a very active topic in image processing with a\nwide variety of approaches and many denoising algorithms available, almost\nalways designed for additive Gaussian noise suppression. This paper proposes a\ngeneral scheme, called MuLoG (MUlti-channel LOgarithm with Gaussian denoising),\nto include such Gaussian denoisers within a multi-channel SAR speckle reduction\ntechnique. A new family of speckle reduction algorithms can thus be obtained,\nbenefiting from the ongoing progress in Gaussian denoising, and offering\nseveral speckle reduction results often displaying method-specific artifacts\nthat can be dismissed by comparison between results.\n",
"title": "MuLoG, or How to apply Gaussian denoisers to multi-channel SAR speckle reduction?"
}
| null | null | null | null | true | null |
14741
| null |
Default
| null | null |
null |
{
"abstract": " We define the Radon transform functor for sheaves and prove that it is an\nequivalence after suitable microlocal localizations. As a result, the sheaf\ncategory associated to a Legendrian is invariant under the Radon transform. We\nalso manage to place the Radon transform and other transforms in microlocal\nsheaf theory altogether in a diagram.\n",
"title": "Radon Transform for Sheaves"
}
| null | null | null | null | true | null |
14742
| null |
Default
| null | null |
null |
{
"abstract": " We propose foundations for a synthetic theory of $(\\infty,1)$-categories\nwithin homotopy type theory. We axiomatize a directed interval type, then\ndefine higher simplices from it and use them to probe the internal categorical\nstructures of arbitrary types. We define Segal types, in which binary\ncomposites exist uniquely up to homotopy; this automatically ensures\ncomposition is coherently associative and unital at all dimensions. We define\nRezk types, in which the categorical isomorphisms are additionally equivalent\nto the type-theoretic identities - a \"local univalence\" condition. And we\ndefine covariant fibrations, which are type families varying functorially over\na Segal type, and prove a \"dependent Yoneda lemma\" that can be viewed as a\ndirected form of the usual elimination rule for identity types. We conclude by\nstudying homotopically correct adjunctions between Segal types, and showing\nthat for a functor between Rezk types to have an adjoint is a mere proposition.\nTo make the bookkeeping in such proofs manageable, we use a three-layered\ntype theory with shapes, whose contexts are extended by polytopes within\ndirected cubes, which can be abstracted over using \"extension types\" that\ngeneralize the path-types of cubical type theory. In an appendix, we describe\nthe motivating semantics in the Reedy model structure on bisimplicial sets, in\nwhich our Segal and Rezk types correspond to Segal spaces and complete Segal\nspaces.\n",
"title": "A type theory for synthetic $\\infty$-categories"
}
| null | null | null | null | true | null |
14743
| null |
Default
| null | null |
null |
{
"abstract": " High-order parametric models that include terms for feature interactions are\napplied to various data mining tasks, where ground truth depends on\ninteractions of features. However, with sparse data, the high- dimensional\nparameters for feature interactions often face three issues: expensive\ncomputation, difficulty in parameter estimation and lack of structure. Previous\nwork has proposed approaches which can partially re- solve the three issues. In\nparticular, models with factorized parameters (e.g. Factorization Machines) and\nsparse learning algorithms (e.g. FTRL-Proximal) can tackle the first two issues\nbut fail to address the third. Regarding to unstructured parameters,\nconstraints or complicated regularization terms are applied such that\nhierarchical structures can be imposed. However, these methods make the\noptimization problem more challenging. In this work, we propose Strongly\nHierarchical Factorization Machines and ANOVA kernel regression where all the\nthree issues can be addressed without making the optimization problem more\ndifficult. Experimental results show the proposed models significantly\noutperform the state-of-the-art in two data mining tasks: cold-start user\nresponse time prediction and stock volatility prediction.\n",
"title": "Strongly Hierarchical Factorization Machines and ANOVA Kernel Regression"
}
| null | null |
[
"Computer Science"
] | null | true | null |
14744
| null |
Validated
| null | null |
null |
{
"abstract": " In this article, I discuss the relationship of mathematics to the physical\nworld, and to other spheres of human knowledge. In particular, I argue that\nMathematics is created by human beings, and the number $\\pi$ can not be said to\nhave existed $100,000$ years ago, using the conventional meaning of the word\n`exist'.\n",
"title": "On the relationship of Mathematics to the real world"
}
| null | null | null | null | true | null |
14745
| null |
Default
| null | null |
null |
{
"abstract": " We extend the constructive dependent type theory of the Logical Framework\n$\\mathsf{LF}$ with monadic, dependent type constructors indexed with predicates\nover judgements, called Locks. These monads capture various possible proof\nattitudes in establishing the judgment of the object logic encoded by an\n$\\mathsf{LF}$ type. Standard examples are factoring-out the verification of a\nconstraint or delegating it to an external oracle, or supplying some\nnon-apodictic epistemic evidence, or simply discarding the proof witness of a\nprecondition deeming it irrelevant. This new framework, called Lax Logical\nFramework, $\\mathsf{LLF}_{\\cal P}$, is a conservative extension of\n$\\mathsf{LF}$, and hence it is the appropriate metalanguage for dealing\nformally with side-conditions in rules or external evidence in logical systems.\n$\\mathsf{LLF}_{\\cal P}$ arises once the monadic nature of the lock\ntype-constructor, ${\\cal L}^{\\cal P}_{M,\\sigma}[\\cdot]$, introduced by the\nauthors in a series of papers, together with Marina Lenisa, is fully exploited.\nThe nature of the lock monads permits to utilize the very Lock destructor,\n${\\cal U}^{\\cal P}_{M,\\sigma}[\\cdot]$, in place of Moggi's monadic $let_T$,\nthus simplifying the equational theory. The rules for ${\\cal U}^{\\cal\nP}_{M,\\sigma}[\\cdot]$ permit also the removal of the monad once the constraint\nis satisfied. We derive the meta-theory of $\\mathsf{LLF}_{\\cal P}$ by a novel\nindirect method based on the encoding of $\\mathsf{LLF}_{\\cal P}$ in\n$\\mathsf{LF}$. We discuss encodings in $\\mathsf{LLF}_{\\cal P}$ of call-by-value\n$\\lambda$-calculi, Hoare's Logic, and Fitch-Prawitz Naive Set Theory.\n",
"title": "$\\mathsf{LLF}_{\\cal P}$: a logical framework for modeling external evidence, side conditions, and proof irrelevance using monads"
}
| null | null | null | null | true | null |
14746
| null |
Default
| null | null |
null |
{
"abstract": " In this article, we give the explicit solutions to the Laplace equations\nassociated to the Dirac operator, Euler operator and the harmonic oscillator in\nR.\n",
"title": "Laplace equation for the Dirac, Euler and the harmonic oscillator"
}
| null | null | null | null | true | null |
14747
| null |
Default
| null | null |
null |
{
"abstract": " We quantify the accuracy of various simulators compared to a real world\nrobotic reaching and interaction task. Simulators are used in robotics to\ndesign solutions for real world hardware without the need for physical access.\nThe `reality gap' prevents solutions developed or learnt in simulation from\nperforming well, or at at all, when transferred to real-world hardware. Making\nuse of a Kinova robotic manipulator and a motion capture system, we record a\nground truth enabling comparisons with various simulators, and present\nquantitative data for various manipulation-oriented robotic tasks. We show the\nrelative strengths and weaknesses of numerous contemporary simulators,\nhighlighting areas of significant discrepancy, and assisting researchers in the\nfield in their selection of appropriate simulators for their use cases. All\ncode and parameter listings are publicly available from:\nthis https URL .\n",
"title": "Quantifying the Reality Gap in Robotic Manipulation Tasks"
}
| null | null | null | null | true | null |
14748
| null |
Default
| null | null |
null |
{
"abstract": " Let $q$ be a power of a prime $p$ and let $U(q)$ be a Sylow $p$-subgroup of a\nfinite Chevalley group $G(q)$ defined over the field with $q$ elements. We\nfirst give a parametrization of the set $\\text{Irr}(U(q))$ of irreducible\ncharacters of $U(q)$ when $G(q)$ is of type $\\mathrm{G}_2$. This is uniform for\nprimes $p \\ge 5$, while the bad primes $p=2$ and $p=3$ have to be considered\nseparately. We then use this result and the contribution of several authors to\nshow a general result, namely that if $G(q)$ is any finite Chevalley group with\n$p$ a bad prime, then there exists a character $\\chi \\in \\text{Irr}(U(q))$ such\nthat $\\chi(1)=q^n/p$ for some $n \\in \\mathbb{Z}_{\\ge_0}$. In particular, for\neach $G(q)$ and every bad prime $p$, we construct a family of characters of\nsuch degree as inflation followed by an induction of linear characters of an\nabelian subquotient $V(q)$ of $U(q)$.\n",
"title": "On the character degrees of a Sylow $p$-subgroup of a finite Chevalley group $G(p^f)$ over a bad prime"
}
| null | null | null | null | true | null |
14749
| null |
Default
| null | null |
null |
{
"abstract": " We deal with zero-delay source coding of a vector-valued Gauss-Markov source\nsubject to a mean-squared error (MSE) fidelity criterion characterized by the\noperational zero-delay vector-valued Gaussian rate distortion function (RDF).\nWe address this problem by considering the nonanticipative RDF (NRDF) which is\na lower bound to the causal optimal performance theoretically attainable (OPTA)\nfunction and operational zero-delay RDF. We recall the realization that\ncorresponds to the optimal \"test-channel\" of the Gaussian NRDF, when\nconsidering a vector Gauss-Markov source subject to a MSE distortion in the\nfinite time horizon. Then, we introduce sufficient conditions to show existence\nof solution for this problem in the infinite time horizon. For the asymptotic\nregime, we use the asymptotic characterization of the Gaussian NRDF to provide\na new equivalent realization scheme with feedback which is characterized by a\nresource allocation (reverse-waterfilling) problem across the dimension of the\nvector source. We leverage the new realization to derive a predictive coding\nscheme via lattice quantization with subtractive dither and joint memoryless\nentropy coding. This coding scheme offers an upper bound to the operational\nzero-delay vector-valued Gaussian RDF. When we use scalar quantization, then\nfor \"r\" active dimensions of the vector Gauss-Markov source the gap between the\nobtained lower and theoretical upper bounds is less than or equal to 0.254r + 1\nbits/vector. We further show that it is possible when we use vector\nquantization, and assume infinite dimensional Gauss-Markov sources to make the\nprevious gap to be negligible, i.e., Gaussian NRDF approximates the operational\nzero-delay Gaussian RDF. We also extend our results to vector-valued Gaussian\nsources of any finite memory under mild conditions. Our theoretical framework\nis demonstrated with illustrative numerical experiments.\n",
"title": "Zero-Delay Rate Distortion via Filtering for Vector-Valued Gaussian Sources"
}
| null | null | null | null | true | null |
14750
| null |
Default
| null | null |
null |
{
"abstract": " We develop a unified description, via the Boltzmann equation, of damping of\ngravitational waves by matter, incorporating collisions. We identify two\nphysically distinct damping mechanisms -- collisional and Landau damping. We\nfirst consider damping in flat spacetime, and then generalize the results to\nallow for cosmological expansion. In the first regime, maximal collisional\ndamping of a gravitational wave, independent of the details of the collisions\nin the matter is, as we show, significant only when its wavelength is\ncomparable to the size of the horizon. Thus damping by intergalactic or\ninterstellar matter for all but primordial gravitational radiation can be\nneglected. Although collisions in matter lead to a shear viscosity, they also\nact to erase anisotropic stresses, thus suppressing the damping of\ngravitational waves. Damping of primordial gravitational waves remains\npossible. We generalize Weinberg's calculation of gravitational wave damping,\nnow including collisions and particles of finite mass, and interpret the\ncollisionless limit in terms of Landau damping. While Landau damping of\ngravitational waves cannot occur in flat spacetime, the expansion of the\nuniverse allows such damping by spreading the frequency of a gravitational wave\nof given wavevector.\n",
"title": "Damping of gravitational waves by matter"
}
| null | null | null | null | true | null |
14751
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we define a notion of calibration for an equivalent approach to\nthe classical Steiner problem in a covering space setting and we give some\nexplicit examples. Moreover we introduce the notion of calibration in families:\nthe idea is to divide the set of competitors in a suitable way, defining an\nappropriate (and weaker) notion of calibration. Then, calibrating the candidate\nminimizers in each family and comparing their perimeter, it is possible to find\nthe minimizers of the minimization problem. Thanks to this procedure we prove\nthe minimality of the Steiner configurations spanning the vertices of a regular\nhexagon and of a regular pentagon.\n",
"title": "Calibrations for minimal networks in a covering space setting"
}
| null | null | null | null | true | null |
14752
| null |
Default
| null | null |
null |
{
"abstract": " We present the design and manufacturing of high fidelity universal 3D\nfingerprint targets, which can be imaged on a variety of fingerprint sensing\ntechnologies, namely capacitive, contact-optical, and contactless-optical.\nUniversal 3D fingerprint targets enable, for the first time, not only a\nrepeatable and controlled evaluation of fingerprint readers, but also the\nability to conduct fingerprint reader interoperability studies. Fingerprint\nreader interoperability refers to how robust fingerprint recognition systems\nare to variations in the images acquired by different types of fingerprint\nreaders. To build universal 3D fingerprint targets, we adopt a molding and\ncasting framework consisting of (i) digital mapping of fingerprint images to a\nnegative mold, (ii) CAD modeling a scaffolding system to hold the negative\nmold, (iii) fabricating the mold and scaffolding system with a high resolution\n3D printer, (iv) producing or mixing a material with similar electrical,\noptical, and mechanical properties to that of the human finger, and (v)\nfabricating a 3D fingerprint target using controlled casting. Our experiments\nconducted with PIV and Appendix F certified optical (contact and contactless)\nand capacitive fingerprint readers demonstrate the usefulness of universal 3D\nfingerprint targets for controlled and repeatable fingerprint reader\nevaluations and also fingerprint reader interoperability studies.\n",
"title": "Universal 3D Wearable Fingerprint Targets: Advancing Fingerprint Reader Evaluations"
}
| null | null | null | null | true | null |
14753
| null |
Default
| null | null |
null |
{
"abstract": " We consider the dynamics of message passing for spatially coupled codes and,\nin particular, the set of density evolution equations that tracks the profile\nof decoding errors along the spatial direction of coupling. It is known that,\nfor suitable boundary conditions and after a transient phase, the error profile\nexhibits a \"solitonic behavior\". Namely, a uniquely-shaped wavelike solution\ndevelops, that propagates with constant velocity. Under this assumption we\nderive an analytical formula for the velocity in the framework of a continuum\nlimit of the spatially coupled system. The general formalism is developed for\nspatially coupled low-density parity-check codes on general binary memoryless\nsymmetric channels which form the main system of interest in this work. We\napply the formula for special channels and illustrate that it matches the\ndirect numerical evaluation of the velocity for a wide range of noise values. A\npossible application of the velocity formula to the evaluation of finite size\nscaling law parameters is also discussed. We conduct a similar analysis for\ngeneral scalar systems and illustrate the findings with applications to\ncompressive sensing and generalized low-density parity-check codes on the\nbinary erasure or binary symmetric channels.\n",
"title": "The Velocity of the Propagating Wave for Spatially Coupled Systems with Applications to LDPC Codes"
}
| null | null |
[
"Computer Science",
"Physics",
"Mathematics"
] | null | true | null |
14754
| null |
Validated
| null | null |
null |
{
"abstract": " In the first part of this paper we will prove the Voevodsky's nilpotence\nconjecture for smooth cubic fourfolds and ordinary generic Gushel-Mukai\nfourfolds. Then, making use of noncommutative motives, we will prove the\nVoevodsky's nilpotence conjecture for generic Gushel-Mukai fourfolds containing\na $\\tau$-plane $\\G(2,3)$ and for ordinary Gushel-Mukai fourfolds containing a\nquintic del Pezzo surface.\n",
"title": "Voevodsky's conjecture for cubic fourfolds and Gushel-Mukai fourfolds via noncommutative K3 surfaces"
}
| null | null | null | null | true | null |
14755
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we propose the first computationally efficient projection-free\nalgorithm for bandit convex optimization (BCO). We show that our algorithm\nachieves a sublinear regret of $O(nT^{4/5})$ (where $T$ is the horizon and $n$\nis the dimension) for any bounded convex functions with uniformly bounded\ngradients. We also evaluate the performance of our algorithm against baselines\non both synthetic and real data sets for quadratic programming, portfolio\nselection and matrix completion problems.\n",
"title": "Projection-Free Bandit Convex Optimization"
}
| null | null | null | null | true | null |
14756
| null |
Default
| null | null |
null |
{
"abstract": " We have already developed the recommendation system of sightseeing\ninformation on SNS by using smartphone based user participatory sensing system.\nThe system can post the attractive information for tourists to the specified\nFacebook page by our developed smartphone application. The users in Facebook,\nwho are interested in sightseeing, can come flocking through information space\nfrom far and near. However, the activities in the community on SNS are only\nsupported by the specified people called a hub. We proposed the method of\nvitalization of tourist behaviors to give a stimulus to the people. We\ndeveloped the simulation system for multi agent system with altruistic\nbehaviors inspired by the Army Ants. The army ant takes feeding action with\naltruistic behaviors to suppress selfish behavior to a common object used by a\nplurality of users in common. In this paper, we introduced the altruism\nbehavior determined by some simulation to vitalize the SNS community. The\nefficiency of the revitalization process of the community was investigated by\nsome experimental simulation results.\n",
"title": "The Recommendation System to SNS Community for Tourists by Using Altruistic Behaviors"
}
| null | null | null | null | true | null |
14757
| null |
Default
| null | null |
null |
{
"abstract": " A seminal result in decentralized control is the development of fixed modes\nby Wang and Davison in 1973 - that plant modes which cannot be moved with a\nstatic decentralized controller cannot be moved by a dynamic one either, and\nthat the other modes which can be moved can be shifted to any chosen location\nwith arbitrary precision. These results were developed for perfectly\ndecentralized, or block diagonal, information structure, where each control\ninput may only depend on a single corresponding measurement. Furthermore, the\nresults were claimed after a preliminary step was demonstrated, omitting a\nrigorous induction for each of these results, and the remaining task is\nnontrivial.\nIn this paper, we consider fixed modes for arbitrary information structures,\nwhere certain control inputs may depend on some measurements but not others. We\nprovide a comprehensive proof that the modes which cannot be altered by a\nstatic controller with the given structure cannot be moved by a dynamic one\neither, and that the modes which can be altered by a static controller with the\ngiven structure can be moved by a dynamic one to any chosen location with\narbitrary precision, thus generalizing and solidifying Wang and Davison's\nresults.\nThis shows that a system can be stabilized by a linear time-invariant\ncontroller with the given information structure as long as all of the modes\nwhich are fixed with respect to that structure are in the left half-plane; an\nalgorithm for synthesizing such a stabilizing decentralized controller is then\ndistilled from the proof.\n",
"title": "Constructive Stabilization and Pole Placement by Arbitrary Decentralized Architectures"
}
| null | null | null | null | true | null |
14758
| null |
Default
| null | null |
null |
{
"abstract": " Optical clocks benefit from tight atomic confinement enabling extended\ninterrogation times as well as Doppler- and recoil-free operation. However,\nthese benefits come at the cost of frequency shifts that, if not properly\ncontrolled, may degrade clock accuracy. Numerous theoretical studies have\npredicted optical lattice clock frequency shifts that scale nonlinearly with\ntrap depth. To experimentally observe and constrain these shifts in an\n$^{171}$Yb optical lattice clock, we construct a lattice enhancement cavity\nthat exaggerates the light shifts. We observe an atomic temperature that is\nproportional to the optical trap depth, fundamentally altering the scaling of\ntrap-induced light shifts and simplifying their parametrization. We identify an\n\"operational\" magic wavelength where frequency shifts are insensitive to\nchanges in trap depth. These measurements and scaling analysis constitute an\nessential systematic characterization for clock operation at the $10^{-18}$\nlevel and beyond.\n",
"title": "Hyperpolarizability and operational magic wavelength in an optical lattice clock"
}
| null | null | null | null | true | null |
14759
| null |
Default
| null | null |
null |
{
"abstract": " A special type of rotary-wing Unmanned Aerial Vehicles (UAV), called\nQuadcopter have prevailed to the civilian use for the past decade. They have\ngained significant amount of attention within the UAV community for their\nredundancy and ease of control, despite the fact that they fall under an\nunder-actuated system category. They come in a variety of configurations. The\n\"+\" and \"x\" configurations were introduced first. Literature pertinent to these\ntwo configurations is vast. However, in this paper, we define 6 additional\npossible configurations for a Quadcopter that can be built under either \"+\" or\n\"x\" setup. These configurations can be achieved by changing the angle that the\naxis of rotation for rotors make with the main body, i.e., fuselage. This would\nalso change the location of the COM with respect to the propellers which can\nadd to the overall stability. A comprehensive dynamic model for all these\nconfigurations is developed for the first time. The overall stability for these\nconfigurations are addressed. In particular, it is shown that one configuration\ncan lead to the most statically-stable platform by adopting damping motion in\nRoll/Pitch/Yaw, which is described for the first time to the best of our\nknowledge.\n",
"title": "A Unified Approach to Configuration-based Dynamic Analysis of Quadcopters for Optimal Stability"
}
| null | null | null | null | true | null |
14760
| null |
Default
| null | null |
null |
{
"abstract": " Traffic congestion is a widespread problem. Dynamic traffic routing systems\nand congestion pricing are getting importance in recent research. Lane\nprediction and vehicle density estimation is an important component of such\nsystems. We introduce a novel problem of vehicle self-positioning which\ninvolves predicting the number of lanes on the road and vehicle's position in\nthose lanes using videos captured by a dashboard camera. We propose an\nintegrated closed-loop approach where we use the presence of vehicles to aid\nthe task of self-positioning and vice-versa. To incorporate multiple factors\nand high-level semantic knowledge into the solution, we formulate this problem\nas a Bayesian framework. In the framework, the number of lanes, the vehicle's\nposition in those lanes and the presence of other vehicles are considered as\nparameters. We also propose a bounding box selection scheme to reduce the\nnumber of false detections and increase the computational efficiency. We show\nthat the number of box proposals decreases by a factor of 6 using the selection\napproach. It also results in large reduction in the number of false detections.\nThe entire approach is tested on real-world videos and is found to give\nacceptable results.\n",
"title": "Improving Vision-based Self-positioning in Intelligent Transportation Systems via Integrated Lane and Vehicle Detection"
}
| null | null | null | null | true | null |
14761
| null |
Default
| null | null |
null |
{
"abstract": " In this note, we expand on some technical issues raised in \\cite{PPV} by the\nauthors, as well as providing a friendly introduction to and summary of our\nprevious work. We construct a set of heterotic string compactifications to 0+1\ndimensions intimately related to the Monstrous moonshine module of Frenkel,\nLepowsky, and Meurman (and orbifolds thereof). Using this model, we review our\nphysical interpretation of the genus zero property of Monstrous moonshine.\nFurthermore, we show that the space of (second-quantized) BPS-states forms a\nmodule over the Monstrous Lie algebras $\\mathfrak{m}_g$---some of the first and\nmost prominent examples of Generalized Kac-Moody algebras---constructed by\nBorcherds and Carnahan. In particular, we clarify the structure of the module\npresent in the second-quantized string theory. We also sketch a proof of our\nmethods in the language of vertex operator algebras, for the interested\nmathematician.\n",
"title": "BPS Algebras, Genus Zero, and the Heterotic Monster"
}
| null | null |
[
"Mathematics"
] | null | true | null |
14762
| null |
Validated
| null | null |
null |
{
"abstract": " Nontrivial connectivity has allowed the training of very deep networks by\naddressing the problem of vanishing gradients and offering a more efficient\nmethod of reusing parameters. In this paper we make a comparison between\nresidual networks, densely-connected networks and highway networks on an image\nclassification task. Next, we show that these methodologies can easily be\ndeployed into automatic speech recognition and provide significant improvements\nto existing models.\n",
"title": "Exploiting Nontrivial Connectivity for Automatic Speech Recognition"
}
| null | null | null | null | true | null |
14763
| null |
Default
| null | null |
null |
{
"abstract": " There is a significant literature on methods for incorporating knowledge into\nmultiple testing procedures so as to improve their power and precision. Some\ncommon forms of prior knowledge include (a) beliefs about which hypotheses are\nnull, modeled by non-uniform prior weights; (b) differing importances of\nhypotheses, modeled by differing penalties for false discoveries; (c) multiple\narbitrary partitions of the hypotheses into (possibly overlapping) groups; and\n(d) knowledge of independence, positive or arbitrary dependence between\nhypotheses or groups, suggesting the use of more aggressive or conservative\nprocedures. We present a unified algorithmic framework called p-filter for\nglobal null testing and false discovery rate (FDR) control that allows the\nscientist to incorporate all four types of prior knowledge (a)-(d)\nsimultaneously, recovering a variety of known algorithms as special cases.\n",
"title": "A unified treatment of multiple testing with prior knowledge using the p-filter"
}
| null | null | null | null | true | null |
14764
| null |
Default
| null | null |
null |
{
"abstract": " Mahlmann and Schindelhauer (2005) defined a Markov chain which they called\n$k$-Flipper, and showed that it is irreducible on the set of all connected\nregular graphs of a given degree (at least 3). We study the 1-Flipper chain,\nwhich we call the flip chain, and prove that the flip chain converges rapidly\nto the uniform distribution over connected $2r$-regular graphs with $n$\nvertices, where $n\\geq 8$ and $r = r(n)\\geq 2$. Formally, we prove that the\ndistribution of the flip chain will be within $\\varepsilon$ of uniform in total\nvariation distance after $\\text{poly}(n,r,\\log(\\varepsilon^{-1}))$ steps. This\npolynomial upper bound on the mixing time is given explicitly, and improves\nmarkedly on a previous bound given by Feder et al.(2006). We achieve this\nimprovement by using a direct two-stage canonical path construction, which we\ndefine in a general setting.\nThis work has applications to decentralised networks based on random regular\nconnected graphs of even degree, as a self-stabilising protocol in which nodes\nspontaneously perform random flips in order to repair the network.\n",
"title": "The flip Markov chain for connected regular graphs"
}
| null | null |
[
"Computer Science",
"Mathematics"
] | null | true | null |
14765
| null |
Validated
| null | null |
null |
{
"abstract": " Simultaneous Localization And Mapping (SLAM) is the problem of constructing\nor updating a map of an unknown environment while simultaneously keeping track\nof an agent's location within it. How to enable SLAM robustly and durably on\nmobile, or even IoT grade devices, is the main challenge faced by the industry\ntoday. The main problems we need to address are: 1.) how to accelerate the SLAM\npipeline to meet real-time requirements; and 2.) how to reduce SLAM energy\nconsumption to extend battery life. After delving into the problem, we found\nout that feature extraction is indeed the bottleneck of performance and energy\nconsumption. Hence, in this paper, we design, implement, and evaluate a\nhardware ORB feature extractor and prove that our design is a great balance\nbetween performance and energy consumption compared with ARM Krait and Intel\nCore i5.\n",
"title": "FPGA-based ORB Feature Extraction for Real-Time Visual SLAM"
}
| null | null | null | null | true | null |
14766
| null |
Default
| null | null |
null |
{
"abstract": " Noting the importance of the latent variables in inference and learning, we\npropose a novel framework for autoencoders based on the homeomorphic\ntransformation of latent variables --- which could reduce the distance between\nvectors in the transformed space, while preserving the topological properties\nof the original space --- and investigate the effect of the transformation in\nboth learning generative models and denoising corrupted data. The results of\nour experiments show that the proposed model can work as both a generative\nmodel and a denoising model with improved performance due to the transformation\ncompared to conventional variational and denoising autoencoders.\n",
"title": "On the Transformation of Latent Space in Autoencoders"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
14767
| null |
Validated
| null | null |
null |
{
"abstract": " In the present work weighted area integral means $M_{p,\\varphi}(f;{\\mathrm\n{Im}}z)$ are studied and it is proved that the function $y\\to \\log\nM_{p,\\varphi}(f;y)$ is convex in the case when $f$ belongs to a Hardy space on\nthe upper half-plane.\n",
"title": "Log-Convexity of Weighted Area Integral Means of $H^p$ Functions on the Upper Half-plan"
}
| null | null | null | null | true | null |
14768
| null |
Default
| null | null |
null |
{
"abstract": " Based on first-principles calculations and effective model analysis, we\npropose that the WC-type HfC, in the absence of spin-orbit coupling (SOC), can\nhost a three-dimensional nodal chain semimetal state. Distinguished from the\nprevious material IrF4 [T. Bzdusek et al., Nature 538, 75 (2016)], the nodal\nchain here is protected by mirror reflection symmetries of a simple space\ngroup, while in IrF4 the nonsymmorphic space group with a glide plane is a\nnecessity. Moreover, in the presence of SOC, the nodal chain in WC type HfC\nevolves into Weyl points. In the Brillouin zone, a total of 30 pairs of Weyl\npoints in three types are obtained through the first-principles calculations.\nBesides, the surface states and the pattern of the surface Fermi arcs\nconnecting these Weyl points are studied, which may be measured by future\nexperiments.\n",
"title": "From Nodal Chain Semimetal To Weyl Semimetal in HfC"
}
| null | null |
[
"Physics"
] | null | true | null |
14769
| null |
Validated
| null | null |
null |
{
"abstract": " Water plays a major role in bio-systems, greatly contributing to determine\ntheir structure, stability and even function. It is well know, for instance,\nthat proteins require a minimum amount of water to be functionally active.\nSince the biological functions of proteins involve changes of conformation, and\nsometimes chemical reactions, it is natural to expect a connection of these\nfunctions with dynamical properties of the coupled system of proteins and their\nhydration water. However, despite many years of intensive research, the\ndetailed nature of protein - hydration water interactions, and their effect on\nthe biochemical activity of proteins through peculiar dynamical effects, is\nstill partly unknown. In particular, models proposed so far, fail to explain\nthe full set of experimental data. The well-accepted 'protein dynamical\ntransition' scenario is based on perfect coupling between the dynamics of\nproteins and of their hydration water, which has never been confuted\nexperimentally. We present high-energy resolution elastic neutron scattering\nmeasurements of the atomistic dynamics of the model protein, lysozyme, in water\nthat were carried out on the IN16B spectrometer at the Institut Laue-Langevin\nin Grenoble, France. These show for the first time that the dynamics of\nproteins and of their hydration water are actually de-coupled. This important\nresult militates against the well-accepted scenario, and requires a new model\nto link protein dynamics to the dynamics of its hydration water, and, in turn,\nto biochemical function.\n",
"title": "Protein and hydration-water dynamics are decoupled: A new model connecting dynamics and biochemical function is required"
}
| null | null | null | null | true | null |
14770
| null |
Default
| null | null |
null |
{
"abstract": " Let $\\mathfrak{g}$ be a hyperbolic Kac-Moody algebra of rank $2$, and set\n$\\lambda: = \\Lambda_1 - \\Lambda_2$, where $\\Lambda_1, \\Lambda_2$ are the\nfundamental weights for $\\mathfrak{g}$; note that $\\lambda$ is neither dominant\nnor antidominant. Let $\\mathbb{B}(\\lambda)$ be the crystal of all\nLakshmibai-Seshadri paths of shape $\\lambda$. We prove that (the crystal graph\nof) $\\mathbb{B}(\\lambda)$ is connected. Furthermore, we give an explicit\ndescription of Lakshmibai-Seshadri paths of shape $\\lambda$.\n",
"title": "Lakshmibai-Seshadri paths for hyperbolic Kac-Moody algebras of rank $2$"
}
| null | null |
[
"Mathematics"
] | null | true | null |
14771
| null |
Validated
| null | null |
null |
{
"abstract": " Affine policies (or control) are widely used as a solution approach in\ndynamic optimization where computing an optimal adjustable solution is usually\nintractable. While the worst case performance of affine policies can be\nsignificantly bad, the empirical performance is observed to be near-optimal for\na large class of problem instances. For instance, in the two-stage dynamic\nrobust optimization problem with linear covering constraints and uncertain\nright hand side, the worst-case approximation bound for affine policies is\n$O(\\sqrt m)$ that is also tight (see Bertsimas and Goyal (2012)), whereas\nobserved empirical performance is near-optimal. In this paper, we aim to\naddress this stark-contrast between the worst-case and the empirical\nperformance of affine policies. In particular, we show that affine policies\ngive a good approximation for the two-stage adjustable robust optimization\nproblem with high probability on random instances where the constraint\ncoefficients are generated i.i.d. from a large class of distributions; thereby,\nproviding a theoretical justification of the observed empirical performance. On\nthe other hand, we also present a distribution such that the performance bound\nfor affine policies on instances generated according to that distribution is\n$\\Omega(\\sqrt m)$ with high probability; however, the constraint coefficients\nare not i.i.d.. This demonstrates that the empirical performance of affine\npolicies can depend on the generative model for instances.\n",
"title": "Beyond Worst-case: A Probabilistic Analysis of Affine Policies in Dynamic Optimization"
}
| null | null | null | null | true | null |
14772
| null |
Default
| null | null |
null |
{
"abstract": " We give the sharp conditions for boundedness of Hausdorff operators on\ncertain modulation and Wiener amalgam spaces.\n",
"title": "Hausdorff operators on modulation and Wiener amalgam spaces"
}
| null | null | null | null | true | null |
14773
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we introduce transformations of deep rectifier networks,\nenabling the conversion of deep rectifier networks into shallow rectifier\nnetworks. We subsequently prove that any rectifier net of any depth can be\nrepresented by a maximum of a number of functions that can be realized by a\nshallow network with a single hidden layer. The transformations of both deep\nrectifier nets and deep residual nets are conducted to demonstrate the\nadvantages of the residual nets over the conventional neural nets and the\nadvantages of the deep neural nets over the shallow neural nets. In summary,\nfor two rectifier nets with different depths but with same total number of\nhidden units, the corresponding single hidden layer representation of the\ndeeper net is much more complex than the corresponding single hidden\nrepresentation of the shallower net. Similarly, for a residual net and a\nconventional rectifier net with the same structure except for the skip\nconnections in the residual net, the corresponding single hidden layer\nrepresentation of the residual net is much more complex than the corresponding\nsingle hidden layer representation of the conventional net.\n",
"title": "From Deep to Shallow: Transformations of Deep Rectifier Networks"
}
| null | null | null | null | true | null |
14774
| null |
Default
| null | null |
null |
{
"abstract": " Convolutional Neural Networks (CNNs) have been successfully applied to many\ncomputer vision tasks, such as image classification. By performing linear\ncombinations and element-wise nonlinear operations, these networks can be\nthought of as extracting solely first-order information from an input image. In\nthe past, however, second-order statistics computed from handcrafted features,\ne.g., covariances, have proven highly effective in diverse recognition tasks.\nIn this paper, we introduce a novel class of CNNs that exploit second-order\nstatistics. To this end, we design a series of new layers that (i) extract a\ncovariance matrix from convolutional activations, (ii) compute a parametric,\nsecond-order transformation of a matrix, and (iii) perform a parametric\nvectorization of a matrix. These operations can be assembled to form a\nCovariance Descriptor Unit (CDU), which replaces the fully-connected layers of\nstandard CNNs. Our experiments demonstrate the benefits of our new\narchitecture, which outperform the first-order CNNs, while relying on up to 90%\nfewer parameters.\n",
"title": "Second-order Convolutional Neural Networks"
}
| null | null | null | null | true | null |
14775
| null |
Default
| null | null |
null |
{
"abstract": " We prove the superhedging duality for a discrete-time financial market with\nproportional transaction costs under portfolio constraints and model\nuncertainty. Frictions are modeled through solvency cones as in the original\nmodel of [Kabanov, Y., Hedging and liquidation under transaction costs in\ncurrency markets. Fin. Stoch., 3(2):237-248, 1999] adapted to the quasi-sure\nsetup of [Bouchard, B. and Nutz, M., Arbitrage and duality in nondominated\ndiscrete-time models. Ann. Appl. Probab., 25(2):823-859, 2015]. Our results\nhold under the condition of No Strict Arbitrage and under the efficient\nfriction hypothesis.\n",
"title": "On the quasi-sure superhedging duality with frictions"
}
| null | null | null | null | true | null |
14776
| null |
Default
| null | null |
null |
{
"abstract": " We take advantage of the Gaia-ESO Survey iDR4 bulge data to search for\nabundance anomalies that could shed light on the composite nature of the Milky\nWay bulge. The alpha-elements (Mg, Si, and whenever available, Ca) abundances,\nand their trends with Fe abundances have been analysed for a total of 776 bulge\nstars. In addition, the aluminum abundances and their ratio to Fe and Mg have\nalso been examined. Our analysis reveals the existence of low-alpha element\nabundance stars with respect to the standard bulge sequence in the [alpha/Fe]\nvs. [Fe/H] plane. 18 objects present deviations in [alpha/Fe] ranging from 2.1\nto 5.3 sigma with respect to the median standard value. Those stars do not show\nMg-Al anti-correlation patterns. Incidentally, this sign of the existence of\nmultiple stellar populations is reported firmly for the first time for the\nbulge globular cluster NGC 6522. The identified low-alpha abundance stars have\nchemical patterns compatible with those of the thin disc. Their link with\nmassive dwarf galaxies accretion seems unlikely, as larger deviations in alpha\nabundance and Al would be expected. The vision of a bulge composite nature and\na complex formation process is reinforced by our results. The used approach, a\nmulti-method and model-driven analysis of high resolution data seems crucial to\nreveal this complexity.\n",
"title": "The Gaia-ESO Survey: low-alpha element stars in the Galactic Bulge"
}
| null | null | null | null | true | null |
14777
| null |
Default
| null | null |
null |
{
"abstract": " In 2016 we proved that for every symmetric, repetition invariant and Jensen\nconcave mean $\\mathscr{M}$ the Kedlaya-type inequality $$\n\\mathscr{A}\\big(x_1,\\mathscr{M}(x_1,x_2),\\ldots,\\mathscr{M}(x_1,\\ldots,x_n)\\big)\\le\n\\mathscr{M} \\big(x_1,\n\\mathscr{A}(x_1,x_2),\\ldots,\\mathscr{A}(x_1,\\ldots,x_n)\\big) $$ holds for an\narbitrary $(x_n)$ ($\\mathscr{A}$ stands for the arithmetic mean). We are going\nto prove the weighted counterpart of this inequality. More precisely, if\n$(x_n)$ is a vector with corresponding (non-normalized) weights $(\\lambda_n)$\nand $\\mathscr{M}_{i=1}^n(x_i,\\lambda_i)$ denotes the weighted mean then, under\nanalogous conditions on $\\mathscr{M}$, the inequality $$ \\mathscr{A}_{i=1}^n\n\\big(\\mathscr{M}_{j=1}^i (x_j,\\lambda_j),\\:\\lambda_i\\big) \\le\n\\mathscr{M}_{i=1}^n \\big(\\mathscr{A}_{j=1}^i (x_j,\\lambda_j),\\:\\lambda_i\\big)\n$$ holds for every $(x_n)$ and $(\\lambda_n)$ such that the sequence\n$(\\frac{\\lambda_k}{\\lambda_1+\\cdots+\\lambda_k})$ is decreasing.\n",
"title": "On Kedlaya type inequalities for weighted means"
}
| null | null | null | null | true | null |
14778
| null |
Default
| null | null |
null |
{
"abstract": " We present three new semi-Lagrangian methods based on radial basis function\n(RBF) interpolation for numerically simulating transport on a sphere. The\nmethods are mesh-free and are formulated entirely in Cartesian coordinates,\nthus avoiding any irregular clustering of nodes at artificial boundaries on the\nsphere and naturally bypassing any apparent artificial singularities associated\nwith surface-based coordinate systems. For problems involving tracer transport\nin a given velocity field, the semi-Lagrangian framework allows these new\nmethods to avoid the use of any stabilization terms (such as hyperviscosity)\nduring time-integration, thus reducing the number of parameters that have to be\ntuned. The three new methods are based on interpolation using 1) global RBFs,\n2) local RBF stencils, and 3) RBF partition of unity. For the latter two of\nthese methods, we find that it is crucial to include some low degree spherical\nharmonics in the interpolants. Standard test cases consisting of solid body\nrotation and deformational flow are used to compare and contrast the methods in\nterms of their accuracy, efficiency, conservation properties, and\ndissipation/dispersion errors. For global RBFs, spectral spatial convergence is\nobserved for smooth solutions on quasi-uniform nodes, while high-order accuracy\nis observed for the local RBF stencil and partition of unity approaches.\n",
"title": "Mesh-free Semi-Lagrangian Methods for Transport on a Sphere Using Radial Basis Functions"
}
| null | null |
[
"Computer Science"
] | null | true | null |
14779
| null |
Validated
| null | null |
null |
{
"abstract": " The Integrated Nested Laplace Approximation (INLA) is a convenient way to\nobtain approximations to the posterior marginals for parameters in Bayesian\nhierarchical models when the latent effects can be expressed as a Gaussian\nMarkov Random Field (GMRF). In addition, its implementation in the R-INLA\npackage for the R statistical software provides an easy way to fit models using\nINLA in practice. R-INLA implements a number of widely used latent models,\nincluding several spatial models. In addition, R-INLA can fit models in a\nfraction of the time than other computer intensive methods (e.g. Markov Chain\nMonte Carlo) take to fit the same model.\nAlthough INLA provides a fast approximation to the marginals of the model\nparameters, it is difficult to use it with models not implemented in R-INLA. It\nis also difficult to make multivariate posterior inference on the parameters of\nthe model as INLA focuses on the posterior marginals and not the joint\nposterior distribution.\nIn this paper we describe how to use INLA within the Metropolis-Hastings\nalgorithm to fit spatial models and estimate the joint posterior distribution\nof a reduced number of parameters. We will illustrate the benefits of this new\nmethod with two examples on spatial econometrics and disease mapping where\ncomplex spatial models with several spatial structures need to be fitted.\n",
"title": "Spatial Models with the Integrated Nested Laplace Approximation within Markov Chain Monte Carlo"
}
| null | null | null | null | true | null |
14780
| null |
Default
| null | null |
null |
{
"abstract": " Supervisory signals can help topic models discover low-dimensional data\nrepresentations that are more interpretable for clinical tasks. We propose a\nframework for training supervised latent Dirichlet allocation that balances two\ngoals: faithful generative explanations of high-dimensional data and accurate\nprediction of associated class labels. Existing approaches fail to balance\nthese goals by not properly handling a fundamental asymmetry: the intended task\nis always predicting labels from data, not data from labels. Our new\nprediction-constrained objective trains models that predict labels from heldout\ndata well while also producing good generative likelihoods and interpretable\ntopic-word parameters. In a case study on predicting depression medications\nfrom electronic health records, we demonstrate improved recommendations\ncompared to previous supervised topic models and high- dimensional logistic\nregression from words alone.\n",
"title": "Prediction-Constrained Topic Models for Antidepressant Recommendation"
}
| null | null | null | null | true | null |
14781
| null |
Default
| null | null |
null |
{
"abstract": " We study the spatial fluctuations of the Casimir-Polder force experienced by\nan atom or a small sphere moved above a metallic plate at fixed separation\ndistance. We demonstrate that unlike the mean force, the magnitude of these\nfluctuations crucially relies on the relaxation of conduction electron in the\nmetallic bulk, and even achieves values that differ by orders of magnitude\ndepending on the amount of dissipation. We also discover that fluctuations\nsuffer a spectacular decrease at large distances in the case of nonzero\ntemperature.\n",
"title": "Casimir-Polder force fluctuations as spatial probes of dissipation in metals"
}
| null | null | null | null | true | null |
14782
| null |
Default
| null | null |
null |
{
"abstract": " Non-linear kernel methods can be approximated by fast linear ones using\nsuitable explicit feature maps allowing their application to large scale\nproblems. To this end, explicit feature maps of kernels for vectorial data have\nbeen extensively studied. As many real-world data is structured, various\nkernels for complex data like graphs have been proposed. Indeed, many of them\ndirectly compute feature maps. However, the kernel trick is employed when the\nnumber of features is very large or the individual vertices of graphs are\nannotated by real-valued attributes.\nCan we still compute explicit feature maps efficiently under these\ncircumstances? Triggered by this question, we investigate how general\nconvolution kernels are composed from base kernels and construct corresponding\nfeature maps. We apply our results to widely used graph kernels and analyze for\nwhich kernels and graph properties computation by explicit feature maps is\nfeasible and actually more efficient. In particular, we derive feature maps for\nrandom walk and subgraph matching kernels and apply them to real-world graphs\nwith discrete labels. Thereby, our theoretical results are confirmed\nexperimentally by observing a phase transition when comparing running time with\nrespect to label diversity, walk lengths and subgraph size, respectively.\nMoreover, we derive approximative, explicit feature maps for state-of-the-art\nkernels supporting real-valued attributes including the GraphHopper and Graph\nInvariant kernels. In extensive experiments we show that our approaches often\nachieve a classification accuracy close to the exact methods based on the\nkernel trick, but require only a fraction of their running time.\n",
"title": "A Unifying View of Explicit and Implicit Feature Maps for Structured Data: Systematic Studies of Graph Kernels"
}
| null | null | null | null | true | null |
14783
| null |
Default
| null | null |
null |
{
"abstract": " Though suicide is a major public health problem in the US, machine learning\nmethods are not commonly used to predict an individual's risk of\nattempting/committing suicide. In the present work, starting with an anonymized\ncollection of electronic health records for 522,056 unique, California-resident\nadolescents, we develop neural network models to predict suicide attempts. We\nframe the problem as a binary classification problem in which we use a\npatient's data from 2006-2009 to predict either the presence (1) or absence (0)\nof a suicide attempt in 2010. After addressing issues such as severely\nimbalanced classes and the variable length of a patient's history, we build\nneural networks with depths varying from two to eight hidden layers. For test\nset observations where we have at least five ED/hospital visits' worth of data\non a patient, our depth-4 model achieves a sensitivity of 0.703, specificity of\n0.980, and AUC of 0.958.\n",
"title": "Predicting Adolescent Suicide Attempts with Neural Networks"
}
| null | null | null | null | true | null |
14784
| null |
Default
| null | null |
null |
{
"abstract": " An RNA sequence is a word over an alphabet on four elements $\\{A,C,G,U\\}$\ncalled bases. RNA sequences fold into secondary structures where some bases\nmatch one another while others remain unpaired. Pseudoknot-free secondary\nstructures can be represented as well-parenthesized expressions with additional\ndots, where pairs of matching parentheses symbolize paired bases and dots,\nunpaired bases. The two fundamental problems in RNA algorithmic are to predict\nhow sequences fold within some model of energy and to design sequences of bases\nwhich will fold into targeted secondary structures. Predicting how a given RNA\nsequence folds into a pseudoknot-free secondary structure is known to be\nsolvable in cubic time since the eighties and in truly subcubic time by a\nrecent result of Bringmann et al. (FOCS 2016). As a stark contrast, it is\nunknown whether or not designing a given RNA secondary structure is a tractable\ntask; this has been raised as a challenging open question by Anne Condon (ICALP\n2003). Because of its crucial importance in a number of fields such as\npharmaceutical research and biochemistry, there are dozens of heuristics and\nsoftware libraries dedicated to RNA secondary structure design. It is therefore\nrather surprising that the computational complexity of this central problem in\nbioinformatics has been unsettled for decades.\nIn this paper we show that, in the simplest model of energy which is the\nWatson-Crick model the design of secondary structures is NP-complete if one\nadds natural constraints of the form: index $i$ of the sequence has to be\nlabeled by base $b$. This negative result suggests that the same lower bound\nholds for more realistic models of energy. It is noteworthy that the additional\nconstraints are by no means artificial: they are provided by all the RNA design\npieces of software and they do correspond to the actual practice.\n",
"title": "Designing RNA Secondary Structures is Hard"
}
| null | null | null | null | true | null |
14785
| null |
Default
| null | null |
null |
{
"abstract": " Motivated by recent findings that human mobility is proxy for crime behavior\nin big cities and that there is a superlinear relationship between the people's\nmovement and crime, this article aims to evaluate the impact of how these\nfindings influence police allocation. More precisely, we shed light on the\ndifferences between an allocation strategy, in which the resources are\ndistributed by clusters of floating population, and conventional allocation\nstrategies, in which the police resources are distributed by an Administrative\nArea (typically based on resident population). We observed a substantial\ndifference in the distributions of police resources allocated following these\nstrategies, what evidences the imprecision of conventional police allocation\nmethods.\n",
"title": "Towards Understanding the Impact of Human Mobility on Police Allocation"
}
| null | null | null | null | true | null |
14786
| null |
Default
| null | null |
null |
{
"abstract": " We present a sufficient condition for irreducibility of forcing algebras and\nstudy the (non)-reducedness phenomenon. Furthermore, we prove a criterion for\nnormality for forcing algebras over a polynomial base ring with coefficients in\na perfect field. This gives a geometrical normality criterion for algebraic\n(forcing) varieties over algebraically closed fields. Besides, we examine in\ndetail an specific (enlightening) example with several forcing equations.\nFinally, we compute explicitly the normalization of a particular forcing\nalgebra by means of finding explicitly the generators of the ideal defining it\nas an affine ring.\n",
"title": "Normality and Related Properties of Forcing Algebras"
}
| null | null | null | null | true | null |
14787
| null |
Default
| null | null |
null |
{
"abstract": " Triplet networks are widely used models that are characterized by good\nperformance in classification and retrieval tasks. In this work we propose to\ntrain a triplet network by putting it as the discriminator in Generative\nAdversarial Nets (GANs). We make use of the good capability of representation\nlearning of the discriminator to increase the predictive quality of the model.\nWe evaluated our approach on Cifar10 and MNIST datasets and observed\nsignificant improvement on the classification performance using the simple k-nn\nmethod.\n",
"title": "Training Triplet Networks with GAN"
}
| null | null | null | null | true | null |
14788
| null |
Default
| null | null |
null |
{
"abstract": " Recovering pairwise interactions, i.e. pairs of input features whose joint\neffect on an output is different from the sum of their marginal effects, is\ncentral in many scientific applications. We conceptualize a solution to this\nproblem as a two-stage procedure: first, we model the relationship between the\nfeatures and the output using a flexible hybrid neural network; second, we\ndetect feature interactions from the trained model. For the second step we\npropose a simple and intuitive interaction measure (IM), which has no specific\nrequirements on the machine learning model used in the first step, only that it\ndefines a mapping from an input to an output. And in a special case it reduces\nto the averaged Hessian of the input-output mapping. Importantly, our method\nupper bounds the interaction recovery error with the error of the learning\nmodel, which ensures that we can improve the recovered interactions by training\na more accurate model. We present analyses of simulated and real-world data\nwhich demonstrate the benefits of our method compared to available\nalternatives, and theoretically analyse its properties and relation to other\nmethods.\n",
"title": "Recovering Pairwise Interactions Using Neural Networks"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
14789
| null |
Validated
| null | null |
null |
{
"abstract": " The boundary algebraic Bethe Ansatz for a supersymmetric nineteen\nvertex-model constructed from a three-dimensional representation of the twisted\nquantum affine Lie superalgebra $U_{q}[\\mathrm{osp}(2|2)^{(2)}]$ is presented.\nThe eigenvalues and eigenvectors of Sklyanin's transfer matrix, with diagonal\nreflection $K$-matrices, are calculated and the corresponding Bethe Ansatz\nequations are obtained.\n",
"title": "Boundary Algebraic Bethe Ansatz for a nineteen vertex model with $U_{q}[\\mathrm{osp}(2|2)^{(2)}] symmetry$"
}
| null | null | null | null | true | null |
14790
| null |
Default
| null | null |
null |
{
"abstract": " Strengthening or destroying a network is a very important issue in designing\nresilient networks or in planning attacks against networks including planning\nstrategies to immunize a network against diseases, viruses etc.. Here we\ndevelop a method for strengthening or destroying a random network with a\nminimum cost. We assume a correlation between the cost required to strengthen\nor destroy a node and the degree of the node. Accordingly, we define a cost\nfunction c(k), which is the cost of strengthening or destroying a node with\ndegree k. Using the degrees $k$ in a network and the cost function c(k), we\ndevelop a method for defining a list of priorities of degrees, and for choosing\nthe right group of degrees to be strengthened or destroyed that minimizes the\ntotal price of strengthening or destroying the entire network. We find that the\nlist of priorities of degrees is universal and independent of the network's\ndegree distribution, for all kinds of random networks. The list of priorities\nis the same for both strengthening a network and for destroying a network with\nminimum cost. However, in spite of this similarity there is a difference\nbetween their p_c - the critical fraction of nodes that has to be functional,\nto guarantee the existence of a giant component in the network.\n",
"title": "Optimal cost for strengthening or destroying a given network"
}
| null | null | null | null | true | null |
14791
| null |
Default
| null | null |
null |
{
"abstract": " Reparameterization of variational auto-encoders with continuous latent spaces\nis an effective method for reducing the variance of their gradient estimates.\nHowever, using the same approach when latent variables are discrete is\nproblematic, due to the resulting non-differentiable objective. In this work,\nwe present a direct optimization method that propagates gradients through a\nnon-differentiable $\\arg \\max$ prediction operation. We apply this method to\ndiscrete variational auto-encoders, by modeling a discrete random variable by\nthe $\\arg \\max$ function of the Gumbel-Max perturbation model.\n",
"title": "Direct Optimization through $\\arg \\max$ for Discrete Variational Auto-Encoder"
}
| null | null |
[
"Statistics"
] | null | true | null |
14792
| null |
Validated
| null | null |
null |
{
"abstract": " The case of the classical Hill problem is numerically investigated by\nperforming a thorough and systematic classification of the initial conditions\nof the orbits. More precisely, the initial conditions of the orbits are\nclassified into four categories: (i) non-escaping regular orbits; (ii) trapped\nchaotic orbits; (iii) escaping orbits; and (iv) collision orbits. In order to\nobtain a more general and complete view of the orbital structure of the\ndynamical system our exploration takes place in both planar (2D) and the\nspatial (3D) version of the Hill problem. For the 2D system we numerically\nintegrate large sets of initial conditions in several types of planes, while\nfor the system with three degrees of freedom, three-dimensional distributions\nof initial conditions of orbits are examined. For distinguishing between\nordered and chaotic bounded motion the Smaller ALingment Index (SALI) method is\nused. We managed to locate the several bounded basins, as well as the basins of\nescape and collision and also to relate them with the corresponding escape and\ncollision time of the orbits. Our numerical calculations indicate that the\noverall orbital dynamics of the Hamiltonian system is a complicated but highly\ninterested problem. We hope our contribution to be useful for a further\nunderstanding of the orbital properties of the classical Hill problem.\n",
"title": "Orbit classification in the Hill problem: I. The classical case"
}
| null | null | null | null | true | null |
14793
| null |
Default
| null | null |
null |
{
"abstract": " This paper addresses a task allocation problem for a large-scale robotic\nswarm, namely swarm distribution guidance problem. Unlike most of the existing\nframeworks handling this problem, the proposed framework suggests utilising\nlocal information available to generate its time-varying stochastic policies.\nAs each agent requires only local consistency on information with neighbouring\nagents, rather than the global consistency, the proposed framework offers\nvarious advantages, e.g., a shorter timescale for using new information and\npotential to incorporate an asynchronous decision-making process. We perform\ntheoretical analysis on the properties of the proposed framework. From the\nanalysis, it is proved that the framework can guarantee the convergence to the\ndesired density distribution even using local information while maintaining\nadvantages of global-information-based approaches. The design requirements for\nthese advantages are explicitly listed in this paper. This paper also provides\nspecific examples of how to implement the framework developed. The results of\nnumerical experiments confirm the effectiveness and comparability of the\nproposed framework, compared with the global-information-based framework.\n",
"title": "Bio-Inspired Local Information-Based Control for Probabilistic Swarm Distribution Guidance"
}
| null | null | null | null | true | null |
14794
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we construct the simultaneous confidence band (SCB) for the\nnonparametric component in partially linear panel data models with fixed\neffects. We remove the fixed effects, and further obtain the estimators of\nparametric and nonparametric components, which do not depend on the fixed\neffects. We establish the asymptotic distribution of their maximum absolute\ndeviation between the estimated nonparametric component and the true\nnonparametric component under some suitable conditions, and hence the result\ncan be used to construct the simultaneous confidence band of the nonparametric\ncomponent. Based on the asymptotic distribution, it becomes difficult for the\nconstruction of the simultaneous confidence band. The reason is that the\nasymptotic distribution involves the estimators of the asymptotic bias and\nconditional variance, and the choice of the bandwidth for estimating the second\nderivative of nonparametric function. Clearly, these will cause computational\nburden and accumulative errors. To overcome these problems, we propose a\nBootstrap method to construct simultaneous confidence band. Simulation studies\nindicate that the proposed Bootstrap method exhibits better performance under\nthe limited samples.\n",
"title": "Simultaneous Confidence Band for Partially Linear Panel Data Models with Fixed Effects"
}
| null | null |
[
"Statistics"
] | null | true | null |
14795
| null |
Validated
| null | null |
null |
{
"abstract": " Nano--metal/semiconductor junction dependent porosification of silicon (Si)\nhas been studied here. The silicon (Si) nanostructures (NS) have been textured\non n-- and p-- type silicon wafers using Ag and Au metal nano particles induced\nchemical etching. The combinations of n--Si/Ag and p--Si/Au form ohmic contact\nand result in the same texturization on the Si surface on porosification where\ntent--shaped morphology has been observed consistently with n-- and p--type Si.\nWhereas, porosification result in different surface texturization for other two\ncombinations (p--Si/Ag and n--Si/Au) where Schottkey contacts are formed.\nQuantitative analysis have been done using ImageJ to process the SEM images of\nSiNS, which confirms that the tent like SiNS are formed when etching of silicon\nwafer is done by AgNPs and AuNPs on n and p type Si wafer respectively. These\neasily prepared sharp tent--shaped Si NSs can be used for enhanced field\nemission applications.\n",
"title": "Tent--Shaped Surface Morphologies of Silicon: Texturization by Metal Induced Etching"
}
| null | null | null | null | true | null |
14796
| null |
Default
| null | null |
null |
{
"abstract": " The paper approaches the problem of image-to-text with attention-based\nencoder-decoder networks that are trained to handle sequences of characters\nrather than words. We experiment on lines of text from a popular handwriting\ndatabase with different attention mechanisms for the decoder. The model trained\nwith softmax attention achieves the lowest test error, outperforming several\nother RNN-based models. Our results show that softmax attention is able to\nlearn a linear alignment whereas the alignment generated by sigmoid attention\nis linear but much less precise.\n",
"title": "Attention networks for image-to-text"
}
| null | null | null | null | true | null |
14797
| null |
Default
| null | null |
null |
{
"abstract": " Humanoid soccer robots perceive their environment exclusively through\ncameras. This paper presents a monocular vision system that was originally\ndeveloped for use in the RoboCup Humanoid League, but is expected to be\ntransferable to other soccer leagues. Recent changes in the Humanoid League\nrules resulted in a soccer environment with less color coding than in previous\nyears, which makes perception of the game situation more challenging. The\nproposed vision system addresses these challenges by using brightness and\ntexture for the detection of the required field features and objects. Our\nsystem is robust to changes in lighting conditions, and is designed for\nreal-time use on a humanoid soccer robot. This paper describes the main\ncomponents of the detection algorithms in use, and presents experimental\nresults from the soccer field, using ROS and the igus Humanoid Open Platform as\na testbed. The proposed vision system was used successfully at RoboCup 2015.\n",
"title": "A Monocular Vision System for Playing Soccer in Low Color Information Environments"
}
| null | null | null | null | true | null |
14798
| null |
Default
| null | null |
null |
{
"abstract": " This paper considers the problem of inferring image labels from images when\nonly a few annotated examples are available at training time. This setup is\noften referred to as low-shot learning, where a standard approach is to\nre-train the last few layers of a convolutional neural network learned on\nseparate classes for which training examples are abundant. We consider a\nsemi-supervised setting based on a large collection of images to support label\npropagation. This is possible by leveraging the recent advances on large-scale\nsimilarity graph construction.\nWe show that despite its conceptual simplicity, scaling label propagation up\nto hundred millions of images leads to state of the art accuracy in the\nlow-shot learning regime.\n",
"title": "Low-shot learning with large-scale diffusion"
}
| null | null | null | null | true | null |
14799
| null |
Default
| null | null |
null |
{
"abstract": " The Fermilab Muon Campus will host the Muon g-2 experiment - a world class\nexperiment dedicated to the search for signals of new physics. Strict demands\nare placed on beam diagnostics in order to ensure delivery of high quality\nbeams to the storage ring with minimal losses. In this study, we briefly\ndescribe the available secondary beam diagnostics for the Fermilab Muon Campus.\nThen, with the aid of numerical simulations we detail their interaction with\nthe secondary beam. Finally, we compare our results against theoretical\nfindings.\n",
"title": "Instrumentation and its Interaction with the Secondary Beam for the Fermilab Muon Campus"
}
| null | null |
[
"Physics"
] | null | true | null |
14800
| null |
Validated
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.