text
null | inputs
dict | prediction
null | prediction_agent
null | annotation
list | annotation_agent
null | multi_label
bool 1
class | explanation
null | id
stringlengths 1
5
| metadata
null | status
stringclasses 2
values | event_timestamp
null | metrics
null |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
null |
{
"abstract": " In this paper, we discuss stochastic comparisons of parallel systems with\nindependent heterogeneous exponentiated Nadarajah-Haghighi (ENH) components in\nterms of the usual stochastic order, dispersive order, convex transform order\nand the likelihood ratio order. In the presence of the Archimedean copula, we\nstudy stochastic comparison of series dependent systems in terms of the usual\nstochastic order.\n",
"title": "Stochastic comparisons of series and parallel systems with heterogeneous components"
}
| null | null | null | null | true | null |
12201
| null |
Default
| null | null |
null |
{
"abstract": " Mendelian randomization (MR) is a method of exploiting genetic variation to\nunbiasedly estimate a causal effect in presence of unmeasured confounding. MR\nis being widely used in epidemiology and other related areas of population\nscience. In this paper, we study statistical inference in the increasingly\npopular two-sample summary-data MR design. We show a linear model for the\nobserved associations approximately holds in a wide variety of settings when\nall the genetic variants satisfy the exclusion restriction assumption, or in\ngenetic terms, when there is no pleiotropy. In this scenario, we derive a\nmaximum profile likelihood estimator with provable consistency and asymptotic\nnormality. However, through analyzing real datasets, we find strong evidence of\nboth systematic and idiosyncratic pleiotropy in MR, echoing the omnigenic model\nof complex traits that is recently proposed in genetics. We model the\nsystematic pleiotropy by a random effects model, where no genetic variant\nsatisfies the exclusion restriction condition exactly. In this case we propose\na consistent and asymptotically normal estimator by adjusting the profile\nscore. We then tackle the idiosyncratic pleiotropy by robustifying the adjusted\nprofile score. We demonstrate the robustness and efficiency of the proposed\nmethods using several simulated and real datasets.\n",
"title": "Statistical inference in two-sample summary-data Mendelian randomization using robust adjusted profile score"
}
| null | null | null | null | true | null |
12202
| null |
Default
| null | null |
null |
{
"abstract": " We study improved approximations to the distribution of the largest\neigenvalue $\\hat{\\ell}$ of the sample covariance matrix of $n$ zero-mean\nGaussian observations in dimension $p+1$. We assume that one population\nprincipal component has variance $\\ell > 1$ and the remaining `noise'\ncomponents have common variance $1$. In the high dimensional limit $p/n \\to\n\\gamma > 0$, we begin study of Edgeworth corrections to the limiting Gaussian\ndistribution of $\\hat{\\ell}$ in the supercritical case $\\ell > 1 + \\sqrt\n\\gamma$. The skewness correction involves a quadratic polynomial as in\nclassical settings, but the coefficients reflect the high dimensional\nstructure. The methods involve Edgeworth expansions for sums of independent\nnon-identically distributed variates obtained by conditioning on the sample\nnoise eigenvalues, and limiting bulk properties \\textit{and} fluctuations of\nthese noise eigenvalues.\n",
"title": "Edgeworth correction for the largest eigenvalue in a spiked PCA model"
}
| null | null |
[
"Mathematics",
"Statistics"
] | null | true | null |
12203
| null |
Validated
| null | null |
null |
{
"abstract": " This paper is devoted to the investigation of the following function $$ f:\nx=\\Delta^{3}_{\\alpha_{1}\\alpha_{2}...\\alpha_{n}...}{\\rightarrow}\n\\Delta^{3}_{\\varphi(\\alpha_{1})\\varphi(\\alpha_{2})...\\varphi(\\alpha_{n})...}=f(x)=y,\n$$ where $\\varphi(i)=\\frac{-3i^{2}+7i}{2}$, $ i \\in N^{0}_{2}=\\{0,1,2\\}$, and\n$\\Delta^{3}_{\\alpha_{1}\\alpha_{2}...\\alpha_{n}...}$ is the ternary\nrepresentation of $x \\in [0;1]$. That is values of this function are obtained\nfrom the ternary representation of the argument by the following change of\ndigits: 0 by 0, 1 by 2, and 2 by 1. This function preserves the ternary digit\n$0$.\nMain mapping properties and differential, integral, fractal properties of the\nfunction are studied. Equivalent representations by additionally defined\nauxiliary functions of this function are proved.\nThis paper is the paper translated from Ukrainian (the Ukrainian variant\navailable at this https URL). In 2012, the\nUkrainian variant of this paper was represented by the author in the\nInternational Scientific Conference \"Asymptotic Methods in the Theory of\nDifferential Equations\" dedicated to 80th anniversary of M. I. Shkil (the\nconference paper available at\nthis https URL). In 2013, the\ninvestigations of the present article were generalized by the author in the\npaper \"One one class of functions with complicated local structure\"\n(this https URL) and in the several conference papers\n(available at: this https URL,\nthis https URL).\n",
"title": "On one nearly everywhere continuous and nowhere differentiable function, that defined by automaton with finite memory"
}
| null | null | null | null | true | null |
12204
| null |
Default
| null | null |
null |
{
"abstract": " We describe the dimensions of low Hochschild cohomology spaces of exceptional\nperiodic representation-infinite algebras of polynomial growth. As an\napplication we obtain that an indecomposable non-standard periodic\nrepresentation-infinite algebra of polynomial growth is not derived equivalent\nto a standard self-injective algebra.\n",
"title": "Hochschild cohomology for periodic algebras of polynomial growth"
}
| null | null | null | null | true | null |
12205
| null |
Default
| null | null |
null |
{
"abstract": " Fracton models, a collection of exotic gapped lattice Hamiltonians recently\ndiscovered in three spatial dimensions, contain some 'topological' features:\nthey support fractional bulk excitations (dubbed fractons), and a ground state\ndegeneracy that is robust to local perturbations. However, because previous\nfracton models have only been defined and analyzed on a cubic lattice with\nperiodic boundary conditions, it is unclear to what extent a notion of topology\nis applicable. In this paper, we demonstrate that the X-cube model, a\nprototypical type-I fracton model, can be defined on general three-dimensional\nmanifolds. Our construction revolves around the notion of a singular compact\ntotal foliation of the spatial manifold, which constructs a lattice from\nintersecting stacks of parallel surfaces called leaves. We find that the ground\nstate degeneracy depends on the topology of the leaves and the pattern of leaf\nintersections. We further show that such a dependence can be understood from a\nrenormalization group transformation for the X-cube model, wherein the system\nsize can be changed by adding or removing 2D layers of topological states. Our\nresults lead to an improved definition of fracton phase and bring to the fore\nthe topological nature of fracton orders.\n",
"title": "Fracton Models on General Three-Dimensional Manifolds"
}
| null | null | null | null | true | null |
12206
| null |
Default
| null | null |
null |
{
"abstract": " Topic modeling enables exploration and compact representation of a corpus.\nThe CaringBridge (CB) dataset is a massive collection of journals written by\npatients and caregivers during a health crisis. Topic modeling on the CB\ndataset, however, is challenging due to the asynchronous nature of multiple\nauthors writing about their health journeys. To overcome this challenge we\nintroduce the Dynamic Author-Persona topic model (DAP), a probabilistic\ngraphical model designed for temporal corpora with multiple authors. The\nnovelty of the DAP model lies in its representation of authors by a persona ---\nwhere personas capture the propensity to write about certain topics over time.\nFurther, we present a regularized variational inference algorithm, which we use\nto encourage the DAP model's personas to be distinct. Our results show\nsignificant improvements over competing topic models --- particularly after\nregularization, and highlight the DAP model's unique ability to capture common\njourneys shared by different authors.\n",
"title": "Topic Modeling on Health Journals with Regularized Variational Inference"
}
| null | null | null | null | true | null |
12207
| null |
Default
| null | null |
null |
{
"abstract": " Let $\\mathcal{V}_p(\\lambda)$ be the collection of all functions $f$ defined\nin the unit disc $\\ID$ having a simple pole at $z=p$ where $0<p<1$ and analytic\nin $\\ID\\setminus\\{p\\}$ with $f(0)=0=f'(0)-1$ and satisfying the differential\ninequality $|(z/f(z))^2 f'(z)-1|< \\lambda $ for $z\\in \\ID$, $0<\\lambda\\leq 1$.\nEach $f\\in\\mathcal{V}_p(\\lambda)$ has the following Taylor expansion:\n$$\nf(z)=z+\\sum_{n=2}^{\\infty}a_n(f) z^n, \\quad |z|<p.\n$$\nIn \\cite{BF-3}, we conjectured that\n$$\n|a_n(f)|\\leq \\frac{1-(\\lambda p^2)^n}{p^{n-1}(1-\\lambda p^2)}\\quad\n\\mbox{for}\\quad n\\geq3. $$ In the present article, we first obtain a\nrepresentation formula for functions in the class $\\mathcal{V}_p(\\lambda)$.\nUsing this representation, we prove the aforementioned conjecture for $n=3,4,5$\nwhenever $p$ belongs to certain subintervals of $(0,1)$. Also we determine non\nsharp bounds for $|a_n(f)|,\\,n\\geq 3$ and for $|a_{n+1}(f)-a_n(f)/p|,\\,n\\geq\n2$.\n",
"title": "On the Taylor coefficients of a subclass of meromorphic univalent functions"
}
| null | null | null | null | true | null |
12208
| null |
Default
| null | null |
null |
{
"abstract": " The loss functions of deep neural networks are complex and their geometric\nproperties are not well understood. We show that the optima of these complex\nloss functions are in fact connected by simple curves over which training and\ntest accuracy are nearly constant. We introduce a training procedure to\ndiscover these high-accuracy pathways between modes. Inspired by this new\ngeometric insight, we also propose a new ensembling method entitled Fast\nGeometric Ensembling (FGE). Using FGE we can train high-performing ensembles in\nthe time required to train a single model. We achieve improved performance\ncompared to the recent state-of-the-art Snapshot Ensembles, on CIFAR-10,\nCIFAR-100, and ImageNet.\n",
"title": "Loss Surfaces, Mode Connectivity, and Fast Ensembling of DNNs"
}
| null | null | null | null | true | null |
12209
| null |
Default
| null | null |
null |
{
"abstract": " Urban areas with larger and more connected populations offer an auspicious\nenvironment for contagion processes such as the spread of pathogens. Empirical\nevidence reveals a systematic increase in the rates of certain sexually\ntransmitted diseases (STDs) with larger urban population size. However, the\nmain drivers of these systemic infection patterns are still not well\nunderstood, and rampant urbanization rates worldwide makes it critical to\nadvance our understanding on this front. Using confirmed-cases data for three\nSTDs in US metropolitan areas, we investigate the scaling patterns of\ninfectious disease incidence in urban areas. The most salient features of these\npatterns are that, on average, the incidence of infectious diseases that\ntransmit with less ease-- either because of a lower inherent transmissibility\nor due to a less suitable environment for transmission-- scale more steeply\nwith population size, are less predictable across time and more variable across\ncities of similar size. These features are explained, first, using a simple\nmathematical model of contagion, and then through the lens of a new theory of\nurban scaling. These theoretical frameworks help us reveal the links between\nthe factors that determine the transmissibility of infectious diseases and the\nproperties of their scaling patterns across cities.\n",
"title": "On the scaling patterns of infectious disease incidence in cities"
}
| null | null | null | null | true | null |
12210
| null |
Default
| null | null |
null |
{
"abstract": " We investigate homological subsets of the prime spectrum of a ring, defined\nby the help of the Ext-family $\\{\\Ext^i_R(-,R)\\}$. We extend Grothendieck's\ncalculation of $\\dim(\\Ext^g_R(M,R))$. We compute support of $\\Ext^i_R(M,R)$ in\nmany cases. Also, we answer a low-dimensional case of a problem posed by\nVasconcelos on the finiteness of associated prime ideals of\n$\\{\\Ext^i_R(M,R)\\}$. An application is given.\n",
"title": "Homological subsets of Spec"
}
| null | null | null | null | true | null |
12211
| null |
Default
| null | null |
null |
{
"abstract": " The phenomenon of self-synchronization in populations of oscillatory units\nappears naturally in neurosciences. However, in some situations, the formation\nof a coherent state is damaging. In this article we study a repulsive\nmean-field Kuramoto model that describes the time evolution of n points on the\nunit circle, which are transformed into incoherent phase-locked states. It has\nbeen recently shown that such systems can be reduced to a three-dimensional\nsystem of ordinary differential equations, whose mathematical structure is\nstrongly related to hyperbolic geometry. The orbits of the Kuramoto dynamical\nsystem are then described by a ow of Möbius transformations. We show this\nunderlying dynamic performs statistical inference by computing dynamically\nM-estimates of scatter matrices. We also describe the limiting phase-locked\nstates for random initial conditions using Tyler's transformation matrix.\nMoreover, we show the repulsive Kuramoto model performs dynamically not only\nrobust covariance matrix estimation, but also data processing: the initial\nconfiguration of the n points is transformed by the dynamic into a limiting\nphase-locked state that surprisingly equals the spatial signs from\nnonparametric statistics. That makes the sign empirical covariance matrix to\nequal 1 2 id2, the variance-covariance matrix of a random vector that is\nuniformly distributed on the unit circle.\n",
"title": "Mean field repulsive Kuramoto models: Phase locking and spatial signs"
}
| null | null | null | null | true | null |
12212
| null |
Default
| null | null |
null |
{
"abstract": " We study causal waveform estimation (tracking) of time-varying signals in a\nparadigmatic atomic sensor, an alkali vapor monitored by Faraday rotation\nprobing. We use Kalman filtering, which optimally tracks known linear Gaussian\nstochastic processes, to estimate stochastic input signals that we generate by\noptical pumping. Comparing the known input to the estimates, we confirm the\naccuracy of the atomic statistical model and the reliability of the Kalman\nfilter, allowing recovery of waveform details far briefer than the sensor's\nintrinsic time resolution. With proper filter choice, we obtain similar\nbenefits when tracking partially-known and non-Gaussian signal processes, as\nare found in most practical sensing applications. The method evades the\ntrade-off between sensitivity and time resolution in coherent sensing.\n",
"title": "Signal tracking beyond the time resolution of an atomic sensor by Kalman filtering"
}
| null | null | null | null | true | null |
12213
| null |
Default
| null | null |
null |
{
"abstract": " In this report, we applied integrated gradients to explaining a neural\nnetwork for diabetic retinopathy detection. The integrated gradient is an\nattribution method which measures the contributions of input to the quantity of\ninterest. We explored some new ways for applying this method such as explaining\nintermediate layers, filtering out unimportant units by their attribution value\nand generating contrary samples. Moreover, the visualization results extend the\nuse of diabetic retinopathy detection model from merely predicting to assisting\nfinding potential lesions.\n",
"title": "Case Study: Explaining Diabetic Retinopathy Detection Deep CNNs via Integrated Gradients"
}
| null | null |
[
"Computer Science"
] | null | true | null |
12214
| null |
Validated
| null | null |
null |
{
"abstract": " The fast detection of terahertz radiation is of great importance for various\napplications such as fast imaging, high speed communications, and spectroscopy.\nMost commercial products capable of sensitively responding the terahertz\nradiation are thermal detectors, i.e., pyroelectric sensors and bolometers.\nThis class of terahertz detectors is normally characterized by low modulation\nfrequency (dozens or hundreds of Hz). Here we demonstrate the first fast\nsemiconductor-based terahertz quantum well photodetectors by carefully\ndesigning the device structure and microwave transmission line for high\nfrequency signal extraction. Modulation response bandwidth of gigahertz level\nis obtained. As an example, the 6.2-GHz modulated terahertz light emitted from\na Fabry-Pérot terahertz quantum cascade laser is successfully detected\nusing the fast terahertz quantum well photodetector. In addition to the fast\nterahertz detection, the technique presented in this work can also facilitate\nthe frequency stability or phase noise characterizations for terahertz quantum\ncascade lasers.\n",
"title": "6.2-GHz modulated terahertz light detection using fast terahertz quantum well photodetectors"
}
| null | null | null | null | true | null |
12215
| null |
Default
| null | null |
null |
{
"abstract": " We investigate the low-dimensional structure of deterministic transformations\nbetween random variables, i.e., transport maps between probability measures. In\nthe context of statistics and machine learning, these transformations can be\nused to couple a tractable \"reference\" measure (e.g., a standard Gaussian) with\na target measure of interest. Direct simulation from the desired measure can\nthen be achieved by pushing forward reference samples through the map. Yet\ncharacterizing such a map---e.g., representing and evaluating it---grows\nchallenging in high dimensions. The central contribution of this paper is to\nestablish a link between the Markov properties of the target measure and the\nexistence of low-dimensional couplings, induced by transport maps that are\nsparse and/or decomposable. Our analysis not only facilitates the construction\nof transformations in high-dimensional settings, but also suggests new\ninference methodologies for continuous non-Gaussian graphical models. For\ninstance, in the context of nonlinear state-space models, we describe new\nvariational algorithms for filtering, smoothing, and sequential parameter\ninference. These algorithms can be understood as the natural\ngeneralization---to the non-Gaussian case---of the square-root\nRauch-Tung-Striebel Gaussian smoother.\n",
"title": "Inference via low-dimensional couplings"
}
| null | null | null | null | true | null |
12216
| null |
Default
| null | null |
null |
{
"abstract": " We show a unified second-order scheme for constructing simple, robust and\naccurate algorithms for typical thermostats for configurational sampling for\nthe canonical ensemble. When Langevin dynamics is used, the scheme leads to the\nBAOAB algorithm that has been recently investigated. We show that the scheme is\nalso useful for other types of thermostat, such as the Andersen thermostat and\nNosé-Hoover chain. Two 1-dimensional models and three typical realistic\nmolecular systems that range from the gas phase, clusters, to the condensed\nphase are used in numerical examples for demonstration. Accuracy may be\nincreased by an order of magnitude for estimating coordinate-dependent\nproperties in molecular dynamics (when the same time interval is used),\nirrespective of which type of thermostat is applied. The scheme is especially\nuseful for path integral molecular dynamics, because it consistently improves\nthe efficiency for evaluating all thermodynamic properties for any type of\nthermostat.\n",
"title": "A unified thermostat scheme for efficient configurational sampling for classical/quantum canonical ensembles via molecular dynamics"
}
| null | null |
[
"Physics"
] | null | true | null |
12217
| null |
Validated
| null | null |
null |
{
"abstract": " It is shown via theory and simulation that the resonant frequency of a Free\nElectron Laser may be modulated to obtain an FEL interaction with a frequency\nbandwidth which is at least an order of magnitude greater than normal FEL\noperation. The system is described in the linear regime by a summation over\nexponential gain modes, allowing the amplification of multiple light\nfrequencies simultaneously. Simulation in 3D demonstrates the process for\nparameters of the UK's CLARA FEL test facility currently under construction.\nThis new mode of FEL operation has close analogies to Frequency Modulation in a\nconventional cavity laser. This new, wide bandwidth mode of FEL operation\nscales well for X-ray generation and offers users a new form of high-power FEL\noutput.\n",
"title": "Wide Bandwidth, Frequency Modulated Free Electron Laser"
}
| null | null |
[
"Physics"
] | null | true | null |
12218
| null |
Validated
| null | null |
null |
{
"abstract": " This paper presents a novel transformation-proximal bundle algorithm to solve\nmultistage adaptive robust mixed-integer linear programs (MARMILPs). By\nexplicitly partitioning recourse decisions into state decisions and local\ndecisions, the proposed algorithm applies affine decision rule only to state\ndecisions and allows local decisions to be fully adaptive. In this way, the\nMARMILP is proved to be transformed into an equivalent two-stage adaptive\nrobust optimization (ARO) problem. The proposed multi-to-two transformation\nscheme remains valid for other types of non-anticipative decision rules besides\nthe affine one, and it is general enough to be employed with existing two-stage\nARO algorithms for solving MARMILPs. The proximal bundle method is developed\nfor the resulting two-stage ARO problem. We perform a theoretical analysis to\nshow finite convergence of the proposed algorithm with any positive tolerance.\nTo quantitatively assess solution quality, we develop a scenario-tree-based\nlower bounding technique. Computational studies on multiperiod inventory\nmanagement and process network planning are presented to demonstrate its\neffectiveness and computational scalability. In the inventory management\napplication, the affine decision rule method suffers from a severe\nsuboptimality with an average gap of 34.88%, while the proposed algorithm\ngenerates near-optimal solutions with an average gap of merely 1.68%.\n",
"title": "A Transformation-Proximal Bundle Algorithm for Solving Large-Scale Multistage Adaptive Robust Optimization Problems"
}
| null | null | null | null | true | null |
12219
| null |
Default
| null | null |
null |
{
"abstract": " We present a dual subspace ascent algorithm for support vector machine\ntraining that respects a budget constraint limiting the number of support\nvectors. Budget methods are effective for reducing the training time of kernel\nSVM while retaining high accuracy. To date, budget training is available only\nfor primal (SGD-based) solvers. Dual subspace ascent methods like sequential\nminimal optimization are attractive for their good adaptation to the problem\nstructure, their fast convergence rate, and their practical speed. By\nincorporating a budget constraint into a dual algorithm, our method enjoys the\nbest of both worlds. We demonstrate considerable speed-ups over primal budget\ntraining methods.\n",
"title": "Dual SVM Training on a Budget"
}
| null | null |
[
"Statistics"
] | null | true | null |
12220
| null |
Validated
| null | null |
null |
{
"abstract": " AWAKE is a proton-driven plasma wakefield acceleration experiment. % We show\nthat the experimental setup briefly described here is ready for systematic\nstudy of the seeded self-modulation of the 400\\,GeV proton bunch in the\n10\\,m-long rubidium plasma with density adjustable from 1 to\n10$\\times10^{14}$\\,cm$^{-3}$. % We show that the short laser pulse used for\nionization of the rubidium vapor propagates all the way along the column,\nsuggesting full ionization of the vapor. % We show that ionization occurs along\nthe proton bunch, at the laser time and that the plasma that follows affects\nthe proton bunch. %\n",
"title": "AWAKE readiness for the study of the seeded self-modulation of a 400\\,GeV proton bunch"
}
| null | null | null | null | true | null |
12221
| null |
Default
| null | null |
null |
{
"abstract": " In lieu of an abstract here is the first paragraph: No other species remotely\napproaches the human capacity for the cultural evolution of novelty that is\naccumulative, adaptive, and open-ended (i.e., with no a priori limit on the\nsize or scope of possibilities). By culture we mean extrasomatic\nadaptations--including behavior and technology--that are socially rather than\nsexually transmitted. This chapter synthesizes research from anthropology,\npsychology, archaeology, and agent-based modeling into a speculative yet\ncoherent account of two fundamental cognitive transitions underlying human\ncultural evolution that is consistent with contemporary psychology. While the\nchapter overlaps with a more technical paper on this topic (Gabora & Smith\n2018), it incorporates new research and elaborates a genetic component to our\noverall argument. The ideas in this chapter grew out of a non-Darwinian\nframework for cultural evolution, referred to as the Self-other Reorganization\n(SOR) theory of cultural evolution (Gabora, 2013, in press; Smith, 2013), which\nwas inspired by research on the origin and earliest stage in the evolution of\nlife (Cornish-Bowden & Cárdenas 2017; Goldenfeld, Biancalani, & Jafarpour,\n2017, Vetsigian, Woese, & Goldenfeld 2006; Woese, 2002). SOR bridges\npsychological research on fundamental aspects of our human nature such as\ncreativity and our proclivity to reflect on ideas from different perspectives,\nwith the literature on evolutionary approaches to cultural evolution that\naspire to synthesize the behavioral sciences much as has been done for the\nbiological scientists. The current chapter is complementary to this effort, but\nless abstract; it attempts to ground the theory of cultural evolution in terms\nof cognitive transitions as suggested by archaeological evidence.\n",
"title": "Exploring the Psychological Basis for Transitions in the Archaeological Record"
}
| null | null | null | null | true | null |
12222
| null |
Default
| null | null |
null |
{
"abstract": " The purpose of this article is to study the role of Gödel's functional\ninterpretation in the extraction of programs from proofs in well quasi-order\ntheory. The main focus is on the interpretation of Nash-Williams' famous\nminimal bad sequence construction, and the exploration of a number of much\nbroader problems which are related to this, particularly the question of the\nconstructive meaning of Zorn's lemma and the notion of recursion over the\nnon-wellfounded lexicographic ordering on infinite sequences.\n",
"title": "Well quasi-orders and the functional interpretation"
}
| null | null |
[
"Computer Science",
"Mathematics"
] | null | true | null |
12223
| null |
Validated
| null | null |
null |
{
"abstract": " As proved by Régnier and Rösler, the number of key comparisons required\nby the randomized sorting algorithm QuickSort to sort a list of $n$ distinct\nitems (keys) satisfies a global distributional limit theorem. Fill and Janson\nproved results about the limiting distribution and the rate of convergence, and\nused these to prove a result part way towards a corresponding local limit\ntheorem. In this paper we use a multi-round smoothing technique to prove the\nfull local limit theorem.\n",
"title": "A local limit theorem for Quicksort key comparisons via multi-round smoothing"
}
| null | null | null | null | true | null |
12224
| null |
Default
| null | null |
null |
{
"abstract": " In this work, we present a method to compute the Kantorovich distance, that\nis, the Wasserstein distance of order one, between a pair of two-dimensional\nhistograms. Recent works in Computer Vision and Machine Learning have shown the\nbenefits of measuring Wasserstein distances of order one between histograms\nwith $N$ bins, by solving a classical transportation problem on (very large)\ncomplete bipartite graphs with $N$ nodes and $N^2$ edges. The main contribution\nof our work is to approximate the original transportation problem by an\nuncapacitated min cost flow problem on a reduced flow network of size $O(N)$.\nMore precisely, when the distance among the bin centers is measured with the\n1-norm or the $\\infty$-norm, our approach provides an optimal solution. When\nthe distance amongst bins is measured with the 2-norm: (i) we derive a\nquantitative estimate on the error between optimal and approximate solution;\n(ii) given the error, we construct a reduced flow network of size $O(N)$. We\nnumerically show the benefits of our approach by computing Wasserstein\ndistances of order one on a set of grey scale images used as benchmarks in the\nliterature. We show how our approach scales with the size of the images with\n1-norm, 2-norm and $\\infty$-norm ground distances.\n",
"title": "On the Computation of Kantorovich-Wasserstein Distances between 2D-Histograms by Uncapacitated Minimum Cost Flows"
}
| null | null | null | null | true | null |
12225
| null |
Default
| null | null |
null |
{
"abstract": " Elastic dissipation through radiation towards the substrate is a major loss\nchannel in micro- and nanomechanical resonators. Engineering the coupling of\nthese resonators with optical cavities further complicates and constrains the\ndesign of low-loss optomechanical devices. In this work we rely on the coherent\ncancellation of mechanical radiation to demonstrate material and surface\nabsorption limited silicon near-field optomechanical resonators oscillating at\ntens of MHz. The effectiveness of our dissipation suppression scheme is\ninvestigated at room and cryogenic temperatures. While at room temperature we\ncan reach a maximum quality factor of 7.61k ($fQ$-product of the order of\n$10^{11}$~Hz), at 22~K the quality factor increases to 37k, resulting in a\n$fQ$-product of $2\\times10^{12}$~Hz.\n",
"title": "Efficient anchor loss suppression in coupled near-field optomechanical resonators"
}
| null | null | null | null | true | null |
12226
| null |
Default
| null | null |
null |
{
"abstract": " In this proceedings application of a fuzzy Support Vector Machine (FSVM)\nlearning algorithm, to classify mid-infrared (MIR) sources from the AKARI NEP\nDeep field into three classes: stars, galaxies and AGNs, is presented. FSVM is\nan improved version of the classical SVM algorithm, incorporating measurement\nerrors into the classification process; this is the first successful\napplication of this algorithm in the astronomy. We created reliable catalogues\nof galaxies, stars and AGNs consisting of objects with MIR measurements, some\nof them with no optical counterparts. Some examples of identified objects are\nshown, among them O-rich and C-rich AGB stars.\n",
"title": "Searching for previously unknown classes of objects in the AKARI-NEP Deep data with fuzzy logic SVM algorithm"
}
| null | null | null | null | true | null |
12227
| null |
Default
| null | null |
null |
{
"abstract": " Cylindrical Couette flow is a subject where the main focus has long been on\nthe onset of turbulence or, more precisely, the limit of stability of the\nsimplest laminar flow. The theoretical framework of this paper is a recently\ndeveloped action principle for hydrodynamics. It incorporates Euler-Lagrange\nequations that are in essential agreement with the Navier-Stokes equation, but\napplicable to the general case of a compressible fluid. The variational\nprinciple incorporates the equation of continuity, a canonical structure and a\nconserved Hamiltonian. The density is compressible, characterized by a general\n(non-polar) equation of state, and homogeneous. The onset of instability is\noften accompanied by bubble formation. It is proposed that the limit of\nstability of laminar Couette flow may some times be related to cavitation. In\ncontrast to traditional stability theory we are not looking for mathematical\ninstabilities of a system of differential equations, but instead for the\npossibility that the system is driven to a metastable or unstable\nconfiguration. The application of this idea to cylindrical Couette flow\nreported here turns out to account rather well for the observations. The\nfailure of a famous criterion due to Rayleigh is well known. It is here shown\nthat it may be due to the use of methods that are appropriate only in the case\nthat the equations of motion are derived from an action principle.\n",
"title": "Stability of laminar Couette flow of compressible fluids"
}
| null | null | null | null | true | null |
12228
| null |
Default
| null | null |
null |
{
"abstract": " Understanding the spatial extent of extreme precipitation is necessary for\ndetermining flood risk and adequately designing infrastructure (e.g.,\nstormwater pipes) to withstand such hazards. While environmental phenomena\ntypically exhibit weakening spatial dependence at increasingly extreme levels,\nlimiting max-stable process models for block maxima have a rigid dependence\nstructure that does not capture this type of behavior. We propose a flexible\nBayesian model from a broader family of max-infinitely divisible processes that\nallows for weakening spatial dependence at increasingly extreme levels, and due\nto a hierarchical representation of the likelihood in terms of random effects,\nour inference approach scales to large datasets. The proposed model is\nconstructed using flexible random basis functions that are estimated from the\ndata, allowing for straightforward inspection of the predominant spatial\npatterns of extremes. In addition, the described process possesses\nmax-stability as a special case, making inference on the tail dependence class\npossible. We apply our model to extreme precipitation in eastern North America,\nand show that the proposed model adequately captures the extremal behavior of\nthe data.\n",
"title": "A Hierarchical Max-infinitely Divisible Process for Extreme Areal Precipitation Over Watersheds"
}
| null | null |
[
"Statistics"
] | null | true | null |
12229
| null |
Validated
| null | null |
null |
{
"abstract": " We present a study of the connection between brightest cluster galaxies\n(BCGs) and their host galaxy clusters. Using galaxy clusters at $0.1<z<0.3$\nfrom the Hectospec Cluster Survey (HeCS) with X-ray information from the\nArchive of {\\it Chandra} Cluster Entropy Profile Tables (ACCEPT), we confirm\nthat BCGs in low central entropy clusters are well aligned with the X-ray\ncenter. Additionally, the magnitude difference between BCG and the 2nd\nbrightest one also correlates with the central entropy of the intracluster\nmedium. From the red-sequence (RS) galaxies, we cannot find significant\ndependence of RS color scatter and stellar population on the central entropy of\nthe intracluster medium of their host cluster. However, BCGs in low entropy\nclusters are systematically less massive than those in high entropy clusters,\nalthough this is dependent on the method used to derive the stellar mass of\nBCGs. In contrast, the stellar velocity dispersion of BCGs shows no dependence\non BCG activity and cluster central entropy. This implies that the potential of\nthe BCG is established earlier and the activity leading to optical emission\nlines is dictated by the properties of the intracluster medium in the cluster\ncore.\n",
"title": "The dependence of cluster galaxy properties on the central entropy of their host cluster"
}
| null | null | null | null | true | null |
12230
| null |
Default
| null | null |
null |
{
"abstract": " We propose a novel architecture for $k$-shot classification on the Omniglot\ndataset. Building on prototypical networks, we extend their architecture to\nwhat we call Gaussian prototypical networks. Prototypical networks learn a map\nbetween images and embedding vectors, and use their clustering for\nclassification. In our model, a part of the encoder output is interpreted as a\nconfidence region estimate about the embedding point, and expressed as a\nGaussian covariance matrix. Our network then constructs a direction and class\ndependent distance metric on the embedding space, using uncertainties of\nindividual data points as weights. We show that Gaussian prototypical networks\nare a preferred architecture over vanilla prototypical networks with an\nequivalent number of parameters. We report state-of-the-art performance in\n1-shot and 5-shot classification both in 5-way and 20-way regime (for 5-shot\n5-way, we are comparable to previous state-of-the-art) on the Omniglot dataset.\nWe explore artificially down-sampling a fraction of images in the training set,\nwhich improves our performance even further. We therefore hypothesize that\nGaussian prototypical networks might perform better in less homogeneous,\nnoisier datasets, which are commonplace in real world applications.\n",
"title": "Gaussian Prototypical Networks for Few-Shot Learning on Omniglot"
}
| null | null | null | null | true | null |
12231
| null |
Default
| null | null |
null |
{
"abstract": " We study the pairs of projections $$ P_If=\\chi_If ,\\ \\ Q_Jf= \\left(\\chi_J\n\\hat{f}\\right)\\check{\\ } , \\ \\ f\\in L^2(\\mathbb{R}^n), $$ where $I, J\\subset\n\\mathbb{R}^n$ are sets of finite Lebesgue measure, $\\chi_I, \\chi_J$ denote the\ncorresponding characteristic functions and $\\hat{\\ } , \\check{\\ }$ denote the\nFourier-Plancherel transformation $L^2(\\mathbb{R}^n)\\to L^2(\\mathbb{R}^n)$ and\nits inverse. These pairs of projections have been widely studied by several\nauthors in connection with the mathematical formulation of Heisenberg's\nuncertainty principle. Our study is done from a differential geometric point of\nview. We apply known results on the Finsler geometry of the Grassmann manifold\n${\\cal P}({\\cal H})$ of a Hilbert space ${\\cal H}$ to establish that there\nexists a unique minimal geodesic of ${\\cal P}({\\cal H})$, which is a curve of\nthe form $$ \\delta(t)=e^{itX_{I,J}}P_Ie^{-itX_{I,J}} $$ which joins $P_I$ and\n$Q_J$ and has length $\\pi/2$. As a consequence we obtain that if $H$ is the\nlogarithm of the Fourier-Plancherel map, then $$ \\|[H,P_I]\\|\\ge \\pi/2. $$ The\nspectrum of $X_{I,J}$ is denumerable and symmetric with respect to the origin,\nit has a smallest positive eigenvalue $\\gamma(X_{I,J})$ which satisfies $$\n\\cos(\\gamma(X_{I,J}))=\\|P_IQ_J\\|. $$\n",
"title": "Uncertainty principle and geometry of the infinite Grassmann manifold"
}
| null | null | null | null | true | null |
12232
| null |
Default
| null | null |
null |
{
"abstract": " This paper is concerned with the following fractional Schrödinger\nequations involving critical exponents: \\begin{eqnarray*}\n(-\\Delta)^{\\alpha}u+V(x)u=k(x)f(u)+\\lambda|u|^{2_{\\alpha}^{*}-2}u\\quad\\quad\n\\mbox{in}\\ \\mathbb{R}^{N}, \\end{eqnarray*} where $(-\\Delta)^{\\alpha}$ is the\nfractional Laplacian operator with $\\alpha\\in(0,1)$, $N\\geq2$, $\\lambda$ is a\npositive real parameter and $2_{\\alpha}^{*}=2N/(N-2\\alpha)$ is the critical\nSobolev exponent, $V(x)$ and $k(x)$ are positive and bounded functions\nsatisfying some extra hypotheses. Based on the principle of concentration\ncompactness in the fractional Sobolev space and the minimax arguments, we\nobtain the existence of a nontrivial radially symmetric weak solution for the\nabove-mentioned equations without assuming the Ambrosetti-Rabinowitz condition\non the subcritical nonlinearity.\n",
"title": "Existence and symmetry of solutions for critical fractional Schrödinger equations with bounded potentials"
}
| null | null | null | null | true | null |
12233
| null |
Default
| null | null |
null |
{
"abstract": " Social media expose millions of users every day to information campaigns ---\nsome emerging organically from grassroots activity, others sustained by\nadvertising or other coordinated efforts. These campaigns contribute to the\nshaping of collective opinions. While most information campaigns are benign,\nsome may be deployed for nefarious purposes. It is therefore important to be\nable to detect whether a meme is being artificially promoted at the very moment\nit becomes wildly popular. This problem has important social implications and\nposes numerous technical challenges. As a first step, here we focus on\ndiscriminating between trending memes that are either organic or promoted by\nmeans of advertisement. The classification is not trivial: ads cause bursts of\nattention that can be easily mistaken for those of organic trends. We designed\na machine learning framework to classify memes that have been labeled as\ntrending on Twitter.After trending, we can rely on a large volume of activity\ndata. Early detection, occurring immediately at trending time, is a more\nchallenging problem due to the minimal volume of activity data that is\navailable prior to trending.Our supervised learning framework exploits hundreds\nof time-varying features to capture changing network and diffusion patterns,\ncontent and sentiment information, timing signals, and user meta-data. We\nexplore different methods for encoding feature time series. Using millions of\ntweets containing trending hashtags, we achieve 75% AUC score for early\ndetection, increasing to above 95% after trending. We evaluate the robustness\nof the algorithms by introducing random temporal shifts on the trend time\nseries. Feature selection analysis reveals that content cues provide\nconsistently useful signals; user features are more informative for early\ndetection, while network and timing features are more helpful once more data is\navailable.\n",
"title": "Early Detection of Promoted Campaigns on Social Media"
}
| null | null | null | null | true | null |
12234
| null |
Default
| null | null |
null |
{
"abstract": " Discrete statistical models supported on labelled event trees can be\nspecified using so-called interpolating polynomials which are generalizations\nof generating functions. These admit a nested representation. A new algorithm\nexploits the primary decomposition of monomial ideals associated with an\ninterpolating polynomial to quickly compute all nested representations of that\npolynomial. It hereby determines an important subclass of all trees\nrepresenting the same statistical model. To illustrate this method we analyze\nthe full polynomial equivalence class of a staged tree representing the best\nfitting model inferred from a real-world dataset.\n",
"title": "Discovery of statistical equivalence classes using computer algebra"
}
| null | null |
[
"Mathematics",
"Statistics"
] | null | true | null |
12235
| null |
Validated
| null | null |
null |
{
"abstract": " Boltzmann machines are physics informed generative models with wide\napplications in machine learning. They can learn the probability distribution\nfrom an input dataset and generate new samples accordingly. Applying them back\nto physics, the Boltzmann machines are ideal recommender systems to accelerate\nMonte Carlo simulation of physical systems due to their flexibility and\neffectiveness. More intriguingly, we show that the generative sampling of the\nBoltzmann Machines can even discover unknown cluster Monte Carlo algorithms.\nThe creative power comes from the latent representation of the Boltzmann\nmachines, which learn to mediate complex interactions and identify clusters of\nthe physical system. We demonstrate these findings with concrete examples of\nthe classical Ising model with and without four spin plaquette interactions.\nOur results endorse a fresh research paradigm where intelligent machines are\ndesigned to create or inspire human discovery of innovative algorithms.\n",
"title": "Can Boltzmann Machines Discover Cluster Updates ?"
}
| null | null | null | null | true | null |
12236
| null |
Default
| null | null |
null |
{
"abstract": " Graph Signal Processing (GSP) is a promising framework to analyze\nmulti-dimensional neuroimaging datasets, while taking into account both the\nspatial and functional dependencies between brain signals. In the present work,\nwe apply dimensionality reduction techniques based on graph representations of\nthe brain to decode brain activity from real and simulated fMRI datasets. We\nintroduce seven graphs obtained from a) geometric structure and/or b)\nfunctional connectivity between brain areas at rest, and compare them when\nperforming dimension reduction for classification. We show that mixed graphs\nusing both a) and b) offer the best performance. We also show that graph\nsampling methods perform better than classical dimension reduction including\nPrincipal Component Analysis (PCA) and Independent Component Analysis (ICA).\n",
"title": "Evaluating Graph Signal Processing for Neuroimaging Through Classification and Dimensionality Reduction"
}
| null | null | null | null | true | null |
12237
| null |
Default
| null | null |
null |
{
"abstract": " We compare various notions of weak subsolutions to degenerate complex\nMonge-Ampère equations, showing that they all coincide. This allows us to\ngive an alternative proof of mixed Monge-Ampère inequalities due to Kolodziej\nand Dinew.\n",
"title": "Weak subsolutions to complex Monge-Ampère equations"
}
| null | null |
[
"Mathematics"
] | null | true | null |
12238
| null |
Validated
| null | null |
null |
{
"abstract": " Efficient assessment of convolved hidden Markov models is discussed. The\nbottom-layer is defined as an unobservable categorical first-order Markov\nchain, while the middle-layer is assumed to be a Gaussian spatial variable\nconditional on the bottom-layer. Hence, this layer appear as a Gaussian mixture\nspatial variable unconditionally. We observe the top-layer as a convolution of\nthe middle-layer with Gaussian errors. Focus is on assessment of the\ncategorical and Gaussian mixture variables given the observations, and we\noperate in a Bayesian inversion framework. The model is defined to make\ninversion of subsurface seismic AVO data into lithology/fluid classes and to\nassess the associated elastic material properties. Due to the spatial coupling\nin the likelihood functions, evaluation of the posterior normalizing constant\nis computationally demanding, and brute-force, single-site updating Markov\nchain Monte Carlo algorithms converges far too slow to be useful. We construct\ntwo classes of approximate posterior models which we assess analytically and\nefficiently using the recursive Forward-Backward algorithm. These approximate\nposterior densities are used as proposal densities in an independent proposal\nMarkov chain Monte Carlo algorithm, to assess the correct posterior model. A\nset of synthetic realistic examples are presented. The proposed approximations\nprovides efficient proposal densities which results in acceptance probabilities\nin the range 0.10-0.50 in the Markov chain Monte Carlo algorithm. A case study\nof lithology/fluid seismic inversion is presented. The lithology/fluid classes\nand the elastic material properties can be reliably predicted.\n",
"title": "Bayesian inversion of convolved hidden Markov models with applications in reservoir prediction"
}
| null | null |
[
"Physics",
"Statistics"
] | null | true | null |
12239
| null |
Validated
| null | null |
null |
{
"abstract": " Detailed numerical analyses of the orbital motion of a test particle around a\nspinning primary are performed. They aim to investigate the possibility of\nusing the post-Keplerian (pK) corrections to the orbiter's periods (draconitic,\nanomalistic and sidereal) as a further opportunity to perform new tests of\npost-Newtonian (pN) gravity. As a specific scenario, the S-stars orbiting the\nMassive Black Hole (MBH) supposedly lurking in Sgr A$^\\ast$ at the center of\nthe Galaxy is adopted. We, first, study the effects of the pK Schwarzchild,\nLense-Thirring and quadrupole moment accelerations experienced by a target star\nfor various possible initial orbital configurations. It turns out that the\nresults of the numerical simulations are consistent with the analytical ones in\nthe small eccentricity approximation for which almost all the latter ones were\nderived. For highly elliptical orbits, the size of all the three pK corrections\nconsidered turn out to increase remarkably. The periods of the observed S2 and\nS0-102 stars as functions of the MBH's spin axis orientation are considered as\nwell. The pK accelerations considered lead to corrections of the orbital\nperiods of the order of 1-100d (Schwarzschild), 0.1-10h (Lense-Thirring) and\n1-10^3s (quadrupole) for a target star with a=300-800~AU and e ~ 0.8, which\ncould be possibly measurable by the future facilities.\n",
"title": "On the post-Keplerian corrections to the orbital periods of a two-body system and their application to the Galactic Center"
}
| null | null | null | null | true | null |
12240
| null |
Default
| null | null |
null |
{
"abstract": " Bottom-up and top-down, as well as low-level and high-level factors influence\nwhere we fixate when viewing natural scenes. However, the importance of each of\nthese factors and how they interact remains a matter of debate. Here, we\ndisentangle these factors by analysing their influence over time. For this\npurpose we develop a saliency model which is based on the internal\nrepresentation of a recent early spatial vision model to measure the low-level\nbottom-up factor. To measure the influence of high-level bottom-up features, we\nuse a recent DNN-based saliency model. To account for top-down influences, we\nevaluate the models on two large datasets with different tasks: first, a\nmemorisation task and, second, a search task. Our results lend support to a\nseparation of visual scene exploration into three phases: The first saccade, an\ninitial guided exploration characterised by a gradual broadening of the\nfixation density, and an steady state which is reached after roughly 10\nfixations. Saccade target selection during the initial exploration and in the\nsteady state are related to similar areas of interest, which are better\npredicted when including high-level features. In the search dataset, fixation\nlocations are determined predominantly by top-down processes. In contrast, the\nfirst fixation follows a different fixation density and contains a strong\ncentral fixation bias. Nonetheless, first fixations are guided strongly by\nimage properties and as early as 200 ms after image onset, fixations are better\npredicted by high-level information. We conclude that any low-level bottom-up\nfactors are mainly limited to the generation of the first saccade. All saccades\nare better explained when high-level features are considered, and later this\nhigh-level bottom-up control can be overruled by top-down influences.\n",
"title": "Disentangling top-down vs. bottom-up and low-level vs. high-level influences on eye movements over time"
}
| null | null | null | null | true | null |
12241
| null |
Default
| null | null |
null |
{
"abstract": " The sharp range of $L^p$-estimates for the class of Hörmander-type\noscillatory integral operators is established in all dimensions under a\npositive-definite assumption on the phase. This is achieved by generalising a\nrecent approach of the first author for studying the Fourier extension\noperator, which utilises polynomial partitioning arguments.\n",
"title": "Sharp estimates for oscillatory integral operators via polynomial partitioning"
}
| null | null | null | null | true | null |
12242
| null |
Default
| null | null |
null |
{
"abstract": " Through the Hasimoto map, various dynamical systems can be mapped to\ndifferent integrodifferential generalizations of Nonlinear Schrodinger (NLS)\nfamily of equations some of which are known to be integrable. Two such\ncontinuum limits, corresponding to the inhomogeneous XXX Heisenberg spin chain\n[Balakrishnan, J. Phys. C 15, L1305 (1982)] and that of a thin vortex filament\nmoving in a superfluid with drag [Shivamoggi, Eur. Phys. J. B 86, 275 (2013)\n86; Van Gorder, Phys. Rev. E 91, 053201 (2015)], are shown to be particular\nnon-holonomic deformations (NHDs) of the standard NLS system involving\ngeneralized parameterizations. Crucially, such NHDs of the NLS system are\nrestricted to specific spectral orders that exactly complements NHDs of the\noriginal physical systems. The specific non-holonomic constraints associated\nwith these integrodifferential generalizations additionally posses distinct\nsemi-classical signature.\n",
"title": "Inhomogeneous Heisenberg Spin Chain and Quantum Vortex Filament as Non-Holonomically Deformed NLS Systems"
}
| null | null | null | null | true | null |
12243
| null |
Default
| null | null |
null |
{
"abstract": " Deduplication finds and removes long-range data duplicates. It is commonly\nused in cloud and enterprise server settings and has been successfully applied\nto primary, backup, and archival storage. Despite its practical importance as a\nsource-coding technique, its analysis from the point of view of information\ntheory is missing. This paper provides such an information-theoretic analysis\nof data deduplication. It introduces a new source model adapted to the\ndeduplication setting. It formalizes the two standard fixed-length and\nvariable-length deduplication schemes, and it introduces a novel multi-chunk\ndeduplication scheme. It then provides an analysis of these three deduplication\nvariants, emphasizing the importance of boundary synchronization between source\nblocks and deduplication chunks. In particular, under fairly mild assumptions,\nthe proposed multi-chunk deduplication scheme is shown to be order optimal.\n",
"title": "An Information-Theoretic Analysis of Deduplication"
}
| null | null | null | null | true | null |
12244
| null |
Default
| null | null |
null |
{
"abstract": " While deep neural networks take loose inspiration from neuroscience, it is an\nopen question how seriously to take the analogies between artificial deep\nnetworks and biological neuronal systems. Interestingly, recent work has shown\nthat deep convolutional neural networks (CNNs) trained on large-scale image\nrecognition tasks can serve as strikingly good models for predicting the\nresponses of neurons in visual cortex to visual stimuli, suggesting that\nanalogies between artificial and biological neural networks may be more than\nsuperficial. However, while CNNs capture key properties of the average\nresponses of cortical neurons, they fail to explain other properties of these\nneurons. For one, CNNs typically require large quantities of labeled input data\nfor training. Our own brains, in contrast, rarely have access to this kind of\nsupervision, so to the extent that representations are similar between CNNs and\nbrains, this similarity must arise via different training paths. In addition,\nneurons in visual cortex produce complex time-varying responses even to static\ninputs, and they dynamically tune themselves to temporal regularities in the\nvisual environment. We argue that these differences are clues to fundamental\ndifferences between the computations performed in the brain and in deep\nnetworks. To begin to close the gap, here we study the emergent properties of a\npreviously-described recurrent generative network that is trained to predict\nfuture video frames in a self-supervised manner. Remarkably, the model is able\nto capture a wide variety of seemingly disparate phenomena observed in visual\ncortex, ranging from single unit response dynamics to complex perceptual motion\nillusions. These results suggest potentially deep connections between recurrent\npredictive neural network models and the brain, providing new leads that can\nenrich both fields.\n",
"title": "A neural network trained to predict future video frames mimics critical properties of biological neuronal responses and perception"
}
| null | null | null | null | true | null |
12245
| null |
Default
| null | null |
null |
{
"abstract": " Among the Milky Way satellites discovered in the past three years, Triangulum\nII has presented the most difficulty in revealing its dynamical status. Kirby\net al. (2015a) identified it as the most dark matter-dominated galaxy known,\nwith a mass-to-light ratio within the half-light radius of 3600 +3500 -2100\nM_sun/L_sun. On the other hand, Martin et al. (2016) measured an outer velocity\ndispersion that is 3.5 +/- 2.1 times larger than the central velocity\ndispersion, suggesting that the system might not be in equilibrium. From new\nmulti-epoch Keck/DEIMOS measurements of 13 member stars in Triangulum II, we\nconstrain the velocity dispersion to be sigma_v < 3.4 km/s (90% C.L.). Our\nprevious measurement of sigma_v, based on six stars, was inflated by the\npresence of a binary star with variable radial velocity. We find no evidence\nthat the velocity dispersion increases with radius. The stars display a wide\nrange of metallicities, indicating that Triangulum II retained supernova ejecta\nand therefore possesses or once possessed a massive dark matter halo. However,\nthe detection of a metallicity dispersion hinges on the membership of the two\nmost metal-rich stars. The stellar mass is lower than galaxies of similar mean\nstellar metallicity, which might indicate that Triangulum II is either a star\ncluster or a tidally stripped dwarf galaxy. Detailed abundances of one star\nshow heavily depressed neutron-capture abundances, similar to stars in most\nother ultra-faint dwarf galaxies but unlike stars in globular clusters.\n",
"title": "Triangulum II: Not Especially Dense After All"
}
| null | null | null | null | true | null |
12246
| null |
Default
| null | null |
null |
{
"abstract": " We improve the best known upper bound on the length of the shortest reset\nwords of synchronizing automata. The new bound is slightly better than $114 n^3\n/ 685 + O(n^2)$. The Černý conjecture states that $(n-1)^2$ is an upper\nbound. So far, the best general upper bound was $(n^3-n)/6-1$ obtained by\nJ.-E.~Pin and P.~Frankl in 1982. Despite a number of efforts, it remained\nunchanged for about 35 years.\nTo obtain the new upper bound we utilize avoiding words. A word is avoiding\nfor a state $q$ if after reading the word the automaton cannot be in $q$. We\nobtain upper bounds on the length of the shortest avoiding words, and using the\napproach of Trahtman from 2011 combined with the well known Frankl theorem from\n1982, we improve the general upper bound on the length of the shortest reset\nwords. For all the bounds, there exist polynomial algorithms finding a word of\nlength not exceeding the bound.\n",
"title": "Improving the upper bound on the length of the shortest reset words"
}
| null | null | null | null | true | null |
12247
| null |
Default
| null | null |
null |
{
"abstract": " Graphs are a fundamental abstraction for modeling relational data. However,\ngraphs are discrete and combinatorial in nature, and learning representations\nsuitable for machine learning tasks poses statistical and computational\nchallenges. In this work, we propose Graphite an algorithmic framework for\nunsupervised learning of representations over nodes in a graph using deep\nlatent variable generative models. Our model is based on variational\nautoencoders (VAE), and uses graph neural networks for parameterizing both the\ngenerative model (i.e., decoder) and inference model (i.e., encoder). The use\nof graph neural networks directly incorporates inductive biases due to the\nspatial, local structure of graphs directly in the generative model. We draw\nnovel connections of our framework with approximate inference via kernel\nembeddings. Empirically, Graphite outperforms competing approaches for the\ntasks of density estimation, link prediction, and node classification on\nsynthetic and benchmark datasets.\n",
"title": "Graphite: Iterative Generative Modeling of Graphs"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
12248
| null |
Validated
| null | null |
null |
{
"abstract": " Characteristic classes in space-time manifolds are discussed for both even-\nand odd-dimensional spacetimes. In particular, it is shown that the\nEinstein--Hilbert action is equivalent to a second Chern-class on a modified\nPoincare bundle in four dimensions. Consequently, the cosmological constant and\nthe trace of an energy-momentum tensor become divisible modulo R/Z.\n",
"title": "Characteristic classes in general relativity on a modified Poincare curvature bundle"
}
| null | null | null | null | true | null |
12249
| null |
Default
| null | null |
null |
{
"abstract": " We describe high resolution observations of a GOES B-class flare\ncharacterized by a circular ribbon at chromospheric level, corresponding to the\nnetwork at photospheric level. We interpret the flare as a consequence of a\nmagnetic reconnection event occurred at a three-dimensional (3D) coronal null\npoint located above the supergranular cell. The potential field extrapolation\nof the photospheric magnetic field indicates that the circular chromospheric\nribbon is cospatial with the fan footpoints, while the ribbons of the inner and\nouter spines look like compact kernels. We found new interesting observational\naspects that need to be explained by models: 1) a loop corresponding to the\nouter spine became brighter a few minutes before the onset of the flare; 2) the\ncircular ribbon was formed by several adjacent compact kernels characterized by\na size of 1\"-2\"; 3) the kernels with stronger intensity emission were located\nat the outer footpoint of the darker filaments departing radially from the\ncenter of the supergranular cell; 4) these kernels start to brighten\nsequentially in clockwise direction; 5) the site of the 3D null point and the\nshape of the outer spine were detected by RHESSI in the low energy channel\nbetween 6.0 and 12.0 keV. Taking into account all these features and the length\nscales of the magnetic systems involved by the event we argued that the low\nintensity of the flare may be ascribed to the low amount of magnetic flux and\nto its symmetric configuration.\n",
"title": "Observation of a 3D magnetic null point"
}
| null | null | null | null | true | null |
12250
| null |
Default
| null | null |
null |
{
"abstract": " Unraveling bacterial strategies for spatial exploration is crucial to\nunderstand the complexity of the organi- zation of life. Currently, a\ncornerstone for quantitative modeling of bacterial transport, is their\nrun-and-tumble strategy to explore their environment. For Escherichia coli, the\nrun time distribution was reported to follow a Poisson process with a single\ncharacteristic time related to the rotational switching of the flagellar motor.\nDirect measurements on flagellar motors show, on the contrary, heavy-tailed\ndistributions of rotation times stemming from the intrinsic noise in the\nchemotactic mechanism. The crucial role of stochasticity on the chemotactic\nresponse has also been highlighted by recent modeling, suggesting its\ndeterminant influence on motility. In stark contrast with the accepted vision\nof run-and-tumble, here we report a large behavioral variability of wild-type\nE. coli, revealed in their three-dimensional trajectories. At short times, a\nbroad distribution of run times is measured on a population and attributed to\nthe slow fluctuations of a signaling protein triggering the flagellar motor\nreversal. Over long times, individual bacteria undergo significant changes in\nmotility. We demonstrate that such a large distribution introduces measurement\nbiases in most practical situations. These results reconcile the notorious\nconundrum between run time observations and motor switching statistics. We\nfinally propose that statistical modeling of transport properties currently\nundertaken in the emerging framework of active matter studies should be\nreconsidered under the scope of this large variability of motility features.\n",
"title": "3D spatial exploration by E. coli echoes motor temporal variability"
}
| null | null | null | null | true | null |
12251
| null |
Default
| null | null |
null |
{
"abstract": " Behavioral annotation using signal processing and machine learning is highly\ndependent on training data and manual annotations of behavioral labels.\nPrevious studies have shown that speech information encodes significant\nbehavioral information and be used in a variety of automated behavior\nrecognition tasks. However, extracting behavior information from speech is\nstill a difficult task due to the sparseness of training data coupled with the\ncomplex, high-dimensionality of speech, and the complex and multiple\ninformation streams it encodes. In this work we exploit the slow varying\nproperties of human behavior. We hypothesize that nearby segments of speech\nshare the same behavioral context and hence share a similar underlying\nrepresentation in a latent space. Specifically, we propose a Deep Neural\nNetwork (DNN) model to connect behavioral context and derive the behavioral\nmanifold in an unsupervised manner. We evaluate the proposed manifold in the\ncouples therapy domain and also provide examples from publicly available data\n(e.g. stand-up comedy). We further investigate training within the couples'\ntherapy domain and from movie data. The results are extremely encouraging and\npromise improved behavioral quantification in an unsupervised manner and\nwarrants further investigation in a range of applications.\n",
"title": "Unsupervised Latent Behavior Manifold Learning from Acoustic Features: audio2behavior"
}
| null | null | null | null | true | null |
12252
| null |
Default
| null | null |
null |
{
"abstract": " Local properties of the fundamental group of a path-connected topological\nspace can pose obstructions to the applicability of covering space theory. A\ngeneralized covering map is a generalization of the classical notion of\ncovering map defined in terms of unique lifting properties. The existence of\ngeneralized covering maps depends entirely on the verification of the unique\npath lifting property for a standard covering construction. Given any\npath-connected metric space $X$, and a subgroup $H\\leq\\pi_1(X,x_0)$, we\ncharacterize the unique path lifting property relative to $H$ in terms of a new\nclosure operator on the $\\pi_1$-subgroup lattice that is induced by maps from a\nfixed \"test\" domain into $X$. Using this test map framework, we develop a\nunified approach to comparing the existence of generalized coverings with a\nnumber of related properties.\n",
"title": "Test map characterizations of local properties of fundamental groups"
}
| null | null | null | null | true | null |
12253
| null |
Default
| null | null |
null |
{
"abstract": " We present adaptive strategies for antenna selection for Direction of Arrival\n(DoA) estimation of a far-field source using TDM MIMO radar with linear arrays.\nOur treatment is formulated within a general adaptive sensing framework that\nuses one-step ahead predictions of the Bayesian MSE using a parametric family\nof Weiss-Weinstein bounds that depend on previous measurements. We compare in\nsimulations our strategy with adaptive policies that optimize the Bobrovsky-\nZaka{\\i} bound and the Expected Cramér-Rao bound, and show the performance\nfor different levels of measurement noise.\n",
"title": "Adaptive channel selection for DOA estimation in MIMO radar"
}
| null | null | null | null | true | null |
12254
| null |
Default
| null | null |
null |
{
"abstract": " We characterize the variation in photometric response of the Dark Energy\nCamera (DECam) across its 520~Mpix science array during 4 years of operation.\nThese variations are measured using high signal-to-noise aperture photometry of\n$>10^7$ stellar images in thousands of exposures of a few selected fields, with\nthe telescope dithered to move the sources around the array. A calibration\nprocedure based on these results brings the RMS variation in aperture\nmagnitudes of bright stars on cloudless nights down to 2--3 mmag, with <1 mmag\nof correlated photometric errors for stars separated by $\\ge20$\". On cloudless\nnights, any departures of the exposure zeropoints from a secant airmass law\nexceeding >1 mmag are plausibly attributable to spatial/temporal variations in\naperture corrections. These variations can be inferred and corrected by\nmeasuring the fraction of stellar light in an annulus between 6\" and 8\"\ndiameter. Key elements of this calibration include: correction of amplifier\nnonlinearities; distinguishing pixel-area variations and stray light from\nquantum-efficiency variations in the flat fields; field-dependent color\ncorrections; and the use of an aperture-correction proxy. The DECam response\npattern across the 2-degree field drifts over months by up to $\\pm7$ mmag, in a\nnearly-wavelength-independent low-order pattern. We find no fundamental\nbarriers to pushing global photometric calibrations toward mmag accuracy.\n",
"title": "Photometric characterization of the Dark Energy Camera"
}
| null | null | null | null | true | null |
12255
| null |
Default
| null | null |
null |
{
"abstract": " Self-healing polymers crosslinked by solely reversible bonds are\nintrinsically weaker than common covalently crosslinked networks. Introducing\ncovalent crosslinks into a reversible network would improve mechanical\nstrength. It is challenging, however, to apply this design concept to dry\nelastomers, largely because reversible crosslinks such as hydrogen bonds are\noften polar motifs, whereas covalent crosslinks are non-polar motifs, and these\ntwo types of bonds are intrinsically immiscible without co-solvents. Here we\ndesign and fabricate a hybrid polymer network by crosslinking randomly branched\npolymers carrying motifs that can form both reversible hydrogen bonds and\npermanent covalent crosslinks. The randomly branched polymer links such two\ntypes of bonds and forces them to mix on the molecular level without\nco-solvents. This allows us to create a hybrid dry elastomer that is very tough\nwith a fracture energy $13,500J/m^2$ comparable to that of natural rubber;\nmoreover, the elastomer can self-heal at room temperature with a recovered\ntensile strength 4 MPa similar to that of existing self-healing elastomers. The\nconcept of forcing covalent and reversible bonds to mix at molecular scale to\ncreate a homogenous network is quite general and should enable development of\ntough, self-healing polymers of practical usage.\n",
"title": "Tough self-healing elastomers by molecular enforced integration of covalent and reversible networks"
}
| null | null |
[
"Physics"
] | null | true | null |
12256
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper, we study an SYK model and an SYK-like tensor model with global\nsymmetry. First, we study the large $N$ expansion of the bi-local collective\naction for the SYK model with manifest global symmetry. We show that the global\nsymmetry is enhanced to a local symmetry at strong coupling limit, and the\ncorresponding symmetry algebra is the Kac-Moody algebra. The emergent local\nsymmetry together with the emergent reparametrization is spontaneously and\nexplicit broken. This leads to a low energy effective action. We evaluate four\npoint functions, and obtain spectrum of our model. We derive the low energy\neffective action and analyze the chaotic behavior of the four point functions.\nWe also consider the recent 3D gravity conjecture for our model.\nWe also introduce an SYK-like tensor model with global symmetry. We first\nstudy chaotic behavior of four point functions in various channels for the\nrank-3 case, and generalize this into a rank-$(q-1)$ tensor model.\n",
"title": "SYK Models and SYK-like Tensor Models with Global Symmetry"
}
| null | null | null | null | true | null |
12257
| null |
Default
| null | null |
null |
{
"abstract": " The common assumption that Theta-1-Ori C is the dominant ionizing source for\nthe Orion Nebula is critically examined. This assumption underlies much of the\nexisting analysis of the nebula. In this paper we establish through comparison\nof the relative strengths of emission lines with expectations from Cloudy\nmodels and through the direction of the bright edges of proplyds that\nTheta-2-Ori-A, which lies beyond the Bright Bar, also plays an important role.\nTheta-1-Ori-C does dominate ionization in the inner part of the Orion Nebula,\nbut outside of the Bright Bar as far as the southeast boundary of the Extended\nOrion Nebula, Theta-2-Ori-A is the dominant source. In addition to identifying\nthe ionizing star in sample regions, we were able to locate those portions of\nthe nebula in 3-D. This analysis illustrates the power of MUSE spectral imaging\nobservations in identifying sources of ionization in extended regions.\n",
"title": "Which Stars are Ionizing the Orion Nebula ?"
}
| null | null | null | null | true | null |
12258
| null |
Default
| null | null |
null |
{
"abstract": " The large majority of high energy sources detected with Fermi-LAT are\nblazars, which are known to be very variable sources. High cadence long-term\nmonitoring simultaneously at different wavelengths being prohibitive, the study\nof their transient activities can help shedding light on our understanding of\nthese objects. The early detection of such potentially fast transient events is\nthe key for triggering follow-up observations at other wavelengths. A Python\ntool, FLaapLUC, built on top of the Science Tools provided by the Fermi Science\nSupport Center and the Fermi-LAT collaboration, has been developed using a\nsimple aperture photometry approach. This tool can effectively detect relative\nflux variations in a set of predefined sources and alert potential users. Such\nalerts can then be used to trigger target of opportunity observations with\nother facilities. It is shown that FLaapLUC is an efficient tool to reveal\ntransient events in Fermi-LAT data, providing quick results which can be used\nto promptly organise follow-up observations. Results from this simple aperture\nphotometry method are also compared to full likelihood analyses. The FLaapLUC\npackage is made available on GitHub and is open to contributions by the\ncommunity.\n",
"title": "FLaapLUC: a pipeline for the generation of prompt alerts on transient Fermi-LAT $γ$-ray sources"
}
| null | null | null | null | true | null |
12259
| null |
Default
| null | null |
null |
{
"abstract": " One of the main computational and scientific challenges in the modern age is\nto extract useful information from unstructured texts. Topic models are one\npopular machine-learning approach which infers the latent topical structure of\na collection of documents. Despite their success --- in particular of its most\nwidely used variant called Latent Dirichlet Allocation (LDA) --- and numerous\napplications in sociology, history, and linguistics, topic models are known to\nsuffer from severe conceptual and practical problems, e.g. a lack of\njustification for the Bayesian priors, discrepancies with statistical\nproperties of real texts, and the inability to properly choose the number of\ntopics. Here we obtain a fresh view on the problem of identifying topical\nstructures by relating it to the problem of finding communities in complex\nnetworks. This is achieved by representing text corpora as bipartite networks\nof documents and words. By adapting existing community-detection methods --\nusing a stochastic block model (SBM) with non-parametric priors -- we obtain a\nmore versatile and principled framework for topic modeling (e.g., it\nautomatically detects the number of topics and hierarchically clusters both the\nwords and documents). The analysis of artificial and real corpora demonstrates\nthat our SBM approach leads to better topic models than LDA in terms of\nstatistical model selection. More importantly, our work shows how to formally\nrelate methods from community detection and topic modeling, opening the\npossibility of cross-fertilization between these two fields.\n",
"title": "A network approach to topic models"
}
| null | null | null | null | true | null |
12260
| null |
Default
| null | null |
null |
{
"abstract": " The propagation of charged cosmic rays through the Galactic environment\ninfluences all aspects of the observation at Earth. Energy spectrum,\ncomposition and arrival directions are changed due to deflections in magnetic\nfields and interactions with the interstellar medium. Today the transport is\nsimulated with different simulation methods either based on the solution of a\ntransport equation (multi-particle picture) or a solution of an equation of\nmotion (single-particle picture).\nWe developed a new module for the publicly available propagation software\nCRPropa 3.1, where we implemented an algorithm to solve the transport equation\nusing stochastic differential equations. This technique allows us to use a\ndiffusion tensor which is anisotropic with respect to an arbitrary magnetic\nbackground field. The source code of CRPropa is written in C++ with python\nsteering via SWIG which makes it easy to use and computationally fast.\nIn this paper, we present the new low-energy propagation code together with\nvalidation procedures that are developed to proof the accuracy of the new\nimplementation. Furthermore, we show first examples of the cosmic ray density\nevolution, which depends strongly on the ratio of the parallel\n$\\kappa_\\parallel$ and perpendicular $\\kappa_\\perp$ diffusion coefficients.\nThis dependency is systematically examined as well the influence of the\nparticle rigidity on the diffusion process.\n",
"title": "CRPropa 3.1 -- A low energy extension based on stochastic differential equations"
}
| null | null |
[
"Physics"
] | null | true | null |
12261
| null |
Validated
| null | null |
null |
{
"abstract": " The space of Kähler potentials in a compact Kähler manifold, endowed with\nMabuchi's metric, is an infinite dimensional Riemannian manifold. We\ncharacterize local isometries between spaces of Kähler potentials, and prove\nexistence and uniqueness for such isometries.\n",
"title": "Isometries in spaces of Kähler potentials"
}
| null | null |
[
"Mathematics"
] | null | true | null |
12262
| null |
Validated
| null | null |
null |
{
"abstract": " Echo state networks are powerful recurrent neural networks. However, they are\noften unstable and shaky, making the process of finding an good ESN for a\nspecific dataset quite hard. Obtaining a superb accuracy by using the Echo\nState Network is a challenging task. We create, develop and implement a family\nof predictably optimal robust and stable ensemble of Echo State Networks via\nregularizing the training and perturbing the input. Furthermore, several\ndistributions of weights have been tried based on the shape to see if the shape\nof the distribution has the impact for reducing the error. We found ESN can\ntrack in short term for most dataset, but it collapses in the long run.\nShort-term tracking with large size reservoir enables ESN to perform strikingly\nwith superior prediction. Based on this scenario, we go a further step to\naggregate many of ESNs into an ensemble to lower the variance and stabilize the\nsystem by stochastic replications and bootstrapping of input data.\n",
"title": "On the Statistical Challenges of Echo State Networks and Some Potential Remedies"
}
| null | null | null | null | true | null |
12263
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we prove that some Gaussian structural equation models with\ndependent errors having equal variances are identifiable from their\ncorresponding Gaussian distributions. Specifically, we prove identifiability\nfor the Gaussian structural equation models that can be represented as\nAndersson-Madigan-Perlman chain graphs (Andersson et al., 2001). These chain\ngraphs were originally developed to represent independence models. However,\nthey are also suitable for representing causal models with additive noise\n(Peña, 2016. Our result implies then that these causal models can be\nidentified from observational data alone. Our result generalizes the result by\nPeters and Bühlmann (2014), who considered independent errors having equal\nvariances. The suitability of the equal error variances assumption should be\nassessed on a per domain basis.\n",
"title": "Identifiability of Gaussian Structural Equation Models with Dependent Errors Having Equal Variances"
}
| null | null |
[
"Statistics"
] | null | true | null |
12264
| null |
Validated
| null | null |
null |
{
"abstract": " The disruptive power of blockchain technologies represents a great\nopportunity to re-imagine standard practices of telecommunication networks and\nto identify critical areas that can benefit from brand new approaches. As a\nstarting point for this debate, we look at the current limits of infrastructure\nsharing, and specifically at the Small-Cell-as-a-Service trend, asking\nourselves how we could push it to its natural extreme: a scenario in which any\nindividual home or business user can become a service provider for mobile\nnetwork operators, freed from all the scalability and legal constraints that\nare inherent to the current modus operandi. We propose the adoption of smart\ncontracts to implement simple but effective Service Level Agreements (SLAs)\nbetween small cell providers and mobile operators, and present an example\ncontract template based on the Ethereum blockchain.\n",
"title": "Smart Contract SLAs for Dense Small-Cell-as-a-Service"
}
| null | null | null | null | true | null |
12265
| null |
Default
| null | null |
null |
{
"abstract": " We investigate the limiting behavior of solutions of nonhomogeneous boundary\nvalue problems for the systems of linear ordinary differential equations. The\ngeneralization of Kiguradze theorem (1987) on passage to the limit is obtained.\n",
"title": "On Kiguradze theorem for linear boundary value problems"
}
| null | null | null | null | true | null |
12266
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we estimate the fidelity of stabilizer and CSS codes. First, we\nderive a lower bound on the fidelity of a stabilizer code via its quantum\nenumerator. Next, we find the average quantum enumerators of the ensembles of\nfinite length stabilizer and CSS codes. We use the average quantum enumerators\nfor obtaining lower bounds on the average fidelity of these ensembles. We\nfurther improve the fidelity bounds by estimating the quantum enumerators of\nexpurgated ensembles of stabilizer and CSS codes. Finally, we derive fidelity\nbounds in the asymptotic regime when the code length tends to infinity.\nThese results tell us which code rate we can afford for achieving a target\nfidelity with codes of a given length. The results also show that in symmetric\ndepolarizing channel a typical stabilizer code has better performance, in terms\nof fidelity and code rate, compared with a typical CSS codes, and that balanced\nCSS codes significantly outperform other CSS codes. Asymptotic results\ndemonstrate that CSS codes have a fundamental performance loss compared to\nstabilizer codes.\n",
"title": "Fidelity Lower Bounds for Stabilizer and CSS Quantum Codes"
}
| null | null | null | null | true | null |
12267
| null |
Default
| null | null |
null |
{
"abstract": " Cross-validation is widely used for selecting among a family of learning\nrules. This paper studies a related method, called aggregated hold-out\n(Agghoo), which mixes cross-validation with aggregation; Agghoo can also be\nrelated to bagging. According to numerical experiments, Agghoo can improve\nsignificantly cross-validation's prediction error, at the same computational\ncost; this makes it very promising as a general-purpose tool for prediction. We\nprovide the first theoretical guarantees on Agghoo, in the supervised\nclassification setting, ensuring that one can use it safely: at worse, Agghoo\nperforms like the hold-out, up to a constant factor. We also prove a\nnon-asymptotic oracle inequality, in binary classification under the margin\ncondition, which is sharp enough to get (fast) minimax rates.\n",
"title": "Cross-validation improved by aggregation: Agghoo"
}
| null | null |
[
"Mathematics",
"Statistics"
] | null | true | null |
12268
| null |
Validated
| null | null |
null |
{
"abstract": " In this work we formulate the problem of image captioning as a multimodal\ntranslation task. Analogous to machine translation, we present a\nsequence-to-sequence recurrent neural networks (RNN) model for image caption\ngeneration. Different from most existing work where the whole image is\nrepresented by convolutional neural network (CNN) feature, we propose to\nrepresent the input image as a sequence of detected objects which feeds as the\nsource sequence of the RNN model. In this way, the sequential representation of\nan image can be naturally translated to a sequence of words, as the target\nsequence of the RNN model. To represent the image in a sequential way, we\nextract the objects features in the image and arrange them in a order using\nconvolutional neural networks. To further leverage the visual information from\nthe encoded objects, a sequential attention layer is introduced to selectively\nattend to the objects that are related to generate corresponding words in the\nsentences. Extensive experiments are conducted to validate the proposed\napproach on popular benchmark dataset, i.e., MS COCO, and the proposed model\nsurpasses the state-of-the-art methods in all metrics following the dataset\nsplits of previous work. The proposed approach is also evaluated by the\nevaluation server of MS COCO captioning challenge, and achieves very\ncompetitive results, e.g., a CIDEr of 1.029 (c5) and 1.064 (c40).\n",
"title": "MAT: A Multimodal Attentive Translator for Image Captioning"
}
| null | null | null | null | true | null |
12269
| null |
Default
| null | null |
null |
{
"abstract": " Mobile gaming has emerged as a promising market with billion-dollar revenues.\nA variety of mobile game platforms and services have been developed around the\nworld. One critical challenge for these platforms and services is to understand\nuser churn behavior in mobile games. Accurate churn prediction will benefit\nmany stakeholders such as game developers, advertisers, and platform operators.\nIn this paper, we present the first large-scale churn prediction solution for\nmobile games. In view of the common limitations of the state-of-the-art methods\nbuilt upon traditional machine learning models, we devise a novel\nsemi-supervised and inductive embedding model that jointly learns the\nprediction function and the embedding function for user-app relationships. We\nmodel these two functions by deep neural networks with a unique edge embedding\ntechnique that is able to capture both contextual information and relationship\ndynamics. We also design a novel attributed random walk technique that takes\ninto consideration both topological adjacency and attribute similarities. To\nevaluate the performance of our solution, we collect real-world data from the\nSamsung Game Launcher platform that includes tens of thousands of games and\nhundreds of millions of user-app interactions. The experimental results with\nthis data demonstrate the superiority of our proposed model against existing\nstate-of-the-art methods.\n",
"title": "A Semi-Supervised and Inductive Embedding Model for Churn Prediction of Large-Scale Mobile Games"
}
| null | null | null | null | true | null |
12270
| null |
Default
| null | null |
null |
{
"abstract": " The seminal work of Morgan and Rubin (2012) considers rerandomization for all\nthe units at one time. In practice, however, experimenters may have to\nrerandomize units sequentially. For example, a clinician studying a rare\ndisease may be unable to wait to perform an experiment until all the\nexperimental units are recruited. Our work offers a mathematical framework for\nsequential rerandomization designs, where the experimental units are enrolled\nin groups. We formulate an adaptive rerandomization procedure for balancing\ntreatment/control assignments over some continuous or binary covariates, using\nMahalanobis distance as the imbalance measure. We prove in our key result,\nTheorem 3, that given the same number of rerandomizations (in expected value),\nunder certain mild assumptions, sequential rerandomization achieves better\ncovariate balance than rerandomization at one time.\n",
"title": "Sequential rerandomization"
}
| null | null | null | null | true | null |
12271
| null |
Default
| null | null |
null |
{
"abstract": " We present an approach for agents to learn representations of a global map\nfrom sensor data, to aid their exploration in new environments. To achieve\nthis, we embed procedures mimicking that of traditional Simultaneous\nLocalization and Mapping (SLAM) into the soft attention based addressing of\nexternal memory architectures, in which the external memory acts as an internal\nrepresentation of the environment. This structure encourages the evolution of\nSLAM-like behaviors inside a completely differentiable deep neural network. We\nshow that this approach can help reinforcement learning agents to successfully\nexplore new environments where long-term memory is essential. We validate our\napproach in both challenging grid-world environments and preliminary Gazebo\nexperiments. A video of our experiments can be found at: this https URL.\n",
"title": "Neural SLAM: Learning to Explore with External Memory"
}
| null | null | null | null | true | null |
12272
| null |
Default
| null | null |
null |
{
"abstract": " To store information at extremely high-density and data-rate, we propose to\nadapt, integrate, and extend the techniques developed by chemists and molecular\nbiologists for the purpose of manipulating biological and other macromolecules.\nIn principle, volumetric densities in excess of 10^21 bits/cm^3 can be achieved\nwhen individual molecules having dimensions below a nanometer or so are used to\nencode the 0's and 1's of a binary string of data. In practice, however, given\nthe limitations of electron-beam lithography, thin film deposition and\npatterning technologies, molecular manipulation in submicron dimensions, etc.,\nwe believe that volumetric storage densities on the order of 10^16 bits/cm^3\n(i.e., petabytes per cubic centimeter) should be readily attainable, leaving\nplenty of room for future growth. The unique feature of the proposed new\napproach is its focus on the feasibility of storing bits of information in\nindividual molecules, each only a few angstroms in size.\n",
"title": "Information Storage and Retrieval using Macromolecules as Storage Media"
}
| null | null | null | null | true | null |
12273
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, Morgan type uncertainty principle and unique continuation\nproperties of abstract Schrödinger equations with time dependent potentials\nin vector-valued classes are obtained. The equation involves a possible linear\noperators considered in the Hilbert spaces. So, by choosing the corresponding\nspaces H and operators we derived unique continuation properties for numerous\nclasses of Schrödinger type equations and its systems which occur in a wide\nvariety of physical systems\n",
"title": "Morgan type uncertainty principle and unique continuation properties for abstract Schrödinger equations"
}
| null | null |
[
"Mathematics"
] | null | true | null |
12274
| null |
Validated
| null | null |
null |
{
"abstract": " Early in researchers' careers, it is difficult to assess how good their work\nis or how important or influential the scholars will eventually be. Hence,\nfunding agencies, academic departments, and others often use the Journal Impact\nFactor (JIF) of where the authors have published to assess their work and\nprovide resources and rewards for future work. The use of JIFs in this way has\nbeen heavily criticized, however. Using a large data set with many thousands of\npublication profiles of individual researchers, this study tests the ability of\nthe JIF (in its normalized variant) to identify, at the beginning of their\ncareers, those candidates who will be successful in the long run. Instead of\nbare JIFs and citation counts, the metrics used here are standardized according\nto Web of Science subject categories and publication years. The results of the\nstudy indicate that the JIF (in its normalized variant) is able to discriminate\nbetween researchers who published papers later on with a citation impact above\nor below average in a field and publication year - not only in the short term,\nbut also in the long term. However, the low to medium effect sizes of the\nresults also indicate that the JIF (in its normalized variant) should not be\nused as the sole criterion for identifying later success: other criteria, such\nas the novelty and significance of the specific research, academic\ndistinctions, and the reputation of previous institutions, should also be\nconsidered.\n",
"title": "Can the Journal Impact Factor Be Used as a Criterion for the Selection of Junior Researchers? A Large-Scale Empirical Study Based on ResearcherID Data"
}
| null | null | null | null | true | null |
12275
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we consider the use of deep neural networks in the context of\nMultiple-Input-Multiple-Output (MIMO) detection. We give a brief introduction\nto deep learning and propose a modern neural network architecture suitable for\nthis detection task. First, we consider the case in which the MIMO channel is\nconstant, and we learn a detector for a specific system. Next, we consider the\nharder case in which the parameters are known yet changing and a single\ndetector must be learned for all multiple varying channels. We demonstrate the\nperformance of our deep MIMO detector using numerical simulations in comparison\nto competing methods including approximate message passing and semidefinite\nrelaxation. The results show that deep networks can achieve state of the art\naccuracy with significantly lower complexity while providing robustness against\nill conditioned channels and mis-specified noise variance.\n",
"title": "Deep MIMO Detection"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
12276
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper, we address the problem of cross-view image geo-localization.\nSpecifically, we aim to estimate the GPS location of a query street view image\nby finding the matching images in a reference database of geo-tagged bird's eye\nview images, or vice versa. To this end, we present a new framework for\ncross-view image geo-localization by taking advantage of the tremendous success\nof deep convolutional neural networks (CNNs) in image classification and object\ndetection. First, we employ the Faster R-CNN to detect buildings in the query\nand reference images. Next, for each building in the query image, we retrieve\nthe $k$ nearest neighbors from the reference buildings using a Siamese network\ntrained on both positive matching image pairs and negative pairs. To find the\ncorrect NN for each query building, we develop an efficient multiple nearest\nneighbors matching method based on dominant sets. We evaluate the proposed\nframework on a new dataset that consists of pairs of street view and bird's eye\nview images. Experimental results show that the proposed method achieves better\ngeo-localization accuracy than other approaches and is able to generalize to\nimages at unseen locations.\n",
"title": "Cross-View Image Matching for Geo-localization in Urban Environments"
}
| null | null | null | null | true | null |
12277
| null |
Default
| null | null |
null |
{
"abstract": " In recent years, deep learning based on artificial neural network (ANN) has\nachieved great success in pattern recognition. However, there is no clear\nunderstanding of such neural computational models. In this paper, we try to\nunravel \"black-box\" structure of Ann model from network flow. Specifically, we\nconsider the feed forward Ann as a network flow model, which consists of many\ndirectional class-pathways. Each class-pathway encodes one class. The\nclass-pathway of a class is obtained by connecting the activated neural nodes\nin each layer from input to output, where activation value of neural node\n(node-value) is defined by the weights of each layer in a trained\nANN-classifier. From the perspective of the class-pathway, training an\nANN-classifier can be regarded as the formulation process of class-pathways of\ndifferent classes. By analyzing the the distances of each two class-pathways in\na trained ANN-classifiers, we try to answer the questions, why the classifier\nperforms so? At last, from the neural encodes view, we define the importance of\neach neural node through the class-pathways, which is helpful to optimize the\nstructure of a classifier. Experiments for two types of ANN model including\nmulti-layer MLP and CNN verify that the network flow based on class-pathway is\na reasonable explanation for ANN models.\n",
"title": "Understanding the Feedforward Artificial Neural Network Model From the Perspective of Network Flow"
}
| null | null | null | null | true | null |
12278
| null |
Default
| null | null |
null |
{
"abstract": " We demonstrate autoparametric excitation of two distinct sub-harmonic\nmechanical modes by the same driven mechanical mode corresponding to different\ndrive frequencies within its resonance dispersion band. This experimental\nobservation is used to motivate a more general physical picture wherein\nmultiple mechanical modes could be excited by the same driven primary mode\nwithin the same device as long as the frequency spacing between the\nsub-harmonic modes is less than half the dispersion bandwidth of the driven\nprimary mode. The excitation of both modes is seen to be threshold-dependent\nand a parametric back-action is observed impacting on the response of the\ndriven primary mode. Motivated by this experimental observation, modified\ndynamical equations specifying 2-mode auto-parametric excitation for such\nsystems are presented.\n",
"title": "Excitation of multiple 2-mode parametric resonances by a single driven mode"
}
| null | null | null | null | true | null |
12279
| null |
Default
| null | null |
null |
{
"abstract": " We present a community-led assessment of the solar system investigations\nachievable with NASA's next-generation space telescope, the Wide Field InfraRed\nSurvey Telescope (WFIRST). WFIRST will provide imaging, spectroscopic, and\ncoronagraphic capabilities from 0.43-2.0 $\\mu$m and will be a potential\ncontemporary and eventual successor to JWST. Surveys of irregular satellites\nand minor bodies are where WFIRST will excel with its 0.28 deg$^2$ field of\nview Wide Field Instrument (WFI). Potential ground-breaking discoveries from\nWFIRST could include detection of the first minor bodies orbiting in the Inner\nOort Cloud, identification of additional Earth Trojan asteroids, and the\ndiscovery and characterization of asteroid binary systems similar to\nIda/Dactyl. Additional investigations into asteroids, giant planet satellites,\nTrojan asteroids, Centaurs, Kuiper Belt Objects, and comets are presented.\nPrevious use of astrophysics assets for solar system science and synergies\nbetween WFIRST, LSST, JWST, and the proposed NEOCam mission are discussed. We\nalso present the case for implementation of moving target tracking, a feature\nthat will benefit from the heritage of JWST and enable a broader range of solar\nsystem observations.\n",
"title": "Solar system science with the Wide-Field InfraRed Survey Telescope (WFIRST)"
}
| null | null |
[
"Physics"
] | null | true | null |
12280
| null |
Validated
| null | null |
null |
{
"abstract": " Unsupervised dependency parsing aims to learn a dependency parser from\nunannotated sentences. Existing work focuses on either learning generative\nmodels using the expectation-maximization algorithm and its variants, or\nlearning discriminative models using the discriminative clustering algorithm.\nIn this paper, we propose a new learning strategy that learns a generative\nmodel and a discriminative model jointly based on the dual decomposition\nmethod. Our method is simple and general, yet effective to capture the\nadvantages of both models and improve their learning results. We tested our\nmethod on the UD treebank and achieved a state-of-the-art performance on thirty\nlanguages.\n",
"title": "Combining Generative and Discriminative Approaches to Unsupervised Dependency Parsing via Dual Decomposition"
}
| null | null |
[
"Computer Science"
] | null | true | null |
12281
| null |
Validated
| null | null |
null |
{
"abstract": " We report the results of broadband (0.95--2.46 $\\mu$m) near-infrared\nspectroscopic observations of the Cassiopeia A supernova remnant. Using a\nclump-finding algorithm in two-dimensional dispersed images, we identify 63\n\"knots\" from eight slit positions and derive their spectroscopic properties.\nAll of the knots emit [Fe II] lines together with other ionic forbidden lines\nof heavy elements, and some of them also emit H and He lines. We identify 46\nemission line features in total from the 63 knots and measure their fluxes and\nradial velocities. The results of our analyses of the emission line features\nbased on principal component analysis show that the knots can be classified\ninto three groups: (1) He-rich, (2) S-rich, and (3) Fe-rich knots. The He-rich\nknots have relatively small, $\\lesssim 200~{\\rm km~s}^{-1}$, line-of-sight\nspeeds and radiate strong He I and [Fe II] lines resembling closely optical\nquasi-stationary flocculi of circumstellar medium, while the S-rich knots show\nstrong lines from O-burning material with large radial velocities up to $\\sim\n2000~{\\rm km~s}^{-1}$ indicating that they are supernova ejecta material known\nas fast-moving knots. The Fe-rich knots also have large radial velocities but\nshow no lines from O-burning material. We discuss the origin of the Fe-rich\nknots and conclude that they are most likely \"pure\" Fe ejecta synthesized in\nthe innermost region during the supernova explosion. The comparison of [Fe II]\nimages with other waveband images shows that these dense Fe ejecta are mainly\ndistributed along the southwestern shell just outside the unshocked $^{44}$Ti\nin the interior, supporting the presence of unshocked Fe associated with\n$^{44}$Ti.\n",
"title": "Near-Infrared Knots and Dense Fe Ejecta in the Cassiopeia A Supernova Remnant"
}
| null | null | null | null | true | null |
12282
| null |
Default
| null | null |
null |
{
"abstract": " We give a survey on some results covering the last 60 years concerning\nJeśmanowicz' conjecture. Moreover, we conclude the survey with a new result\nby showing that the special Diophantine equation $$(20k)^x+(99k)^y=(101k)^z$$\nhas no solution other than $(x,y,z)=(2,2,2)$.\n",
"title": "On the conjecture of Jeśmanowicz"
}
| null | null | null | null | true | null |
12283
| null |
Default
| null | null |
null |
{
"abstract": " Real time evolution of classical gauge fields is relevant for a number of\napplications in particle physics and cosmology, ranging from the early Universe\nto dynamics of quark-gluon plasma. We present a lattice formulation of the\ninteraction between a $shift$-symmetric field and some $U(1)$ gauge sector,\n$a(x)\\tilde{F}_{\\mu\\nu}F^{\\mu\\nu}$, reproducing the continuum limit to order\n$\\mathcal{O}(dx_\\mu^2)$ and obeying the following properties: (i) the system is\ngauge invariant and (ii) shift symmetry is exact on the lattice. For this end\nwe construct a definition of the {\\it topological number density} $Q =\n\\tilde{F}_{\\mu\\nu}F^{\\mu\\nu}$ that admits a lattice total derivative\nrepresentation $Q = \\Delta_\\mu^+ K^\\mu$, reproducing to order\n$\\mathcal{O}(dx_\\mu^2)$ the continuum expression $Q = \\partial_\\mu K^\\mu\n\\propto \\vec E \\cdot \\vec B$. If we consider a homogeneous field $a(x) = a(t)$,\nthe system can be mapped into an Abelian gauge theory with Hamiltonian\ncontaining a Chern-Simons term for the gauge fields. This allow us to study in\nan accompanying paper the real time dynamics of fermion number non-conservation\n(or chirality breaking) in Abelian gauge theories at finite temperature. When\n$a(x) = a(\\vec x,t)$ is inhomogeneous, the set of lattice equations of motion\ndo not admit however a simple explicit local solution (while preserving an\n$\\mathcal{O}(dx_\\mu^2)$ accuracy). We discuss an iterative scheme allowing to\novercome this difficulty.\n",
"title": "Lattice implementation of Abelian gauge theories with Chern-Simons number and an axion field"
}
| null | null | null | null | true | null |
12284
| null |
Default
| null | null |
null |
{
"abstract": " Dermoscopy image detection stays a tough task due to the weak distinguishable\nproperty of the object.Although the deep convolution neural network\nsignifigantly boosted the performance on prevelance computer vision tasks in\nrecent years,there remains a room to explore more robust and precise models to\nthe problem of low contrast image segmentation.Towards the challenge of Lesion\nSegmentation in ISBI 2017,we built a symmetrical identity inception fully\nconvolution network which is based on only 10 reversible inception blocks,every\nblock composed of four convolution branches with combination of different layer\ndepth and kernel size to extract sundry semantic features.Then we proposed an\napproximate loss function for jaccard index metrics to train our model.To\novercome the drawbacks of traditional convolution,we adopted the dilation\nconvolution and conditional random field method to rectify our segmentation.We\nalso introduced multiple ways to prevent the problem of overfitting.The\nexperimental results shows that our model achived jaccard index of 0.82 and\nkept learning from epoch to epoch.\n",
"title": "II-FCN for skin lesion analysis towards melanoma detection"
}
| null | null | null | null | true | null |
12285
| null |
Default
| null | null |
null |
{
"abstract": " The current dominant paradigm for imitation learning relies on strong\nsupervision of expert actions to learn both 'what' and 'how' to imitate. We\npursue an alternative paradigm wherein an agent first explores the world\nwithout any expert supervision and then distills its experience into a\ngoal-conditioned skill policy with a novel forward consistency loss. In our\nframework, the role of the expert is only to communicate the goals (i.e., what\nto imitate) during inference. The learned policy is then employed to mimic the\nexpert (i.e., how to imitate) after seeing just a sequence of images\ndemonstrating the desired task. Our method is 'zero-shot' in the sense that the\nagent never has access to expert actions during training or for the task\ndemonstration at inference. We evaluate our zero-shot imitator in two\nreal-world settings: complex rope manipulation with a Baxter robot and\nnavigation in previously unseen office environments with a TurtleBot. Through\nfurther experiments in VizDoom simulation, we provide evidence that better\nmechanisms for exploration lead to learning a more capable policy which in turn\nimproves end task performance. Videos, models, and more details are available\nat this https URL\n",
"title": "Zero-Shot Visual Imitation"
}
| null | null | null | null | true | null |
12286
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we extend and complement previous works about propagation in\nkinetic reaction-transport equations. The model we study describes particles\nmoving according to a velocity-jump process, and proliferating according to a\nreaction term of monostable type. We focus on the case of bounded velocities,\nhaving dimension higher than one. We extend previous results obtained by the\nfirst author with Calvez and Nadin in dimension one. We study the large\ntime/large scale hyperbolic limit via an Hamilton-Jacobi framework together\nwith the half-relaxed limits method. We deduce spreading results and the\nexistence of travelling wave solutions. A crucial difference with the\nmono-dimensional case is the resolution of the spectral problem at the edge of\nthe front, that yields potential singular velocity distributions. As a\nconsequence, the minimal speed of propagation may not be determined by a first\norder condition.\n",
"title": "Spreading in kinetic reaction-transport equations in higher velocity dimensions"
}
| null | null | null | null | true | null |
12287
| null |
Default
| null | null |
null |
{
"abstract": " Let $F$ be a non-Archimedan local field, $G$ a connected reductive group\ndefined and split over $F$, and $T$ a maximal $F$-split torus in $G$. Let\n$\\chi_0$ be a depth zero character of the maximal compact subgroup\n$\\mathcal{T}$ of $T(F)$. It gives by inflation a character $\\rho$ of an Iwahori\nsubgroup $\\mathcal{I}$ of $G(F)$ containing $\\mathcal{T}$. From Roche, $\\chi_0$\ndefines a split endoscopic group $G'$ of $G$, and there is an injective\nmorphism of ${\\Bbb C}$-algebras $\\mathcal{H}(G(F),\\rho) \\rightarrow\n\\mathcal{H}(G'(F),1_{\\mathcal{I}'})$ where $\\mathcal{H}(G(F),\\rho)$ is the\nHecke algebra of compactly supported $\\rho^{-1}$-spherical functions on $G(F)$\nand $\\mathcal{I}'$ is an Iwahori subgroup of $G'(F)$. This morphism restricts\nto an injective morphism $\\zeta: \\mathcal{Z}(G(F),\\rho)\\rightarrow\n\\mathcal{Z}(G'(F),1_{\\mathcal{I}'})$ between the centers of the Hecke algebras.\nWe prove here that a certain linear combination of morphisms analogous to\n$\\zeta$ realizes the transfer (matching of strongly $G$-regular semisimple\norbital integrals). If ${\\rm char}(F)=p>0$, our result is unconditional only if\n$p$ is large enough.\n",
"title": "Matching of orbital integrals (transfer) and Roche Hecke algebra isomorphisms"
}
| null | null | null | null | true | null |
12288
| null |
Default
| null | null |
null |
{
"abstract": " Accurate rates for energy-degenerate l-changing collisions are needed to\ndetermine cosmological abundances and recombination. There are now several\ncompeting theories for the treatment of this process, and it is not possible to\ntest these experimentally. We show that the H I two-photon continuum produced\nby astrophysical nebulae is strongly affected by l-changing collisions. We\nperform an analysis of the different underlying atomic processes and simulate\nthe recombination and two-photon spectrum of a nebula containing H and He. We\nprovide an extended set of effective recombination coefficients and updated\nl-changing 2s-2p transition rates using several competing theories. In\nprinciple, accurate astronomical observations could determine which theory is\ncorrect.\n",
"title": "Testing atomic collision theory with the two-photon continuum of astrophysical nebulae"
}
| null | null | null | null | true | null |
12289
| null |
Default
| null | null |
null |
{
"abstract": " We study theoretically the usefulness of spin-1 Bose condensates with\nmacroscopic magnetization in a homogeneous magnetic field for quantum\nmetrology. We demonstrate Heisenberg scaling of the quantum Fisher information\nfor states in thermal equilibrium. The scaling applies to both\nantiferromagnetic and ferromagnetic interactions. The effect preserves as long\nas fluctuations of magnetization are sufficiently small. Scaling of the quantum\nFisher information with the total particle number is derived within the\nmean-field approach in the zero temperature limit and exactly in the high\nmagnetic field limit for any temperature. The precision gain is intuitively\nexplained owing to subtle features of the quasi-distribution function in phase\nspace.\n",
"title": "Metrologically useful states of spin-1 Bose condensates with macroscopic magnetization"
}
| null | null | null | null | true | null |
12290
| null |
Default
| null | null |
null |
{
"abstract": " The magnetic insulator Yttrium Iron Garnet can be grown with exceptional\nquality, has a ferrimagnetic transition temperature of nearly 600 K, and is\nused in microwave and spintronic devices that can operate at room temperature.\nThe most accurate prior measurements of the magnon spectrum date back nearly 40\nyears, but cover only 3 of the lowest energy modes out of 20 distinct magnon\nbranches. Here we have used time-of-flight inelastic neutron scattering to\nmeasure the full magnon spectrum throughout the Brillouin zone. We find that\nthe existing model of the excitation spectrum, well known from an earlier work\ntitled \"The Saga of YIG\", fails to describe the optical magnon modes. Using a\nvery general spin Hamiltonian, we show that the magnetic interactions are both\nlonger-ranged and more complex than was previously understood. The results\nprovide the basis for accurate microscopic models of the finite temperature\nmagnetic properties of Yttrium Iron Garnet, necessary for next-generation\nelectronic devices.\n",
"title": "The Final Chapter In The Saga Of YIG"
}
| null | null |
[
"Physics"
] | null | true | null |
12291
| null |
Validated
| null | null |
null |
{
"abstract": " We characterize a multi tier network with classical macro cells, and multi\nradio access technology (RAT) small cells, which are able to operate in\nmicrowave and millimeter-wave (mm-wave) bands. The small cells are assumed to\nbe deployed along roads modeled as a Poisson line process. This\ncharacterization is more realistic as compared to the classical Poisson point\nprocesses typically used in literature. In this context, we derive the\nassociation and RAT selection probabilities of the typical user under various\nsystem parameters such as the small cell deployment density and mm-wave antenna\ngain, and with varying street densities. Finally, we calculate the signal to\ninterference plus noise ratio (SINR) coverage probability for the typical user\nconsidering a tractable dominant interference based model for mm-wave\ninterference. Our analysis reveals the need of deploying more small cells per\nstreet in cities with more streets to maintain coverage, and highlights that\nmm-wave RAT in small cells can help to improve the SINR performance of the\nusers.\n",
"title": "Modeling and Analysis of HetNets with mm-Wave Multi-RAT Small Cells Deployed Along Roads"
}
| null | null | null | null | true | null |
12292
| null |
Default
| null | null |
null |
{
"abstract": " A class of nonlinear Schrödinger equations involving a triad of power law\nterms together with a de Broglie-Bohm potential is shown to admit symmetry\nreduction to a hybrid Ermakov-Painlevé II equation which is linked, in turn,\nto the integrable Painlevé XXXIV equation. A nonlinear Schrödinger\nencapsulation of a Korteweg-type capillary system is thereby used in the\nisolation of such a Ermakov-Painlevé II reduction valid for a multi-parameter\nclass of free energy functions. Iterated application of a Bäcklund\ntransformation then allows the construction of novel classes of exact solutions\nof the nonlinear capillarity system in terms of Yablonskii-Vorob'ev polynomials\nor classical Airy functions. A Painlevé XXXIV equation is derived for the\ndensity in the capillarity system and seen to correspond to the symmetry\nreduction of its Bernoulli integral of motion.\n",
"title": "Ermakov-Painlevé II Symmetry Reduction of a Korteweg Capillarity System"
}
| null | null | null | null | true | null |
12293
| null |
Default
| null | null |
null |
{
"abstract": " Matrix completion models are among the most common formulations of\nrecommender systems. Recent works have showed a boost of performance of these\ntechniques when introducing the pairwise relationships between users/items in\nthe form of graphs, and imposing smoothness priors on these graphs. However,\nsuch techniques do not fully exploit the local stationarity structures of\nuser/item graphs, and the number of parameters to learn is linear w.r.t. the\nnumber of users and items. We propose a novel approach to overcome these\nlimitations by using geometric deep learning on graphs. Our matrix completion\narchitecture combines graph convolutional neural networks and recurrent neural\nnetworks to learn meaningful statistical graph-structured patterns and the\nnon-linear diffusion process that generates the known ratings. This neural\nnetwork system requires a constant number of parameters independent of the\nmatrix size. We apply our method on both synthetic and real datasets, showing\nthat it outperforms state-of-the-art techniques.\n",
"title": "Geometric Matrix Completion with Recurrent Multi-Graph Neural Networks"
}
| null | null | null | null | true | null |
12294
| null |
Default
| null | null |
null |
{
"abstract": " This paper is concerned with two-person dynamic zero-sum games. Let games for\nsome family have common dynamics, running costs and capabilities of players,\nand let these games differ in densities only. We show that the Dynamic\nProgramming Principle directly leads to the General Tauberian Theorem---that\nthe existence of a uniform limit of the value functions for uniform\ndistribution or for exponential distribution implies that the value functions\nuniformly converge to the same limit for arbitrary distribution from large\nclass. No assumptions on strategies are necessary. Applications to differential\ngames and stochastic statement are considered.\n",
"title": "Value Asymptotics in Dynamic Games on Large Horizons"
}
| null | null | null | null | true | null |
12295
| null |
Default
| null | null |
null |
{
"abstract": " In the context of dissipative systems, we show that for any quantum chaotic\nattractor a corre- sponding classical chaotic attractor can always be found. We\nprovide with a general way to locate them, rooted in the structure of the\nparameter space (which is typically bidimensional, accounting for the forcing\nstrength and dissipation parameters). In the cases where an approximate point\nlike quantum distribution is found, it can be associated to exceptionally large\nregular structures. Moreover, supposedly anomalous quantum chaotic behaviour\ncan be very well reproduced by the classical dynamics plus Gaussian noise of\nthe size of an effective Planck constant $\\hbar_{\\rm eff}$. We give support to\nour conjectures by means of two paradigmatic examples of quantum chaos and\ntransport theory. In particular, a dissipative driven system becomes\nfundamental in order to extend their validity to generic cases.\n",
"title": "Classical counterparts of quantum attractors in generic dissipative systems"
}
| null | null |
[
"Physics"
] | null | true | null |
12296
| null |
Validated
| null | null |
null |
{
"abstract": " In a recent paper [1] we introduced the Fuzzy Bayesian Learning (FBL)\nparadigm where expert opinions can be encoded in the form of fuzzy rule bases\nand the hyper-parameters of the fuzzy sets can be learned from data using a\nBayesian approach. The present paper extends this work for selecting the most\nappropriate rule base among a set of competing alternatives, which best\nexplains the data, by calculating the model evidence or marginal likelihood. We\nexplain why this is an attractive alternative over simply minimizing a mean\nsquared error metric of prediction and show the validity of the proposition\nusing synthetic examples and a real world case study in the financial services\nsector.\n",
"title": "Marginal likelihood based model comparison in Fuzzy Bayesian Learning"
}
| null | null | null | null | true | null |
12297
| null |
Default
| null | null |
null |
{
"abstract": " The excitement and convergence of tweets on specific topics are well studied.\nHowever, by utilizing the position information of Tweet, it is also possible to\nanalyze the position-sensitive tweet. In this research, we focus on bomb\nterrorist attacks and propose a method for separately analyzing the number of\ntweets at the place where the incident occurred, nearby, and far. We made\nmeasurements of position-sensitive tweets and suggested a theory to explain it.\nThis theory is an extension of the mathematical model of the hit phenomenon.\n",
"title": "Position-sensitive propagation of information on social media using social physics approach"
}
| null | null | null | null | true | null |
12298
| null |
Default
| null | null |
null |
{
"abstract": " Monte Carlo method is a broad class of computational algorithms that rely on\nrepeated random sampling to obtain numerical results. They are often used in\nphysical and mathematical problems and are most useful when it is difficult or\nimpossible to use other mathematical methods. Basically, many statisticians\nhave been increasingly drawn to Monte Carlo method in three distinct problem\nclasses: optimization, numerical integration, and generating draws from a\nprobability distribution. In this paper, we will introduce the Monte Carlo\nmethod for calculating coefficients in Generalized Linear Model(GLM),\nespecially for Logistic Regression. Our main methods are Metropolis\nHastings(MH) Algorithms and Stochastic Approximation in Monte Carlo\nComputation(SAMC). For comparison, we also get results automatically using MLE\nmethod in R software.\n",
"title": "The application of Monte Carlo methods for learning generalized linear model"
}
| null | null |
[
"Statistics"
] | null | true | null |
12299
| null |
Validated
| null | null |
null |
{
"abstract": " We address the problem of bootstrapping language acquisition for an\nartificial system similarly to what is observed in experiments with human\ninfants. Our method works by associating meanings to words in manipulation\ntasks, as a robot interacts with objects and listens to verbal descriptions of\nthe interactions. The model is based on an affordance network, i.e., a mapping\nbetween robot actions, robot perceptions, and the perceived effects of these\nactions upon objects. We extend the affordance model to incorporate spoken\nwords, which allows us to ground the verbal symbols to the execution of actions\nand the perception of the environment. The model takes verbal descriptions of a\ntask as the input and uses temporal co-occurrence to create links between\nspeech utterances and the involved objects, actions, and effects. We show that\nthe robot is able form useful word-to-meaning associations, even without\nconsidering grammatical structure in the learning process and in the presence\nof recognition errors. These word-to-meaning associations are embedded in the\nrobot's own understanding of its actions. Thus, they can be directly used to\ninstruct the robot to perform tasks and also allow to incorporate context in\nthe speech recognition task. We believe that the encouraging results with our\napproach may afford robots with a capacity to acquire language descriptors in\ntheir operation's environment as well as to shed some light as to how this\nchallenging process develops with human infants.\n",
"title": "Language Bootstrapping: Learning Word Meanings From Perception-Action Association"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
12300
| null |
Validated
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.