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": " Zero-shot learning (ZSL) endows the computer vision system with the\ninferential capability to recognize instances of a new category that has never\nseen before. Two fundamental challenges in it are visual-semantic embedding and\ndomain adaptation in cross-modality learning and unseen class prediction steps,\nrespectively. To address both challenges, this paper presents two corresponding\nmethods named Adaptive STructural Embedding (ASTE) and Self-PAsed Selective\nStrategy (SPASS), respectively. Specifically, ASTE formulates the\nvisualsemantic interactions in a latent structural SVM framework to adaptively\nadjust the slack variables to embody the different reliableness among training\ninstances. In this way, the reliable instances are imposed with small\npunishments, wheras the less reliable instances are imposed with more severe\npunishments. Thus, it ensures a more discriminative embedding. On the other\nhand, SPASS offers a framework to alleviate the domain shift problem in ZSL,\nwhich exploits the unseen data in an easy to hard fashion. Particularly, SPASS\nborrows the idea from selfpaced learning by iteratively selecting the unseen\ninstances from reliable to less reliable to gradually adapt the knowledge from\nthe seen domain to the unseen domain. Subsequently, by combining SPASS and\nASTE, we present a self-paced Transductive ASTE (TASTE) method to progressively\nreinforce the classification capacity. Extensive experiments on three benchmark\ndatasets (i.e., AwA, CUB, and aPY) demonstrate the superiorities of ASTE and\nTASTE. Furthermore, we also propose a fast training (FT) strategy to improve\nthe efficiency of most of existing ZSL methods. The FT strategy is surprisingly\nsimple and general enough, which can speed up the training time of most\nexisting methods by 4~300 times while holding the previous performance.\n",
"title": "Transductive Zero-Shot Learning with Adaptive Structural Embedding"
}
| null | null | null | null | true | null |
19201
| null |
Default
| null | null |
null |
{
"abstract": " Grid cells in the medial entorhinal cortex (mEC) respond when an animal\noccupies a periodic lattice of \"grid fields\" in the environment. The grids are\norganized in modules with spatial periods clustered around discrete values\nseparated by constant ratios reported in the range 1.3-1.8. We propose a\nmechanism for dynamical self-organization in the mEC that can produce this\nmodular structure. In attractor network models of grid formation, the period of\na single module is set by the length scale of recurrent inhibition between\nneurons. We show that grid cells will instead form a hierarchy of discrete\nmodules if a continuous increase in inhibition distance along the dorso-ventral\naxis of the mEC is accompanied by excitatory interactions along this axis.\nMoreover, constant scale ratios between successive modules arise through\ngeometric relationships between triangular grids, whose lattice constants are\nseparated by $\\sqrt{3} \\approx 1.7$, $\\sqrt{7}/2 \\approx 1.3$, or other ratios.\nWe discuss how the interactions required by our model might be tested\nexperimentally and realized by circuits in the mEC.\n",
"title": "A geometric attractor mechanism for self-organization of entorhinal grid modules"
}
| null | null | null | null | true | null |
19202
| null |
Default
| null | null |
null |
{
"abstract": " Online social network (OSN) discussion groups are exerting significant\neffects on political dialogue. In the absence of access control mechanisms, any\nuser can contribute to any OSN thread. Individuals can exploit this\ncharacteristic to execute targeted attacks, which increases the potential for\nsubsequent malicious behaviors such as phishing and malware distribution. These\nkinds of actions will also disrupt bridges among the media, politicians, and\ntheir constituencies.\nFor the concern of Security Management, blending malicious cyberattacks with\nonline social interactions has introduced a brand new challenge. In this paper\nwe describe our proposal for a novel approach to studying and understanding the\nstrategies that attackers use to spread malicious URLs across Facebook\ndiscussion groups. We define and analyze problems tied to predicting the\npotential for attacks focused on threads created by news media organizations.\nWe use a mix of macro static features and the micro dynamic evolution of posts\nand threads to identify likely targets with greater than 90% accuracy. One of\nour secondary goals is to make such predictions within a short (10 minute) time\nframe. It is our hope that the data and analyses presented in this paper will\nsupport a better understanding of attacker strategies and footprints, thereby\ndeveloping new system management methodologies in handing cyber attacks on\nsocial networks.\n",
"title": "Attacking Strategies and Temporal Analysis Involving Facebook Discussion Groups"
}
| null | null | null | null | true | null |
19203
| null |
Default
| null | null |
null |
{
"abstract": " Image stitching is challenging in consumer-level photography, due to\nalignment difficulties in unconstrained shooting environment. Recent studies\nshow that seam-cutting approaches can effectively relieve artifacts generated\nby local misalignment. Normally, seam-cutting is described in terms of energy\nminimization, however, few of existing methods consider human perception in\ntheir energy functions, which sometimes causes that a seam with minimum energy\nis not most invisible in the overlapping region. In this paper, we propose a\nnovel perception-based energy function in the seam-cutting framework, which\nconsiders the nonlinearity and the nonuniformity of human perception in energy\nminimization. Our perception-based approach adopts a sigmoid metric to\ncharacterize the perception of color discrimination, and a saliency weight to\nsimulate that human eyes incline to pay more attention to salient objects. In\naddition, our seam-cutting composition can be easily implemented into other\nstitching pipelines. Experiments show that our method outperforms the\nseam-cutting method of the normal energy function, and a user study\ndemonstrates that our composed results are more consistent with human\nperception.\n",
"title": "Perception-based energy functions in seam-cutting"
}
| null | null | null | null | true | null |
19204
| null |
Default
| null | null |
null |
{
"abstract": " Probabilistic modeling provides the capability to represent and manipulate\nuncertainty in data, models, predictions and decisions. We are concerned with\nthe problem of learning probabilistic models of dynamical systems from measured\ndata. Specifically, we consider learning of probabilistic nonlinear state-space\nmodels. There is no closed-form solution available for this problem, implying\nthat we are forced to use approximations. In this tutorial we will provide a\nself-contained introduction to one of the state-of-the-art methods---the\nparticle Metropolis--Hastings algorithm---which has proven to offer a practical\napproximation. This is a Monte Carlo based method, where the particle filter is\nused to guide a Markov chain Monte Carlo method through the parameter space.\nOne of the key merits of the particle Metropolis--Hastings algorithm is that it\nis guaranteed to converge to the \"true solution\" under mild assumptions,\ndespite being based on a particle filter with only a finite number of\nparticles. We will also provide a motivating numerical example illustrating the\nmethod using a modeling language tailored for sequential Monte Carlo methods.\nThe intention of modeling languages of this kind is to open up the power of\nsophisticated Monte Carlo methods---including particle\nMetropolis--Hastings---to a large group of users without requiring them to know\nall the underlying mathematical details.\n",
"title": "Probabilistic learning of nonlinear dynamical systems using sequential Monte Carlo"
}
| null | null | null | null | true | null |
19205
| null |
Default
| null | null |
null |
{
"abstract": " Semantic segmentation of 3D point clouds is a challenging problem with\nnumerous real-world applications. While deep learning has revolutionized the\nfield of image semantic segmentation, its impact on point cloud data has been\nlimited so far. Recent attempts, based on 3D deep learning approaches\n(3D-CNNs), have achieved below-expected results. Such methods require\nvoxelizations of the underlying point cloud data, leading to decreased spatial\nresolution and increased memory consumption. Additionally, 3D-CNNs greatly\nsuffer from the limited availability of annotated datasets.\nIn this paper, we propose an alternative framework that avoids the\nlimitations of 3D-CNNs. Instead of directly solving the problem in 3D, we first\nproject the point cloud onto a set of synthetic 2D-images. These images are\nthen used as input to a 2D-CNN, designed for semantic segmentation. Finally,\nthe obtained prediction scores are re-projected to the point cloud to obtain\nthe segmentation results. We further investigate the impact of multiple\nmodalities, such as color, depth and surface normals, in a multi-stream network\narchitecture. Experiments are performed on the recent Semantic3D dataset. Our\napproach sets a new state-of-the-art by achieving a relative gain of 7.9 %,\ncompared to the previous best approach.\n",
"title": "Deep Projective 3D Semantic Segmentation"
}
| null | null | null | null | true | null |
19206
| null |
Default
| null | null |
null |
{
"abstract": " Motivated by posted price auctions where buyers are grouped in an unknown\nnumber of latent types characterized by their private values for the good on\nsale, we investigate revenue maximization in stochastic dynamic pricing when\nthe distribution of buyers' private values is supported on an unknown set of\npoints in [0,1] of unknown cardinality K. This setting can be viewed as an\ninstance of a stochastic K-armed bandit problem where the location of the arms\n(the K unknown valuations) must be learned as well. In the distribution-free\ncase, we show that our setting is just as hard as K-armed stochastic bandits:\nwe prove that no algorithm can achieve a regret significantly better than\n$\\sqrt{KT}$, (where T is the time horizon) and present an efficient algorithm\nmatching this lower bound up to logarithmic factors. In the\ndistribution-dependent case, we show that for all K>2 our setting is strictly\nharder than K-armed stochastic bandits by proving that it is impossible to\nobtain regret bounds that grow logarithmically in time or slower. On the other\nhand, when a lower bound $\\gamma>0$ on the smallest drop in the demand curve is\nknown, we prove an upper bound on the regret of order $(1/\\Delta+(\\log \\log\nT)/\\gamma^2)(K\\log T)$. This is a significant improvement on previously known\nregret bounds for discontinuous demand curves, that are at best of order\n$(K^{12}/\\gamma^8)\\sqrt{T}$. When K=2 in the distribution-dependent case, the\nhardness of our setting reduces to that of a stochastic 2-armed bandit: we\nprove that an upper bound of order $(\\log T)/\\Delta$ (up to $\\log\\log$ factors)\non the regret can be achieved with no information on the demand curve. Finally,\nwe show a $O(\\sqrt{T})$ upper bound on the regret for the setting in which the\nbuyers' decisions are nonstochastic, and the regret is measured with respect to\nthe best between two fixed valuations one of which is known to the seller.\n",
"title": "Dynamic Pricing with Finitely Many Unknown Valuations"
}
| null | null | null | null | true | null |
19207
| null |
Default
| null | null |
null |
{
"abstract": " We analyse certain Haar systems associated to groupoids obtained by certain\nnatural equivalence relations of dynamical nature on sets like\n$\\{1,2,...,d\\}^\\mathbb{Z}$, $\\{1,2,...,d\\}^\\mathbb{N}$, $S^1\\times S^1$, or\n$(S^1)^\\mathbb{N}$, where $S^1$ is the unitary circle. We also describe\nproperties of transverse functions, quasi-invariant probabilities and KMS\nstates for some examples of von Neumann algebras (and also $C^*$-Algebras)\nassociated to these groupoids. We relate some of these KMS states with Gibbs\nstates of Thermodynamic Formalism. While presenting new results, we will also\ndescribe in detail several examples and basic results on the above topics. In\nother words it is also a survey paper. Some known results on non-commutative\nintegration are presented, more precisely, the relation of transverse measures,\ncocycles and quasi-invariant probabilities.\nWe describe the results in a language which is more familiar to people in\nDynamical Systems. Our intention is to study Haar systems, quasi-invariant\nprobabilities and von Neumann algebras as a topic on measure theory\n(intersected with ergodic theory) avoiding questions of algebraic nature\n(which, of course, are also extremely important).\n",
"title": "Haar systems, KMS states on von Neumann algebras and $C^*$-algebras on dynamically defined groupoids and Noncommutative Integration"
}
| null | null | null | null | true | null |
19208
| null |
Default
| null | null |
null |
{
"abstract": " Numerical (and experimental) data analysis often requires the restoration of\na smooth function from a set of sampled integrals over finite bins. We present\nthe bin hierarchy method that efficiently computes the maximally smooth\nfunction from the sampled integrals using essentially all the information\ncontained in the data. We perform extensive tests with different classes of\nfunctions and levels of data quality, including Monte Carlo data suffering from\na severe sign problem and physical data for the Green's function of the\nFröhlich polaron.\n",
"title": "Restoring a smooth function from its noisy integrals"
}
| null | null | null | null | true | null |
19209
| null |
Default
| null | null |
null |
{
"abstract": " We propose a new active learning strategy designed for deep neural networks.\nThe goal is to minimize the number of data annotation queried from an oracle\nduring training. Previous active learning strategies scalable for deep networks\nwere mostly based on uncertain sample selection. In this work, we focus on\nexamples lying close to the decision boundary. Based on theoretical works on\nmargin theory for active learning, we know that such examples may help to\nconsiderably decrease the number of annotations. While measuring the exact\ndistance to the decision boundaries is intractable, we propose to rely on\nadversarial examples. We do not consider anymore them as a threat instead we\nexploit the information they provide on the distribution of the input space in\norder to approximate the distance to decision boundaries. We demonstrate\nempirically that adversarial active queries yield faster convergence of CNNs\ntrained on MNIST, the Shoe-Bag and the Quick-Draw datasets.\n",
"title": "Adversarial Active Learning for Deep Networks: a Margin Based Approach"
}
| null | null | null | null | true | null |
19210
| null |
Default
| null | null |
null |
{
"abstract": " Nowadays, the Security Information and Event Management (SIEM) systems take\non great relevance in handling security issues for critical infrastructures as\nInternet Service Providers. Basically, a SIEM has two main functions: i) the\ncollection and the aggregation of log data and security information from\ndisparate network devices (routers, firewalls, intrusion detection systems, ad\nhoc probes and others) and ii) the analysis of the gathered data by\nimplementing a set of correlation rules aimed at detecting potential suspicious\nevents as the presence of encrypted real-time traffic. In the present work, the\nauthors propose an enhanced implementation of a SIEM where a particular focus\nis given to the detection of encrypted Skype traffic by using an ad-hoc\ndeveloped enhanced probe (ESkyPRO) conveniently governed by the SIEM itself.\nSuch enhanced probe, able to interact with an agent counterpart deployed into\nthe SIEM platform, is designed by exploiting some machine learning concepts.\nThe main purpose of the proposed ad-hoc SIEM is to correlate the information\nreceived by ESkyPRO and other types of data obtained by an Intrusion Detection\nSystem (IDS) probe in order to make the encrypted Skype traffic detection as\naccurate as possible.\n",
"title": "Improving SIEM capabilities through an enhanced probe for encrypted Skype traffic detection"
}
| null | null | null | null | true | null |
19211
| null |
Default
| null | null |
null |
{
"abstract": " Adaptive methods such as Adam and RMSProp are widely used in deep learning\nbut are not well understood. In this paper, we seek a crisp, clean and precise\ncharacterization of their behavior in nonconvex settings. To this end, we first\nprovide a novel view of adaptive methods as preconditioned SGD, where the\npreconditioner is estimated in an online manner. By studying the preconditioner\non its own, we elucidate its purpose: it rescales the stochastic gradient noise\nto be isotropic near stationary points, which helps escape saddle points.\nFurthermore, we show that adaptive methods can efficiently estimate the\naforementioned preconditioner. By gluing together these two components, we\nprovide the first (to our knowledge) second-order convergence result for any\nadaptive method. The key insight from our analysis is that, compared to SGD,\nadaptive methods escape saddle points faster, and can converge faster overall\nto second-order stationary points.\n",
"title": "Escaping Saddle Points with Adaptive Gradient Methods"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
19212
| null |
Validated
| null | null |
null |
{
"abstract": " Films of Cu-K-In-Se were co-evaporated at varied K/(K+Cu) compositions and\nsubstrate temperatures (with constant (K+Cu)/In ~ 0.85). Increased Na\ncomposition on the substrate's surface and decreased growth temperature were\nboth found to favor Cu1-xKxInSe2 (CKIS) alloy formation, relative to\nmixed-phase CuInSe2 + KInSe2 formation. Structures from X-ray diffraction\n(XRD), band gaps, resistivities, minority carrier lifetimes and carrier\nconcentrations from time-resolved photoluminescence were in agreement with\nprevious reports, where low K/(K+Cu) composition films exhibited properties\npromising for photovoltaic (PV) absorbers. Films grown at 400-500 C were then\nannealed to 600 C under Se, which caused K loss by evaporation in proportion to\ninitial K/(K+Cu) composition. Similar to growth temperature, annealing drove\nCKIS alloy consumption and CuInSe2 + KInSe2 production, as evidenced by high\ntemperature XRD. Annealing also decomposed KInSe2 and formed K2In12Se19. At\nhigh temperature the KInSe2 crystal lattice gradually contracted as temperature\nand time increased, as well as just time. Evaporative loss of K during\nannealing could accompany the generation of vacancies on K lattice sites, and\nmay explain the KInSe2 lattice contraction. This knowledge of Cu-K-In-Se\nmaterial chemistry may be used to predict and control minor phase impurities in\nCu(In,Ga)(Se,S)2 PV absorbers-where impurities below typical detection limits\nmay have played a role in recent world record PV efficiencies that utilized KF\npost-deposition treatments.\n",
"title": "The Effect of Temperature on Cu-K-In-Se Thin Films"
}
| null | null |
[
"Physics"
] | null | true | null |
19213
| null |
Validated
| null | null |
null |
{
"abstract": " Over the course of last decade, the Nice model has dramatically changed our\nview of the solar system's formation and early evolution. Within the context of\nthis model, a transient period of planet-planet scattering is triggered by\ngravitational interactions between the giant planets and a massive primordial\nplanetesimal disk, leading to a successful reproduction of the solar system's\npresent-day architecture. In typical realizations of the Nice model,\nself-gravity of the planetesimal disk is routinely neglected, as it poses a\ncomputational bottleneck to the calculations. Recent analyses have shown,\nhowever, that a self-gravitating disk can exhibit behavior that is dynamically\ndistinct, and this disparity may have significant implications for the solar\nsystem's evolutionary path. In this work, we explore this discrepancy utilizing\na large suite of Nice odel simulations with and without a self-gravitating\nplanetesimal disk, taking advantage of the inherently parallel nature of\ngraphic processing units. Our simulations demonstrate that self-consistent\nmodeling of particle interactions does not lead to significantly different\nfinal planetary orbits from those obtained within conventional simulations.\nMoreover, self-gravitating calculations show similar planetesimal evolution to\nnon-self-gravitating numerical experiments after dynamical instability is\ntriggered, suggesting that the orbital clustering observed in the distant\nKuiper belt is unlikely to have a self-gravitational origin.\n",
"title": "Simulations of the Solar System's Early Dynamical Evolution with a Self-Gravitating Planetesimal Disk"
}
| null | null | null | null | true | null |
19214
| null |
Default
| null | null |
null |
{
"abstract": " Using the natural action of $S_\\infty$ we show that a countable hereditary\nclass $\\cC$ of finitely generated structures has the joint embedding property\n(JEP) and the weak amalgamation property (WAP) if and only if there is a\nstructure $M$ whose isomorphism type is comeager in the space of all countable,\ninfinitely generated structures with age in $\\cC$. In this case, $M$ is the\nweak Fraïssé limit of $\\cC$.\nThis applies in particular to countable structures with generic automorphisms\nand recovers a result by Kechris and Rosendal [\\textit{Proc.\\ Lond.\\ Math.\\\nSoc.,\\ 2007}].\n",
"title": "On weak Fraisse limits"
}
| null | null |
[
"Mathematics"
] | null | true | null |
19215
| null |
Validated
| null | null |
null |
{
"abstract": " The cost of attending college has been steadily rising and in 10 years is\nestimated to reach $140,000 for a 4-year public university. Recent surveys\nestimate just over half of US families are saving for college. State-operated\n529 college savings plans are an effective way for families to plan and save\nfor future college costs, but only 3% of families currently use them. The\nOffice of the Illinois State Treasurer (Treasurer) administers two 529 plans to\nhelp its residents save for college. In order to increase the number of\nfamilies saving for college, the Treasurer and Civis Analytics used data\nscience techniques to identify the people most likely to sign up for a college\nsavings plan. In this paper, we will discuss the use of person matching to join\naccountholder data from the Treasurer to the Civis National File, as well as\nthe use of lookalike modeling to identify new potential signups. In order to\navoid reinforcing existing demographic imbalances in who saves for college, the\nlookalike models used were ensured to be racially and economically balanced. We\nwill also discuss how these new signup targets were then individually served\ndigital ads to encourage opening college savings accounts.\n",
"title": "Promoting Saving for College Through Data Science"
}
| null | null | null | null | true | null |
19216
| null |
Default
| null | null |
null |
{
"abstract": " Users of Virtual Reality (VR) systems often experience vection, the\nperception of self-motion in the absence of any physical movement. While\nvection helps to improve presence in VR, it often leads to a form of motion\nsickness called cybersickness. Cybersickness is a major deterrent to large\nscale adoption of VR.\nPrior work has discovered that changing vection (changing the perceived speed\nor moving direction) causes more severe cybersickness than steady vection\n(walking at a constant speed or in a constant direction). Based on this idea,\nwe try to reduce the cybersickness caused by character movements in a First\nPerson Shooter (FPS) game in VR. We propose Rotation Blurring (RB), uniformly\nblurring the screen during rotational movements to reduce cybersickness. We\nperformed a user study to evaluate the impact of RB in reducing cybersickness.\nWe found that the blurring technique led to an overall reduction in sickness\nlevels of the participants and delayed its onset. Participants who experienced\nacute levels of cybersickness benefited significantly from this technique.\n",
"title": "Rotation Blurring: Use of Artificial Blurring to Reduce Cybersickness in Virtual Reality First Person Shooters"
}
| null | null | null | null | true | null |
19217
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we present a comprehensive view of prominent causal discovery\nalgorithms, categorized into two main categories (1) assuming acyclic and no\nlatent variables, and (2) allowing both cycles and latent variables, along with\nexperimental results comparing them from three perspectives: (a) structural\naccuracy, (b) standard predictive accuracy, and (c) accuracy of counterfactual\ninference. For (b) and (c) we train causal Bayesian networks with structures as\npredicted by each causal discovery technique to carry out counterfactual or\nstandard predictive inference. We compare causal algorithms on two pub- licly\navailable and one simulated datasets having different sample sizes: small,\nmedium and large. Experiments show that structural accuracy of a technique does\nnot necessarily correlate with higher accuracy of inferencing tasks. Fur- ther,\nsurveyed structure learning algorithms do not perform well in terms of\nstructural accuracy in case of datasets having large number of variables.\n",
"title": "Comparative Benchmarking of Causal Discovery Techniques"
}
| null | null | null | null | true | null |
19218
| null |
Default
| null | null |
null |
{
"abstract": " We examine the sensitivity of the Love and the quasi-Rayleigh waves to model\nparameters. Both waves are guided waves that propagate in the same model of an\nelastic layer above an elastic halfspace. We study their dispersion curves\nwithout any simplifying assumptions, beyond the standard approach of elasticity\ntheory in isotropic media. We examine the sensitivity of both waves to\nelasticity parameters, frequency and layer thickness, for varying frequency and\ndifferent modes. In the case of Love waves, we derive and plot the absolute\nvalue of a dimensionless sensitivity coefficient in terms of partial\nderivatives, and perform an analysis to find the optimum frequency for\ndetermining the layer thickness. For a coherency of the background information,\nwe briefly review the Love-wave dispersion relation and provide details of the\nless common derivation of the quasi-Rayleigh relation in an appendix. We\ncompare that derivation to past results in the literature, finding certain\ndiscrepancies among them.\n",
"title": "Sensitivity of Love and quasi-Rayleigh waves to model parameters"
}
| null | null | null | null | true | null |
19219
| null |
Default
| null | null |
null |
{
"abstract": " Let $\\varphi:\\mathbb{F}_q\\to\\mathbb{F}_q$ be a rational map on a fixed finite\nfield. We give explicit asymptotic formulas for the size of image sets\n$\\varphi^n(\\mathbb{F}_q)$ as a function of $n$. This is done by using\nproperties of the Galois groups of iterated maps, whose connection to the\nquestion of the size of image sets is established via Chebotarev's Density\nTheorem. We then apply these results to provide explicit bounds on the\nproportion of periodic points in $\\mathbb{F}_q$ in terms of $q$ for certain\nrational maps.\n",
"title": "The image size of iterated rational maps over finite fields"
}
| null | null | null | null | true | null |
19220
| null |
Default
| null | null |
null |
{
"abstract": " We form real-analytic Eisenstein series twisted by Manin's noncommutative\nmodular symbols. After developing their basic properties, these series are\nshown to have meromorphic continuations to the entire complex plane and satisfy\nfunctional equations in some cases. This theory neatly contains and generalizes\nearlier work in the literature on the properties of Eisenstein series twisted\nby classical modular symbols.\n",
"title": "Noncommutative modular symbols and Eisenstein series"
}
| null | null | null | null | true | null |
19221
| null |
Default
| null | null |
null |
{
"abstract": " We study closed $n$-dimensional manifolds of which the metrics are critical\nfor quadratic curvature functionals involving the Ricci curvature, the scalar\ncurvature and the Riemannian curvature tensor on the space of Riemannian\nmetrics with unit volume. Under some additional integral conditions, we\nclassify such manifolds. Moreover, under some curvature conditions, the result\nthat a critical metric must be Einstein is proved.\n",
"title": "Some rigidity characterizations on critical metrics for quadratic curvature functionals"
}
| null | null | null | null | true | null |
19222
| null |
Default
| null | null |
null |
{
"abstract": " Using quantum representations of mapping class groups we prove that profinite\ncompletions of Burnside-type surface group quotients are not virtually\nprosolvable, in general. Further, we construct infinitely many finite simple\ncharacteristic quotients of surface groups.\n",
"title": "Profinite completions of Burnside-type quotients of surface groups"
}
| null | null | null | null | true | null |
19223
| null |
Default
| null | null |
null |
{
"abstract": " In this work, we propose an infinite restricted Boltzmann machine~(RBM),\nwhose maximum likelihood estimation~(MLE) corresponds to a constrained convex\noptimization. We consider the Frank-Wolfe algorithm to solve the program, which\nprovides a sparse solution that can be interpreted as inserting a hidden unit\nat each iteration, so that the optimization process takes the form of a\nsequence of finite models of increasing complexity. As a side benefit, this can\nbe used to easily and efficiently identify an appropriate number of hidden\nunits during the optimization. The resulting model can also be used as an\ninitialization for typical state-of-the-art RBM training algorithms such as\ncontrastive divergence, leading to models with consistently higher test\nlikelihood than random initialization.\n",
"title": "Learning Infinite RBMs with Frank-Wolfe"
}
| null | null | null | null | true | null |
19224
| null |
Default
| null | null |
null |
{
"abstract": " We present the discovery of four low-mass ($M<0.6$ $M_\\odot$) eclipsing\nbinary (EB) systems in the sub-Gyr old Praesepe open cluster using Kepler/K2\ntime-series photometry and Keck/HIRES spectroscopy. We present a new Gaussian\nprocess eclipsing binary model, GP-EBOP, as well as a method of simultaneously\ndetermining effective temperatures and distances for EBs. Three of the reported\nsystems (AD 3814, AD 2615 and AD 1508) are detached and double-lined, and\nprecise solutions are presented for the first two. We determine masses and\nradii to 1-3% precision for AD 3814 and to 5-6% for AD 2615. Together with\neffective temperatures determined to $\\sim$50 K precision, we test the PARSEC\nv1.2 and BHAC15 stellar evolution models. Our EB parameters are more consistent\nwith the PARSEC models, primarily because the BHAC15 temperature scale is\nhotter than our data over the mid M-dwarf mass range probed. Both ADs 3814 and\n2615, which have orbital periods of 6.0 and 11.6 days, are circularized but not\nsynchronized. This suggests that either synchronization proceeds more slowly in\nfully convective stars than the theory of equilibrium tides predicts or\nmagnetic braking is currently playing a more important role than tidal forces\nin the spin evolution of these binaries. The fourth system (AD 3116) comprises\na brown dwarf transiting a mid M-dwarf, which is the first such system\ndiscovered in a sub-Gyr open cluster. Finally, these new discoveries increase\nthe number of characterized EBs in sub-Gyr open clusters by 20% (40%) below\n$M<1.5$ $M_{\\odot}$ ($M<0.6$ $M_{\\odot}$).\n",
"title": "New low-mass eclipsing binary systems in Praesepe discovered by K2"
}
| null | null |
[
"Physics"
] | null | true | null |
19225
| null |
Validated
| null | null |
null |
{
"abstract": " Perovskite solar cells with record power conversion efficiency are fabricated\nby alloying both hybrid and fully inorganic compounds. While the basic\nelectronic properties of the hybrid perovskites are now well understood, key\nelectronic parameters for solar cell performance, such as the exciton binding\nenergy of fully inorganic perovskites, are still unknown. By performing magneto\ntransmission measurements, we determine with high accuracy the exciton binding\nenergy and reduced mass of fully inorganic CsPbX$_3$ perovskites (X=I, Br, and\nan alloy of these). The well behaved (continuous) evolution of the band gap\nwith temperature in the range $4-270$\\,K suggests that fully inorganic\nperovskites do not undergo structural phase transitions like their hybrid\ncounterparts. The experimentally determined dielectric constants indicate that\nat low temperature, when the motion of the organic cation is frozen, the\ndielectric screening mechanism is essentially the same both for hybrid and\ninorganic perovskites, and is dominated by the relative motion of atoms within\nthe lead-halide cage.\n",
"title": "The impact of the halide cage on the electronic properties of fully inorganic caesium lead halide perovskites"
}
| null | null | null | null | true | null |
19226
| null |
Default
| null | null |
null |
{
"abstract": " Giant radio galaxies (GRGs) are one of the largest astrophysical sources in\nthe Universe with an overall projected linear size of ~0.7 Mpc or more. Last\nsix decades of radio astronomy research has led to the detection of thousands\nof radio galaxies. But only ~ 300 of them can be classified as GRGs. The\nreasons behind their large size and rarity are unknown. We carried out a\nsystematic search for these radio giants and found a large sample of GRGs. In\nthis paper, we report the discovery of 25 GRGs from NVSS, in the redshift range\n(z) ~ 0.07 to 0.67. Their physical sizes range from ~0.8 Mpc to ~4 Mpc. Eight\nof these GRGs have sizes greater than 2Mpc which is a rarity. In this paper,\nfor the first time, we investigate the mid-IR properties of the optical hosts\nof the GRGs and classify them securely into various AGN types using the WISE\nmid-IR colours. Using radio and IR data, four of the hosts of GRGs were\nobserved to be radio loud quasars that extend up to 2 Mpc in radio size. These\nGRGs missed detection in earlier searches possibly because of their highly\ndiffuse nature, low surface brightness and lack of optical data. The new GRGs\nare a significant addition to the existing sample that will contribute to\nbetter understanding of the physical properties of radio giants.\n",
"title": "Discovery of Giant Radio Galaxies from NVSS: Radio & Infrared Properties"
}
| null | null | null | null | true | null |
19227
| null |
Default
| null | null |
null |
{
"abstract": " We propose a general yet simple theorem describing the convergence of SGD\nunder the arbitrary sampling paradigm. Our theorem describes the convergence of\nan infinite array of variants of SGD, each of which is associated with a\nspecific probability law governing the data selection rule used to form\nmini-batches. This is the first time such an analysis is performed, and most of\nour variants of SGD were never explicitly considered in the literature before.\nOur analysis relies on the recently introduced notion of expected smoothness\nand does not rely on a uniform bound on the variance of the stochastic\ngradients. By specializing our theorem to different mini-batching strategies,\nsuch as sampling with replacement and independent sampling, we derive exact\nexpressions for the stepsize as a function of the mini-batch size. With this we\ncan also determine the mini-batch size that optimizes the total complexity, and\nshow explicitly that as the variance of the stochastic gradient evaluated at\nthe minimum grows, so does the optimal mini-batch size. For zero variance, the\noptimal mini-batch size is one. Moreover, we prove insightful\nstepsize-switching rules which describe when one should switch from a constant\nto a decreasing stepsize regime.\n",
"title": "SGD: General Analysis and Improved Rates"
}
| null | null | null | null | true | null |
19228
| null |
Default
| null | null |
null |
{
"abstract": " We study an extreme scenario in multi-label learning where each training\ninstance is endowed with a single one-bit label out of multiple labels. We\nformulate this problem as a non-trivial special case of one-bit rank-one matrix\nsensing and develop an efficient non-convex algorithm based on alternating\npower iteration. The proposed algorithm is able to recover the underlying\nlow-rank matrix model with linear convergence. For a rank-$k$ model with $d_1$\nfeatures and $d_2$ classes, the proposed algorithm achieves $O(\\epsilon)$\nrecovery error after retrieving $O(k^{1.5}d_1 d_2/\\epsilon)$ one-bit labels\nwithin $O(kd)$ memory. Our bound is nearly optimal in the order of\n$O(1/\\epsilon)$. This significantly improves the state-of-the-art sampling\ncomplexity of one-bit multi-label learning. We perform experiments to verify\nour theory and evaluate the performance of the proposed algorithm.\n",
"title": "Nonconvex One-bit Single-label Multi-label Learning"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
19229
| null |
Validated
| null | null |
null |
{
"abstract": " To explore large-scale population indoor interactions, we analyze 18,715\nusers' WiFi access logs recorded in a Chinese university campus during 3\nmonths, and define two categories of human interactions, the event interaction\n(EI) and the temporal interaction (TI). The EI helps construct a transmission\ngraph, and the TI helps build an interval graph. The dynamics of EIs show that\ntheir active durations are truncated power-law distributed, which is\nindependent on the number of involved individuals. The transmission duration\npresents a truncated power-law behavior at the daily timescale with weekly\nperiodicity. Besides, those `leaf' individuals in the aggregated contact\nnetwork may participate in the `super-connecting cliques' in the aggregated\ntransmission graph. Analyzing the dynamics of the interval graph, we find that\nthe probability distribution of TIs' inter-event duration also displays a\ntruncated power-law pattern at the daily timescale with weekly periodicity,\nwhile the pairwise individuals with burst interactions are prone to randomly\nselect their interactive locations, and those individuals with periodic\ninteractions have preferred interactive locations.\n",
"title": "Two Categories of Indoor Interactive Dynamics of a Large-scale Human Population in a WiFi covered university campus"
}
| null | null |
[
"Computer Science",
"Physics"
] | null | true | null |
19230
| null |
Validated
| null | null |
null |
{
"abstract": " Highly robust and efficient estimators for the generalized linear model with\na dispersion parameter are proposed. The estimators are based on three steps.\nIn the first step the maximum rank correlation estimator is used to\nconsistently estimate the slopes up to a scale factor. In the second step, the\nscale factor, the intercept, and the dispersion parameter are consistently\nestimated using a MT-estimator of a simple regression model. The combined\nestimator is highly robust but inefficient. Then, randomized quantile residuals\nbased on the initial estimators are used to detect outliers to be rejected and\nto define a set S of observations to be retained. Finally, a conditional\nmaximum likelihood (CML) estimator given the observations in S is computed. We\nshow that, under the model, S tends to the complete sample for increasing\nsample size. Therefore, the CML tends to the unconditional maximum likelihood\nestimator. It is therefore highly efficient, while maintaining the high degree\nof robustness of the initial estimator. The case of the negative binomial\nregression model is studied in detail.\n",
"title": "Robust estimators for generalized linear models with a dispersion parameter"
}
| null | null |
[
"Statistics"
] | null | true | null |
19231
| null |
Validated
| null | null |
null |
{
"abstract": " We present lifestate rules--an approach for abstracting event-driven object\nprotocols. Developing applications against event-driven software frameworks is\nnotoriously difficult. One reason why is that to create functioning\napplications, developers must know about and understand the complex protocols\nthat abstract the internal behavior of the framework. Such protocols intertwine\nthe proper registering of callbacks to receive control from the framework with\nappropriate application programming interface (API) calls to delegate back to\nit. Lifestate rules unify lifecycle and typestate constraints in one common\nspecification language. Our primary contribution is a model of event-driven\nsystems from which lifestate rules can be derived. We then apply specification\nmining techniques to learn lifestate specifications for Android framework\ntypes. In the end, our implementation is able to find several rules that\ncharacterize actual behavior of the Android framework.\n",
"title": "Abstracting Event-Driven Systems with Lifestate Rules"
}
| null | null | null | null | true | null |
19232
| null |
Default
| null | null |
null |
{
"abstract": " Restricted Boltzmann Machines (RBMs) are a class of generative neural network\nthat are typically trained to maximize a log-likelihood objective function. We\nargue that likelihood-based training strategies may fail because the objective\ndoes not sufficiently penalize models that place a high probability in regions\nwhere the training data distribution has low probability. To overcome this\nproblem, we introduce Boltzmann Encoded Adversarial Machines (BEAMs). A BEAM is\nan RBM trained against an adversary that uses the hidden layer activations of\nthe RBM to discriminate between the training data and the probability\ndistribution generated by the model. We present experiments demonstrating that\nBEAMs outperform RBMs and GANs on multiple benchmarks.\n",
"title": "Boltzmann Encoded Adversarial Machines"
}
| null | null | null | null | true | null |
19233
| null |
Default
| null | null |
null |
{
"abstract": " It is becoming increasingly clear that complex interactions among genes and\nenvironmental factors play crucial roles in triggering complex diseases. Thus,\nunderstanding such interactions is vital, which is possible only through\nstatistical models that adequately account for such intricate, albeit unknown,\ndependence structures. Bhattacharya & Bhattacharya (2016b) attempt such\nmodeling, relating finite mixtures composed of Dirichlet processes that\nrepresent unknown number of genetic sub-populations through a hierarchical\nmatrix-normal structure that incorporates gene-gene interactions, and possible\nmutations, induced by environmental variables. However, the product dependence\nstructure implied by their matrix-normal model seems to be too simple to be\nappropriate for general complex, realistic situations. In this article, we\npropose and develop a novel nonparametric Bayesian model for case-control\ngenotype data using hierarchies of Dirichlet processes that offers a more\nrealistic and nonparametric dependence structure between the genes, induced by\nthe environmental variables. In this regard, we propose a novel and highly\nparallelisable MCMC algorithm that is rendered quite efficient by the\ncombination of modern parallel computing technology, effective Gibbs sampling\nsteps, retrospective sampling and Transformation based Markov Chain Monte Carlo\n(TMCMC). We use appropriate Bayesian hypothesis testing procedures to detect\nthe roles of genes and environment in case-control studies. We apply our ideas\nto 5 biologically realistic case-control genotype datasets simulated under\ndistinct set-ups, and obtain encouraging results in each case. We finally apply\nour ideas to a real, myocardial infarction dataset, and obtain interesting\nresults on gene-gene and gene-environment interaction, while broadly agreeing\nwith the results reported in the literature.\n",
"title": "A Non-Gaussian, Nonparametric Structure for Gene-Gene and Gene-Environment Interactions in Case-Control Studies Based on Hierarchies of Dirichlet Processes"
}
| null | null | null | null | true | null |
19234
| null |
Default
| null | null |
null |
{
"abstract": " Principal component regression is a linear regression model with principal\ncomponents as regressors. This type of modelling is particularly useful for\nprediction in settings with high-dimensional covariates. Surprisingly, the\nexisting literature treating of Bayesian approaches is relatively sparse. In\nthis paper, we aim at filling some gaps through the following practical\ncontribution: we introduce a Bayesian approach with detailed guidelines for a\nstraightforward implementation. The approach features two characteristics that\nwe believe are important. First, it effectively involves the relevant principal\ncomponents in the prediction process. This is achieved in two steps. The first\none is model selection; the second one is to average out the predictions\nobtained from the selected models according to model averaging mechanisms,\nallowing to account for model uncertainty. The model posterior probabilities\nare required for model selection and model averaging. For this purpose, we\ninclude a procedure leading to an efficient reversible jump algorithm. The\nsecond characteristic of our approach is whole robustness, meaning that the\nimpact of outliers on inference gradually vanishes as they approach plus or\nminus infinity. The conclusions obtained are consequently consistent with the\nmajority of observations (the bulk of the data).\n",
"title": "An Efficient Bayesian Robust Principal Component Regression"
}
| null | null | null | null | true | null |
19235
| null |
Default
| null | null |
null |
{
"abstract": " We propose an efficient meta-algorithm for Bayesian estimation problems that\nis based on low-degree polynomials, semidefinite programming, and tensor\ndecomposition. The algorithm is inspired by recent lower bound constructions\nfor sum-of-squares and related to the method of moments. Our focus is on sample\ncomplexity bounds that are as tight as possible (up to additive lower-order\nterms) and often achieve statistical thresholds or conjectured computational\nthresholds.\nOur algorithm recovers the best known bounds for community detection in the\nsparse stochastic block model, a widely-studied class of estimation problems\nfor community detection in graphs. We obtain the first recovery guarantees for\nthe mixed-membership stochastic block model (Airoldi et el.) in constant\naverage degree graphs---up to what we conjecture to be the computational\nthreshold for this model. We show that our algorithm exhibits a sharp\ncomputational threshold for the stochastic block model with multiple\ncommunities beyond the Kesten--Stigum bound---giving evidence that this task\nmay require exponential time.\nThe basic strategy of our algorithm is strikingly simple: we compute the\nbest-possible low-degree approximation for the moments of the posterior\ndistribution of the parameters and use a robust tensor decomposition algorithm\nto recover the parameters from these approximate posterior moments.\n",
"title": "Bayesian estimation from few samples: community detection and related problems"
}
| null | null | null | null | true | null |
19236
| null |
Default
| null | null |
null |
{
"abstract": " We consider the boundary rigidity problem for asymptotically hyperbolic\nmanifolds. We show injectivity of the X-ray transform in several cases and\nconsider the non-linear inverse problem which consists of recovering a metric\nfrom boundary measurements for the geodesic flow.\n",
"title": "X-ray Transform and Boundary Rigidity for Asymptotically Hyperbolic Manifolds"
}
| null | null | null | null | true | null |
19237
| null |
Default
| null | null |
null |
{
"abstract": " We consider the spontaneous breaking of translational symmetry and identify\nthe associated Goldstone mode -- a longitudinal phonon -- in a holographic\nmodel with Bianchi VII helical symmetry. For the first time in holography, we\nobserve the pinning of this mode after introducing a source for explicit\nbreaking compatible with the helical symmetry of our setup. We study the\ndispersion relation of the resulting pseudo-Goldstone mode, uncovering how its\nspeed and mass gap depend on the amplitude of the source and temperature. In\naddition, we extract the optical conductivity as a function of frequency, which\nreveals a metal-insulator transition as a consequence of the pinning.\n",
"title": "Pinning of longitudinal phonons in holographic spontaneous helices"
}
| null | null | null | null | true | null |
19238
| null |
Default
| null | null |
null |
{
"abstract": " The method of brackets is an efficient method for the evaluation of a large\nclass of definite integrals on the half-line. It is based on a small collection\nof rules, some of which are heuristic. The extension discussed here is based on\nthe concepts of null and divergent series. These are formal representations of\nfunctions, whose coefficients $a_{n}$ have meromorphic representations for $n\n\\in \\mathbb{C}$, but might vanish or blow up when $n \\in \\mathbb{N}$. These\nideas are illustrated with the evaluation of a variety of entries from the\nclassical table of integrals by Gradshteyn and Ryzhik.\n",
"title": "An Extension of the Method of Brackets. Part 1"
}
| null | null | null | null | true | null |
19239
| null |
Default
| null | null |
null |
{
"abstract": " The R package optimParallel provides a parallel version of the gradient-based\noptimization methods of optim(). The main function of the package is\noptimParallel(), which has the same usage and output as optim(). Using\noptimParallel() can significantly reduce optimization times. We introduce the R\npackage and illustrate its implementation, which takes advantage of the lexical\nscoping mechanism of R.\n",
"title": "optimParallel: an R Package Providing Parallel Versions of the Gradient-Based Optimization Methods of optim()"
}
| null | null | null | null | true | null |
19240
| null |
Default
| null | null |
null |
{
"abstract": " Some key results obtained in joint research projects with Alex Müller are\nsummarized, concentrating on the invention of the barocaloric effect and its\napplication for cooling as well as on important findings in the field of\nhigh-temperature superconductivity resulting from neutron scattering\nexperiments.\n",
"title": "The barocaloric effect: A Spin-off of the Discovery of High-Temperature Superconductivity"
}
| null | null |
[
"Physics"
] | null | true | null |
19241
| null |
Validated
| null | null |
null |
{
"abstract": " We show that Rashba spin-orbit coupling at the interface between a\nsuperconductor and a ferromagnet should produce a spontaneous current in the\natomic thickness region near the interface. This current is counter-balanced by\nthe superconducting screening current flowing in the region of the width of the\nLondon penetration depth near the interface. Such current carrying state\ncreates a magnetic field near the superconductor surface, generates a stray\nmagnetic field outside the sample edges, changes the slope of the temperature\ndependence of the critical field $H_{c3}$ and may generate the spontaneous\nAbrikosov vortices near the interface.\n",
"title": "Spontaneous currents in superconducting systems with strong spin-orbit coupling"
}
| null | null |
[
"Physics"
] | null | true | null |
19242
| null |
Validated
| null | null |
null |
{
"abstract": " Fraud has severely detrimental impacts on the business of social networks and\nother online applications. A user can become a fake celebrity by purchasing\n\"zombie followers\" on Twitter. A merchant can boost his reputation through fake\nreviews on Amazon. This phenomenon also conspicuously exists on Facebook, Yelp\nand TripAdvisor, etc. In all the cases, fraudsters try to manipulate the\nplatform's ranking mechanism by faking interactions between the fake accounts\nthey control and the target customers.\n",
"title": "Catching Loosely Synchronized Behavior in Face of Camouflage"
}
| null | null | null | null | true | null |
19243
| null |
Default
| null | null |
null |
{
"abstract": " Systems which can spontaneously reveal periodic evolution are dubbed time\ncrystals. This is in analogy with space crystals that display periodic behavior\nin configuration space. While space crystals are modelled with the help of\nspace periodic potentials, crystalline phenomena in time can be modelled by\nperiodically driven systems. Disorder in the periodic driving can lead to\nAnderson localization in time: the probability for detecting a system at a\nfixed point of configuration space becomes exponentially localized around a\ncertain moment in time. We here show that a three-dimensional system exposed to\na properly disordered pseudo-periodic driving may display a\nlocalized-delocalized Anderson transition in the time domain, in strong analogy\nwith the usual three-dimensional Anderson transition in disordered systems.\nSuch a transition could be experimentally observed with ultra-cold atomic\ngases.\n",
"title": "Three-dimensional localized-delocalized Anderson transition in the time domain"
}
| null | null | null | null | true | null |
19244
| null |
Default
| null | null |
null |
{
"abstract": " The analysis of the demographic transition of the past century and a half,\nusing both empirical data and mathematical models, has rendered a wealth of\nwell-established facts, including the dramatic increases in life expectancy.\nDespite these insights, such analyses have also occasionally triggered debates\nwhich spill over many disciplines, from genetics, to biology, or demography.\nPerhaps the hottest discussion is happening around the question of maximum\nhuman lifespan, which --besides its fascinating historical and philosophical\ninterest-- poses urgent pragmatic warnings on a number of issues in public and\nprivate decision-making. In this paper, we add to the controversy some results\nwhich, based on purely statistical grounds, suggest that the maximum human\nlifespan is not fixed, or has not reached yet a plateau. Quite the contrary,\nanalysis on reliable data for over 150 years in more than 20 industrialized\ncountries point at a sustained increase in the maximum age at death.\nFurthermore, were this trend to continue, a limitless lifespan could be\nachieved by 2102. Finally, we quantify the dependence of increases in the\nmaximum lifespan on socio-economic factors. Our analysis indicates that in some\ncountries the observed rising patterns can only be sustained by progressively\nlarger increases in GDP, setting the problem of longevity in a context of\ndiminishing returns.\n",
"title": "Socio-economic constraints to maximum human lifespan"
}
| null | null | null | null | true | null |
19245
| null |
Default
| null | null |
null |
{
"abstract": " Cyber-Physical Systems (CPS) revolutionize various application domains with\nintegration and interoperability of networking, computing systems, and\nmechanical devices. Due to its scale and variety, CPS faces a number of\nchallenges and opens up a few research questions in terms of management,\nfault-tolerance, and scalability. We propose a software-defined approach\ninspired by Software-Defined Networking (SDN), to address the challenges for a\nwider CPS adoption. We thus design a middleware architecture for the correct\nand resilient operation of CPS, to manage and coordinate the interacting\ndevices centrally in the cyberspace whilst not sacrificing the functionality\nand performance benefits inherent to a distributed execution.\n",
"title": "SD-CPS: Taming the Challenges of Cyber-Physical Systems with a Software-Defined Approach"
}
| null | null |
[
"Computer Science"
] | null | true | null |
19246
| null |
Validated
| null | null |
null |
{
"abstract": " This paper gives a self-contained group-theoretic proof of a dual version of\na theorem of Ore on distributive intervals of finite groups. We deduce a bridge\nbetween combinatorics and representations in finite group theory.\n",
"title": "Dual Ore's theorem on distributive intervals of finite groups"
}
| null | null | null | null | true | null |
19247
| null |
Default
| null | null |
null |
{
"abstract": " A measurable function $\\mu$ on the unit disk $\\mathbb{D}$ of the complex\nplane with $\\|\\mu\\|_\\infty<1$ is sometimes called a Beltrami coefficient. We\nsay that $\\mu$ is trivial if it is the complex dilatation $f_{\\bar z}/f_z$ of a\nquasiconformal automorphism $f$ of $\\mathbb{D}$ satisfying the trivial boundary\ncondition $f(z)=z,~|z|=1.$ Since it is not easy to solve the Beltrami equation\nexplicitly, to detect triviality of a given Beltrami coefficient is a hard\nproblem, in general. In the present article, we offer a sufficient condition\nfor a Beltrami coefficient to be trivial. Our proof is based on Betker's\ntheorem on Löwner chains.\n",
"title": "A construction of trivial Beltrami coefficients"
}
| null | null |
[
"Mathematics"
] | null | true | null |
19248
| null |
Validated
| null | null |
null |
{
"abstract": " We give a detailed account of the so-called \"universal construction\" that\naims to extend invariants of closed manifolds, possibly with additional\nstructure, to topological field theories and show that it amounts to a\ngeneralization of the GNS construction. We apply this construction to an\ninvariant defined in terms of the groupoid cardinality of groupoids of bundles\nto recover Dijkgraaf-Witten theories, including the vector spaces obtained as a\nlinearization of spaces of principal bundles.\n",
"title": "A GNS construction of three-dimensional abelian Dijkgraaf-Witten theories"
}
| null | null |
[
"Mathematics"
] | null | true | null |
19249
| null |
Validated
| null | null |
null |
{
"abstract": " An asymmetric resonant cavity can be used to form a path that is much longer\nthan the cavity size. We demonstrate this capability for a deformed microdisk\nequipped with two linear waveguides, by constructing a multiply reflected\nperiodic orbit that is confined by total internal reflection within the\ndeformed microdisk and outcoupled by the two linear waveguides. Resonant mode\nanalysis reveals that the modes corresponding to the periodic orbit are\ncharacterized by high quality factors. From measured spectral and far-field\ndata, we confirm that the fabricated devices can form a path about 9.3 times\nlonger than the average diameter of the deformed microdisk.\n",
"title": "Long-path formation in a deformed microdisk laser"
}
| null | null | null | null | true | null |
19250
| null |
Default
| null | null |
null |
{
"abstract": " The 3.6 and 4.5 micron characteristics of AGB variables in the LMC and IC1613\nare discussed. For C-rich Mira variables there is a very clear\nperiod-luminosity-colour relation, where the [3.6]-[4.5] colour is associated\nwith the amount of circumstellar material and correlated with the pulsation\namplitude. The [4.5] period-luminosity relation for dusty stars is\napproximately one mag brighter than for their naked counterparts with\ncomparable periods.\n",
"title": "Spitzer Observations of Large Amplitude Variables in the LMC and IC 1613"
}
| null | null | null | null | true | null |
19251
| null |
Default
| null | null |
null |
{
"abstract": " Effective teams are crucial for organisations, especially in environments\nthat require teams to be constantly created and dismantled, such as software\ndevelopment, scientific experiments, crowd-sourcing, or the classroom. Key\nfactors influencing team performance are competences and personality of team\nmembers. Hence, we present a computational model to compose proficient and\ncongenial teams based on individuals' personalities and their competences to\nperform tasks of different nature. With this purpose, we extend Wilde's\npost-Jungian method for team composition, which solely employs individuals'\npersonalities. The aim of this study is to create a model to partition agents\ninto teams that are balanced in competences, personality and gender. Finally,\nwe present some preliminary empirical results that we obtained when analysing\nstudent performance. Results show the benefits of a more informed team\ncomposition that exploits individuals' competences besides information about\ntheir personalities.\n",
"title": "Synergistic Team Composition"
}
| null | null | null | null | true | null |
19252
| null |
Default
| null | null |
null |
{
"abstract": " Given a semigroup S with zero, which is left-cancellative in the sense that\nst=sr \\neq 0 implies that t=r, we construct an inverse semigroup called the\ninverse hull of S, denoted H(S). When S admits least common multiples, in a\nprecise sense defined below, we study the idempotent semilattice of H(S), with\na focus on its spectrum. When S arises as the language semigroup for a subsift\nX on a finite alphabet, we discuss the relationship between H(S) and several\nC*-algebras associated to X appearing in the literature.\n",
"title": "The inverse hull of 0-left cancellative semigroups"
}
| null | null | null | null | true | null |
19253
| null |
Default
| null | null |
null |
{
"abstract": " A generalization of classical determinant inequalities like Hadamard's\ninequality and Fischer's inequality is studied. For a version of the\ninequalities originally proved by Arveson for positive operators in von Neumann\nalgebras with a tracial state, we give a different proof. We also improve and\ngeneralize to the setting of finite von Neumann algebras, some `Fischer-type'\ninequalities by Matic for determinants of perturbed positive-definite matrices.\nIn the process, a conceptual framework is established for viewing these\ninequalities as manifestations of Jensen's inequality in conjunction with the\ntheory of operator monotone and operator convex functions on $[0,\\infty)$. We\nplace emphasis on documenting necessary and sufficient conditions for equality\nto hold.\n",
"title": "The Hadamard Determinant Inequality - Extensions to Operators on a Hilbert Space"
}
| null | null | null | null | true | null |
19254
| null |
Default
| null | null |
null |
{
"abstract": " We propose a new reinforcement learning algorithm for partially observable\nMarkov decision processes (POMDP) based on spectral decomposition methods.\nWhile spectral methods have been previously employed for consistent learning of\n(passive) latent variable models such as hidden Markov models, POMDPs are more\nchallenging since the learner interacts with the environment and possibly\nchanges the future observations in the process. We devise a learning algorithm\nrunning through epochs, in each epoch we employ spectral techniques to learn\nthe POMDP parameters from a trajectory generated by a fixed policy. At the end\nof the epoch, an optimization oracle returns the optimal memoryless planning\npolicy which maximizes the expected reward based on the estimated POMDP model.\nWe prove an order-optimal regret bound with respect to the optimal memoryless\npolicy and efficient scaling with respect to the dimensionality of observation\nand action spaces.\n",
"title": "Experimental results : Reinforcement Learning of POMDPs using Spectral Methods"
}
| null | null | null | null | true | null |
19255
| null |
Default
| null | null |
null |
{
"abstract": " We introduce the notion of ST-pairs of triangulated subcategories, a\nprototypical example of which is the pair of the bound homotopy category and\nthe bound derived category of a finite-dimensional algebra. For an ST-pair\n$(\\C,\\D)$, we construct an order-preserving map from silting objects in $\\C$ to\nbounded $t$-structures on $\\D$ and show that the map is bijective if and only\nif $\\C$ is silting-discrete if and only if $\\D$ is $t$-discrete. Based on a\nwork of Qiu and Woolf, the above result is applied to show that if $\\C$ is\nsilting-discrete then the stability space of $\\D$ is contractible. This is used\nto obtain the contractibility of the stability spaces of some Calabi--Yau\ntriangulated categories associated to Dynkin quivers.\n",
"title": "Discreteness of silting objects and t-structures in triangulated categories"
}
| null | null | null | null | true | null |
19256
| null |
Default
| null | null |
null |
{
"abstract": " We introduce features for massive data streams. These stream features can be\nthought of as \"ordered moments\" and generalize stream sketches from \"moments of\norder one\" to \"ordered moments of arbitrary order\". In analogy to classic\nmoments, they have theoretical guarantees such as universality that are\nimportant for learning algorithms.\n",
"title": "Sketching the order of events"
}
| null | null | null | null | true | null |
19257
| null |
Default
| null | null |
null |
{
"abstract": " We study the problem of learning to rank from multiple sources. Though\nmulti-view learning and learning to rank have been studied extensively leading\nto a wide range of applications, multi-view learning to rank as a synergy of\nboth topics has received little attention. The aim of the paper is to propose a\ncomposite ranking method while keeping a close correlation with the individual\nrankings simultaneously. We propose a multi-objective solution to ranking by\ncapturing the information of the feature mapping from both within each view as\nwell as across views using autoencoder-like networks. Moreover, a novel\nend-to-end solution is introduced to enhance the joint ranking with minimum\nview-specific ranking loss, so that we can achieve the maximum global view\nagreements within a single optimization process. The proposed method is\nvalidated on a wide variety of ranking problems, including university ranking,\nmulti-view lingual text ranking and image data ranking, providing superior\nresults.\n",
"title": "Deep Multi-view Learning to Rank"
}
| null | null | null | null | true | null |
19258
| null |
Default
| null | null |
null |
{
"abstract": " In order for a robot to be a generalist that can perform a wide range of\njobs, it must be able to acquire a wide variety of skills quickly and\nefficiently in complex unstructured environments. High-capacity models such as\ndeep neural networks can enable a robot to represent complex skills, but\nlearning each skill from scratch then becomes infeasible. In this work, we\npresent a meta-imitation learning method that enables a robot to learn how to\nlearn more efficiently, allowing it to acquire new skills from just a single\ndemonstration. Unlike prior methods for one-shot imitation, our method can\nscale to raw pixel inputs and requires data from significantly fewer prior\ntasks for effective learning of new skills. Our experiments on both simulated\nand real robot platforms demonstrate the ability to learn new tasks,\nend-to-end, from a single visual demonstration.\n",
"title": "One-Shot Visual Imitation Learning via Meta-Learning"
}
| null | null | null | null | true | null |
19259
| null |
Default
| null | null |
null |
{
"abstract": " We give an explicit formula for the second variation of the logarithm of the\nSelberg zeta function, $Z(s)$, on Teichmüller space. We then use this formula\nto determine the asymptotic behavior as $s \\to \\infty$ of the second variation.\nAs a consequence, we determine the signature of the Hessian of $\\log Z(s)$ for\nsufficiently large $s$. As a further consequence, the asymptotic behavior of\nthe second variation of $\\log Z(s)$ shows that the Ricci curvature of the Hodge\nbundle $H^0(\\mathcal K^m_t)\\mapsto t$ over Teichmüller space agrees with the\nQuillen curvature up to a term of exponential decay, $O(s^2 e^{-l_0 s}),$ where\n$l_0$ is the length of the shortest closed hyperbolic geodesic.\n",
"title": "Second variation of Selberg zeta functions and curvature asymptotics"
}
| null | null | null | null | true | null |
19260
| null |
Default
| null | null |
null |
{
"abstract": " The idea that black hole spin is instrumental in the generation of powerful\njets in active galactic nuclei and X-ray binaries is arguably the most\ncontentious claim in black hole astrophysics. Because jets are thought to\noriginate in the context of electromagnetism, and the modeling of Maxwell\nfields in curved spacetime around black holes is challenging, various\napproximations are made in numerical simulations that fall under the guise of\n'ideal magnetohydrodynamics'. But the simplifications of this framework may\nstruggle to capture relevant details of real astrophysical environments near\nblack holes. In this work, we highlight tension between analytic and numerical\nresults, specifically between the analytically derived conserved Noether\ncurrents for rotating black hole spacetimes and the results of general\nrelativistic numerical simulations (GRMHD). While we cannot definitively\nattribute the issue to any specific approximation used in the numerical\nschemes, there seem to be natural candidates, which we explore. GRMHD\nnotwithstanding, if electromagnetic fields around rotating black holes are\nbrought to the hole by accretion, we show from first principles that prograde\naccreting disks likely experience weaker large-scale black hole-threading\nfields, implying weaker jets than in retrograde configurations.\n",
"title": "Magnetic Fields Threading Black Holes: restrictions from general relativity and implications for astrophysical black holes"
}
| null | null | null | null | true | null |
19261
| null |
Default
| null | null |
null |
{
"abstract": " The distance covariance of two random vectors is a measure of their\ndependence. The empirical distance covariance and correlation can be used as\nstatistical tools for testing whether two random vectors are independent. We\npropose an analogs of the distance covariance for two stochastic processes\ndefined on some interval. Their empirical analogs can be used to test the\nindependence of two processes.\n",
"title": "Distance covariance for stochastic processes"
}
| null | null | null | null | true | null |
19262
| null |
Default
| null | null |
null |
{
"abstract": " We investigate nonparametric regression methods based on statistical depth\nfunctions. These nonparametric regression procedures can be used in situations,\nwhere the response is multivariate and the covariate is a random element in a\nmetric space. This includes regression with functional covariate as a special\ncase. Our objective is to study different features of the conditional\ndistribution of the response given the covariate. We construct measures of the\ncenter and the spread of the conditional distribution using depth based\nnonparametric regression procedures. We establish the asymptotic consistency of\nthose measures and develop a test for heteroscedasticity based on the measure\nof conditional spread. The usefulness of the methodology is demonstrated in\nsome real datasets. In one dataset consisting of Italian household expenditure\ndata for the period 1973 to 1992, we regress the expenditure for different\nitems on their prices. In another dataset, our responses are the nutritional\ncontents of different meat samples measured by their protein, fat and moisture\ncontents, and the functional covariate is the absorbance spectra of the meat\nsamples.\n",
"title": "On Nonparametric Regression using Data Depth"
}
| null | null | null | null | true | null |
19263
| null |
Default
| null | null |
null |
{
"abstract": " A number of optimal decision problems with uncertainty can be formulated into\na stochastic optimal control framework. The Least-Squares Monte Carlo (LSMC)\nalgorithm is a popular numerical method to approach solutions of such\nstochastic control problems as analytical solutions are not tractable in\ngeneral. This paper generalizes the LSMC algorithm proposed in Shen and Weng\n(2017) to solve a wide class of stochastic optimal control models. Our\nalgorithm has three pillars: a construction of auxiliary stochastic control\nmodel, an artificial simulation of the post-action value of state process, and\na shape-preserving sieve estimation method which equip the algorithm with a\nnumber of merits including bypassing forward simulation and control\nrandomization, evading extrapolating the value function, and alleviating\ncomputational burden of the tuning parameter selection. The efficacy of the\nalgorithm is corroborated by an application to pricing equity-linked insurance\nproducts.\n",
"title": "A Backward Simulation Method for Stochastic Optimal Control Problems"
}
| null | null | null | null | true | null |
19264
| null |
Default
| null | null |
null |
{
"abstract": " The iterated posterior linearization filter (IPLF) is an algorithm for\nBayesian state estimation that performs the measurement update using iterative\nstatistical regression. The main result behind IPLF is that the posterior\napproximation is more accurate when the statistical regression of measurement\nfunction is done in the posterior instead of the prior as is done in\nnon-iterative Kalman filter extensions. In IPLF, each iteration in principle\ngives a better posterior estimate to obtain a better statistical regression and\nmore accurate posterior estimate in the next iteration. However, IPLF may\ndiverge. IPLF's fixed- points are not described as solutions to an optimization\nproblem, which makes it challenging to improve its convergence properties. In\nthis letter, we introduce a double-loop version of IPLF, where the inner loop\ncomputes the posterior mean using an optimization algorithm. Simulation results\nare presented to show that the proposed algorithm has better convergence than\nIPLF and its accuracy is similar to or better than other state-of-the-art\nalgorithms.\n",
"title": "Damped Posterior Linearization Filter"
}
| null | null | null | null | true | null |
19265
| null |
Default
| null | null |
null |
{
"abstract": " We consider a system where randomly generated updates are to be transmitted\nto a monitor, but only a single update can be in the transmission service at a\ntime. Therefore, the source has to prioritize between the two possible\ntransmission policies: preempting the current update or discarding the new one.\nWe consider Poisson arrivals and general service time, and refer to this system\nas the M/G/1/1 queue. We start by studying the average status update age and\nthe optimal update arrival rate for these two schemes under general service\ntime distribution. We then apply these results on two practical scenarios in\nwhich updates are sent through an erasure channel using (a) an infinite\nincremental redundancy (IIR) HARQ system and (b) a fixed redundancy (FR) HARQ\nsystem. We show that in both schemes the best strategy would be not to preempt.\nMoreover, we also prove that, from an age point of view, IIR is better than FR.\n",
"title": "Status updates through M/G/1/1 queues with HARQ"
}
| null | null | null | null | true | null |
19266
| null |
Default
| null | null |
null |
{
"abstract": " Most analyses of randomised trials with incomplete outcomes make untestable\nassumptions and should therefore be subjected to sensitivity analyses. However,\nmethods for sensitivity analyses are not widely used. We propose a mean score\napproach for exploring global sensitivity to departures from missing at random\nor other assumptions about incomplete outcome data in a randomised trial. We\nassume a single outcome analysed under a generalised linear model. One or more\nsensitivity parameters, specified by the user, measure the degree of departure\nfrom missing at random in a pattern mixture model. Advantages of our method are\nthat its sensitivity parameters are relatively easy to interpret and so can be\nelicited from subject matter experts; it is fast and non-stochastic; and its\npoint estimate, standard error and confidence interval agree perfectly with\nstandard methods when particular values of the sensitivity parameters make\nthose standard methods appropriate. We illustrate the method using data from a\nmental health trial.\n",
"title": "A mean score method for sensitivity analysis to departures from the missing at random assumption in randomised trials"
}
| null | null | null | null | true | null |
19267
| null |
Default
| null | null |
null |
{
"abstract": " Let R be a family of n axis-parallel rectangles with packing number p-1,\nmeaning that among any p of the rectangles, there are two with a non-empty\nintersection. We show that the union complexity of R is at most O(n+p^2), and\nthat the (<=k)-level complexity of R is at most O(kn+k^2p^2). Both upper bounds\nare tight.\n",
"title": "On the union complexity of families of axis-parallel rectangles with a low packing number"
}
| null | null | null | null | true | null |
19268
| null |
Default
| null | null |
null |
{
"abstract": " We present a new method to derive oxygen and carbon abundances using the\nultraviolet (UV) lines emitted by the gas-phase ionised by massive stars. The\nmethod is based on the comparison of the nebular emission-line ratios with\nthose predicted by a large grid of photo-ionisation models. Given the large\ndispersion in the O/H - C/O plane, our method firstly fixes C/O using ratios of\nappropriate emission lines and, in a second step, calculates O/H and the\nionisation parameter from carbon lines in the UV. We find abundances totally\nconsistent with those provided by the direct method when we apply this method\nto a sample of objects with an empirical determination of the electron\ntemperature using optical emission lines. The proposed methodology appears as a\npowerful tool for systematic studies of nebular abundances in star-forming\ngalaxies at high redshift.\n",
"title": "Using photo-ionisation models to derive carbon and oxygen gas-phase abundances in the rest UV"
}
| null | null |
[
"Physics"
] | null | true | null |
19269
| null |
Validated
| null | null |
null |
{
"abstract": " We study the extent to which we can infer users' geographical locations from\nsocial media. Location inference from social media can benefit many\napplications, such as disaster management, targeted advertising, and news\ncontent tailoring. In recent years, a number of algorithms have been proposed\nfor identifying user locations on social media platforms such as Twitter and\nFacebook from message contents, friend networks, and interactions between\nusers. In this paper, we propose a novel probabilistic model based on factor\ngraphs for location inference that offers several unique advantages for this\ntask. First, the model generalizes previous methods by incorporating content,\nnetwork, and deep features learned from social context. The model is also\nflexible enough to support both supervised learning and semi-supervised\nlearning. Second, we explore several learning algorithms for the proposed\nmodel, and present a Two-chain Metropolis-Hastings (MH+) algorithm, which\nimproves the inference accuracy. Third, we validate the proposed model on three\ndifferent genres of data - Twitter, Weibo, and Facebook - and demonstrate that\nthe proposed model can substantially improve the inference accuracy (+3.3-18.5%\nby F1-score) over that of several state-of-the-art methods.\n",
"title": "A Probabilistic Framework for Location Inference from Social Media"
}
| null | null | null | null | true | null |
19270
| null |
Default
| null | null |
null |
{
"abstract": " The ABALONE Photosensor Technology (U.S. Patent 9064678 2015) has the\ncapability of supplanting the expensive 80 year old Photomultiplier Tube (PMT)\nmanufacture by providing a modern and cost effective alternative product. An\nABALONE Photosensor comprises only three monolithic glass components, sealed\ntogether by our new thin film adhesive. In 2013, we left one of the early\nABALONE Photosensor prototypes intact for continuous stress testing, and here\nwe report its long term vacuum integrity. The exceptionally low ion\nafterpulsing rate (approximately two orders of magnitude lower than in PMTs)\nhas been constantly improving. We explain the physical and technological\nreasons for this achievement. Due to the cost-effectiveness and the specific\ncombination of features, including low level of radioactivity, integration into\nlarge-area panels, and robustness, this technology can open new horizons in the\nfields of fundamental physics, functional medical imaging, and nuclear\nsecurity.\n",
"title": "The Novel ABALONE Photosensor Technology: 4-Year Long Tests of Vacuum Integrity, Internal Pumping and Afterpulsing"
}
| null | null | null | null | true | null |
19271
| null |
Default
| null | null |
null |
{
"abstract": " In large scale coverage operations, such as marine exploration or aerial\nmonitoring, single robot approaches are not ideal, as they may take too long to\ncover a large area. In such scenarios, multi-robot approaches are preferable.\nFurthermore, several real world vehicles are non-holonomic, but can be modeled\nusing Dubins vehicle kinematics. This paper focuses on environmental monitoring\nof aquatic environments using Autonomous Surface Vehicles (ASVs). In\nparticular, we propose a novel approach for solving the problem of complete\ncoverage of a known environment by a multi-robot team consisting of Dubins\nvehicles. It is worth noting that both multi-robot coverage and Dubins vehicle\ncoverage are NP-complete problems. As such, we present two heuristics methods\nbased on a variant of the traveling salesman problem -- k-TSP -- formulation\nand clustering algorithms that efficiently solve the problem. The proposed\nmethods are tested both in simulations to assess their scalability and with a\nteam of ASVs operating on a lake to ensure their applicability in real world.\n",
"title": "Multi-robot Dubins Coverage with Autonomous Surface Vehicles"
}
| null | null | null | null | true | null |
19272
| null |
Default
| null | null |
null |
{
"abstract": " We study fairness in collaborative-filtering recommender systems, which are\nsensitive to discrimination that exists in historical data. Biased data can\nlead collaborative filtering methods to make unfair predictions against\nminority groups of users. We identify the insufficiency of existing fairness\nmetrics and propose four new metrics that address different forms of\nunfairness. These fairness metrics can be optimized by adding fairness terms to\nthe learning objective. Experiments on synthetic and real data show that our\nnew metrics can better measure fairness than the baseline, and that the\nfairness objectives effectively help reduce unfairness.\n",
"title": "New Fairness Metrics for Recommendation that Embrace Differences"
}
| null | null |
[
"Computer Science"
] | null | true | null |
19273
| null |
Validated
| null | null |
null |
{
"abstract": " Graph inference methods have recently attracted a great interest from the\nscientific community, due to the large value they bring in data interpretation\nand analysis. However, most of the available state-of-the-art methods focus on\nscenarios where all available data can be explained through the same graph, or\ngroups corresponding to each graph are known a priori. In this paper, we argue\nthat this is not always realistic and we introduce a generative model for mixed\nsignals following a heat diffusion process on multiple graphs. We propose an\nexpectation-maximisation algorithm that can successfully separate signals into\ncorresponding groups, and infer multiple graphs that govern their behaviour. We\ndemonstrate the benefits of our method on both synthetic and real data.\n",
"title": "Graph heat mixture model learning"
}
| null | null | null | null | true | null |
19274
| null |
Default
| null | null |
null |
{
"abstract": " This paper presents a provably correct method for robot navigation in 2D\nenvironments cluttered with familiar but unexpected non-convex, star-shaped\nobstacles as well as completely unknown, convex obstacles. We presuppose a\nlimited range onboard sensor, capable of recognizing, localizing and\n(leveraging ideas from constructive solid geometry) generating online from its\ncatalogue of the familiar, non-convex shapes an implicit representation of each\none. These representations underlie an online change of coordinates to a\ncompletely convex model planning space wherein a previously developed online\nconstruction yields a provably correct reactive controller that is pulled back\nto the physically sensed representation to generate the actual robot commands.\nWe extend the construction to differential drive robots, and suggest the\nempirical utility of the proposed control architecture using both formal proofs\nand numerical simulations.\n",
"title": "Technical Report: Reactive Navigation in Partially Known Non-Convex Environments"
}
| null | null | null | null | true | null |
19275
| null |
Default
| null | null |
null |
{
"abstract": " The resolvent Krylov subspace method builds approximations to operator\nfunctions $f(A)$ times a vector $v$. For the semigroup and related operator\nfunctions, this method is proved to possess the favorable property that the\nconvergence is automatically faster when the vector $v$ is smoother. The user\nof the method does not need to know the presented theory and alterations of the\nmethod are not necessary in order to adapt to the (possibly unknown) smoothness\nof $v$. The findings are illustrated by numerical experiments.\n",
"title": "Automatic smoothness detection of the resolvent Krylov subspace method for the approximation of $C_0$-semigroups"
}
| null | null | null | null | true | null |
19276
| null |
Default
| null | null |
null |
{
"abstract": " The best known manifestation of the Fermi-Dirac statistics is the Pauli\nexclusion principle: no two identical fermions can occupy the same one-particle\nstate. This principle enforces high order correlations in systems of many\nidentical fermions and is responsible for a particular geometric arrangement of\ntrapped particles even when all mutual interactions are absent [1]. These\ngeometric structures, called Pauli crystals, are predicted for a system of $N$\nidentical atoms trapped in a harmonic potential. They emerge as the most\nfrequent configurations in a collection of single-shot pictures of the system.\nHere we study how fragile Pauli crystals are when realistic experimental\nlimitations are taken into account. The influence of the number of single-shots\npictures available to analysis, thermal fluctuations and finite efficiency of\ndetection are considered. The role of these sources of noise on the possibility\nof experimental observation of Pauli crystals is shown and conditions necessary\nfor the detection of the geometrical arrangements of particles are identified.\n",
"title": "On the observability of Pauli crystals"
}
| null | null | null | null | true | null |
19277
| null |
Default
| null | null |
null |
{
"abstract": " This paper proves the approximate intermediate value theorem, constructively\nand from notably weak hypotheses: from pointwise rather than uniform\ncontinuity, without assuming that reals are presented with rational\napproximants, and without using countable choice. The theorem is that if a\npointwise continuous function has both a negative and a positive value, then it\nhas values arbitrarily close to 0. The proof builds on the usual classical\nproof by bisection, which repeatedly selects the left or right half of an\ninterval; the algorithm here selects an interval of half the size in a\ncontinuous way, interpolating between those two possibilities.\n",
"title": "Interpolating Between Choices for the Approximate Intermediate Value Theorem"
}
| null | null |
[
"Computer Science",
"Mathematics"
] | null | true | null |
19278
| null |
Validated
| null | null |
null |
{
"abstract": " The ultraviolet (UV) light from a host star influences a planet's atmospheric\nphotochemistry and will affect interpretations of exoplanetary spectra from\nfuture missions like the James Webb Space Telescope. These effects will be\nparticularly critical in the study of planetary atmospheres around M dwarfs,\nincluding Earth-sized planets in the habitable zone. Given the higher activity\nlevels of M dwarfs compared to Sun-like stars, time resolved UV data are needed\nfor more accurate input conditions for exoplanet atmospheric modeling. The\nGalaxy Evolution Explorer (\\emph{GALEX}) provides multi-epoch photometric\nobservations in two UV bands: near-ultraviolet (NUV; 1771 -- 2831 \\AA) and\nfar-ultraviolet (FUV; 1344 -- 1786 \\AA). Within 30 pc of Earth, there are 357\nand 303 M dwarfs in the NUV and FUV bands, respectively, with multiple\\GALEX\nobservations. Simultaneous NUV and FUV detections exist for 145 stars in\nboth\\GALEX bands. Our analyses of these data show that low-mass stars are\ntypically more variable in the FUV than the NUV. Median variability increases\nwith later spectral types in the NUV with no clear trend in the FUV. We find\nevidence that flares increase the FUV flux density far more than the NUV flux\ndensity, leading to variable FUV to NUV flux density ratios in the \\GALEX\\\nbandpasses.The ratio of FUV to NUV flux is important for interpreting the\npresence of atmospheric molecules in planetary atmospheres such as oxygen and\nmethane as a high FUV to NUV ratio may cause false-positive biosignature\ndetections. This ratio of flux density in the\\GALEX\\ bands spans three orders\nof magnitude in our sample, from 0.008 to 4.6, and is 1 to 2 orders of\nmagnitude higher than for G dwarfs like the Sun. These results characterize the\nUV behavior for the largest set of low-mass stars to date.\n",
"title": "HAZMAT II: Ultraviolet Variability of Low-Mass Stars in the GALEX Archive"
}
| null | null | null | null | true | null |
19279
| null |
Default
| null | null |
null |
{
"abstract": " We construct a double-well potential for which the Schrödinger equation can\nbe exactly solved via reducing to the confluent Heun's one. Thus the wave\nfunction is expressed via the confluent Heun's function. The latter is\ntabulated in {\\sl {Maple}} so that the obtained solution is easily treated. The\npotential is infinite at the boundaries of the final interval that makes it to\nbe highly suitable for modeling hydrogen bonds (both ordinary and low-barrier\nones). We exemplify theoretical results by detailed treating the hydrogen bond\nin $KHCO_3$ and show their good agreement with literature experimental data.\n",
"title": "Exactly solvable Schrödinger equation with double-well potential for hydrogen bond"
}
| null | null | null | null | true | null |
19280
| null |
Default
| null | null |
null |
{
"abstract": " Social ties are strongly related to well-being. But what characterizes this\nrelationship? This study investigates social mechanisms explaining how social\nties affect well-being through social integration and social influence, and how\nwell-being affects social ties through social selection. We hypothesize that\nhighly integrated individuals - those with more extensive and dense friendship\nnetworks - report higher emotional well-being than others. Moreover, emotional\nwell-being should be influenced by the well-being of close friends. Finally,\nwell-being should affect friendship selection when individuals prefer others\nwith higher levels of well-being, and others whose well-being is similar to\ntheirs. We test our hypotheses using longitudinal social network and well-being\ndata of 117 individuals living in a graduate housing community. The application\nof a novel extension of Stochastic Actor-Oriented Models for ordered networks\n(ordered SAOMs) allows us to detail and test our hypotheses for weak- and\nstrong-tied friendship networks simultaneously. Results do not support our\nsocial integration and social influence hypotheses but provide evidence for\nselection: individuals with higher emotional well-being tend to have more\nstrong-tied friends, and there are homophily processes regarding emotional\nwell-being in strong-tied networks. Our study highlights the two-directional\nrelationship between social ties and well-being, and demonstrates the\nimportance of considering different tie strengths for various social processes.\n",
"title": "The co-evolution of emotional well-being with weak and strong friendship ties"
}
| null | null | null | null | true | null |
19281
| null |
Default
| null | null |
null |
{
"abstract": " We analyze spectral properties of two mutually related families of magnetic\nSchrödinger operators, $H_{\\mathrm{Sm}}(A)=(i \\nabla +A)^2+\\omega^2\ny^2+\\lambda y \\delta(x)$ and $H(A)=(i \\nabla +A)^2+\\omega^2 y^2+ \\lambda y^2\nV(x y)$ in $L^2(R^2)$, with the parameters $\\omega>0$ and $\\lambda<0$, where\n$A$ is a vector potential corresponding to a homogeneous magnetic field\nperpendicular to the plane and $V$ is a regular nonnegative and compactly\nsupported potential. We show that the spectral properties of the operators\ndepend crucially on the one-dimensional Schrödinger operators $L=\n-\\frac{\\mathrm{d}^2}{\\mathrm{d}x^2} +\\omega^2 +\\lambda \\delta (x)$ and $L (V)=\n- \\frac{\\mathrm{d}^2}{\\mathrm{d}x^2} +\\omega^2 +\\lambda V(x)$, respectively.\nDepending on whether the operators $L$ and $L(V)$ are positive or not, the\nspectrum of $H_{\\mathrm{Sm}}(A)$ and $H(V)$ exhibits a sharp transition.\n",
"title": "A magnetic version of the Smilansky-Solomyak model"
}
| null | null | null | null | true | null |
19282
| null |
Default
| null | null |
null |
{
"abstract": " Given a positive function $u\\in W^{1,n}$, we define its John-Nirenberg radius\nat point $x$ to be the supreme of the radius such that $\\int_{B_t}|\\nabla\\log\nu|^n<\\epsilon_0^n$ when $n>2$, and $\\int_{B_t}|\\nabla u|^2<\\epsilon_0^2$ when\n$n=2$. We will show that for a collapsing sequence in a fixed conformal class\nunder some curvature conditions, the radius is bounded below by a positive\nconstant. As applications, we will study the convergence of a conformal metric\nsequence on a $4$-manifold with bounded $\\|K\\|_{W^{1,2}}$, and prove a\ngeneralized Hélein's Convergence Theorem.\n",
"title": "John-Nirenberg Radius and Collapse in Conformal Geometry"
}
| null | null | null | null | true | null |
19283
| null |
Default
| null | null |
null |
{
"abstract": " The recent literature on deep learning offers new tools to learn a rich\nprobability distribution over high dimensional data such as images or sounds.\nIn this work we investigate the possibility of learning the prior distribution\nover neural network parameters using such tools. Our resulting variational\nBayes algorithm generalizes well to new tasks, even when very few training\nexamples are provided. Furthermore, this learned prior allows the model to\nextrapolate correctly far from a given task's training data on a meta-dataset\nof periodic signals.\n",
"title": "Deep Prior"
}
| null | null | null | null | true | null |
19284
| null |
Default
| null | null |
null |
{
"abstract": " Recent advances show that two-dimensional linear discriminant analysis\n(2DLDA) is a successful matrix based dimensionality reduction method. However,\n2DLDA may encounter the singularity issue theoretically and the sensitivity to\noutliers. In this paper, a generalized Lp-norm 2DLDA framework with\nregularization for an arbitrary $p>0$ is proposed, named G2DLDA. There are\nmainly two contributions of G2DLDA: one is G2DLDA model uses an arbitrary\nLp-norm to measure the between-class and within-class scatter, and hence a\nproper $p$ can be selected to achieve the robustness. The other one is that by\nintroducing an extra regularization term, G2DLDA achieves better generalization\nperformance, and solves the singularity problem. In addition, G2DLDA can be\nsolved through a series of convex problems with equality constraint, and it has\nclosed solution for each single problem. Its convergence can be guaranteed\ntheoretically when $1\\leq p\\leq2$. Preliminary experimental results on three\ncontaminated human face databases show the effectiveness of the proposed\nG2DLDA.\n",
"title": "Generalized two-dimensional linear discriminant analysis with regularization"
}
| null | null | null | null | true | null |
19285
| null |
Default
| null | null |
null |
{
"abstract": " Distinct striation patterns are observed in the spectrograms of speech and\nmusic. This motivated us to propose three novel time-frequency features for\nspeech-music classification. These features are extracted in two stages. First,\na preset number of prominent spectral peak locations are identified from the\nspectra of each frame. These important peak locations obtained from each frame\nare used to form Spectral peak sequences (SPS) for an audio interval. In second\nstage, these SPS are treated as time series data of frequency locations. The\nproposed features are extracted as periodicity, average frequency and\nstatistical attributes of these spectral peak sequences. Speech-music\ncategorization is performed by learning binary classifiers on these features.\nWe have experimented with Gaussian mixture models, support vector machine and\nrandom forest classifiers. Our proposal is validated on four datasets and\nbenchmarked against three baseline approaches. Experimental results establish\nthe validity of our proposal.\n",
"title": "Time-Frequency Audio Features for Speech-Music Classification"
}
| null | null | null | null | true | null |
19286
| null |
Default
| null | null |
null |
{
"abstract": " This paper presents a motion planner for systems subject to kinematic and\ndynamic constraints. The former appear when kinematic loops are present in the\nsystem, such as in parallel manipulators, in robots that cooperate to achieve a\ngiven task, or in situations involving contacts with the environment. The\nlatter are necessary to obtain realistic trajectories, taking into account the\nforces acting on the system. The kinematic constraints make the state space\nbecome an implicitly-defined manifold, which complicates the application of\ncommon motion planning techniques. To address this issue, the planner\nconstructs an atlas of the state space manifold incrementally, and uses this\natlas both to generate random states and to dynamically simulate the steering\nof the system towards such states. The resulting tools are then exploited to\nconstruct a rapidly-exploring random tree (RRT) over the state space. To the\nbest of our knowledge, this is the first randomized kinodynamic planner for\nimplicitly-defined state spaces. The test cases presented in this paper\nvalidate the approach in significantly-complex systems.\n",
"title": "Kinodynamic Planning on Constraint Manifolds"
}
| null | null | null | null | true | null |
19287
| null |
Default
| null | null |
null |
{
"abstract": " Can we learn a binary classifier from only positive data, without any\nnegative data or unlabeled data? We show that if one can equip positive data\nwith confidence (positive-confidence), one can successfully learn a binary\nclassifier, which we name positive-confidence (Pconf) classification. Our work\nis related to one-class classification which is aimed at \"describing\" the\npositive class by clustering-related methods, but one-class classification does\nnot have the ability to tune hyper-parameters and their aim is not on\n\"discriminating\" positive and negative classes. For the Pconf classification\nproblem, we provide a simple empirical risk minimization framework that is\nmodel-independent and optimization-independent. We theoretically establish the\nconsistency and an estimation error bound, and demonstrate the usefulness of\nthe proposed method for training deep neural networks through experiments.\n",
"title": "Binary Classification from Positive-Confidence Data"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
19288
| null |
Validated
| null | null |
null |
{
"abstract": " It is well-known that kernel regression estimators do not produce a constant\nestimator variance over a domain. To correct this problem, Nishida and Kanazawa\n(2015) proposed a variance-stabilizing (VS) local variable bandwidth for Local\nLinear (LL) regression estimator. In contrast, Choi and Hall (1998) proposed\nthe skewing (SK) methods for a univariate LL estimator and constructed a convex\ncombination of one LL estimator and two SK estimators that are symmetrically\nplaced on both sides of the LL estimator (the convex combination (CC)\nestimator) to eliminate higher-order terms in its asymptotic bias. To obtain a\nCC estimator with a constant estimator variance without employing the VS local\nvariable bandwidth, the weight in the convex combination must be determined\nlocally to produce a constant estimator variance. In this study, we compare the\nperformances of two VS methods for a CC estimator and find cases in which the\nweighting method can superior to the VS bandwidth method in terms of the degree\nof variance stabilization.\n",
"title": "Skewing Methods for Variance-Stabilizing Local Linear Regression Estimation"
}
| null | null | null | null | true | null |
19289
| null |
Default
| null | null |
null |
{
"abstract": " To harness the complexity of their high-dimensional bodies during\nsensorimotor development, infants are guided by patterns of freezing and\nfreeing of degrees of freedom. For instance, when learning to reach, infants\nfree the degrees of freedom in their arm proximodistally, i.e. from joints that\nare closer to the body to those that are more distant. Here, we formulate and\nstudy computationally the hypothesis that such patterns can emerge\nspontaneously as the result of a family of stochastic optimization processes\n(evolution strategies with covariance-matrix adaptation), without an innate\nencoding of a maturational schedule. In particular, we present simulated\nexperiments with an arm where a computational learner progressively acquires\nreaching skills through adaptive exploration, and we show that a proximodistal\norganization appears spontaneously, which we denote PDFF (ProximoDistal\nFreezing and Freeing of degrees of freedom). We also compare this emergent\norganization between different arm morphologies -- from human-like to quite\nunnatural ones -- to study the effect of different kinematic structures on the\nemergence of PDFF. Keywords: human motor learning; proximo-distal exploration;\nstochastic optimization; modelling; evolution strategies; cross-entropy\nmethods; policy search; morphology.}\n",
"title": "Proximodistal Exploration in Motor Learning as an Emergent Property of Optimization"
}
| null | null | null | null | true | null |
19290
| null |
Default
| null | null |
null |
{
"abstract": " We prove that $\\tilde{\\Theta}(k d^2 / \\varepsilon^2)$ samples are necessary\nand sufficient for learning a mixture of $k$ Gaussians in $\\mathbb{R}^d$, up to\nerror $\\varepsilon$ in total variation distance. This improves both the known\nupper bounds and lower bounds for this problem. For mixtures of axis-aligned\nGaussians, we show that $\\tilde{O}(k d / \\varepsilon^2)$ samples suffice,\nmatching a known lower bound. Moreover, these results hold in the\nagnostic-learning/robust-estimation setting as well, where the target\ndistribution is only approximately a mixture of Gaussians.\nThe upper bound is shown using a novel technique for distribution learning\nbased on a notion of `compression.' Any class of distributions that allows such\na compression scheme can also be learned with few samples. Moreover, if a class\nof distributions has such a compression scheme, then so do the classes of\nproducts and mixtures of those distributions. The core of our main result is\nshowing that the class of Gaussians in $\\mathbb{R}^d$ admits a small-sized\ncompression scheme.\n",
"title": "Near-optimal Sample Complexity Bounds for Robust Learning of Gaussians Mixtures via Compression Schemes"
}
| null | null | null | null | true | null |
19291
| null |
Default
| null | null |
null |
{
"abstract": " We present a heuristic based algorithm to induce \\textit{nonmonotonic} logic\nprograms that will explain the behavior of XGBoost trained classifiers. We use\nthe technique based on the LIME approach to locally select the most important\nfeatures contributing to the classification decision. Then, in order to explain\nthe model's global behavior, we propose the LIME-FOLD algorithm ---a\nheuristic-based inductive logic programming (ILP) algorithm capable of learning\nnon-monotonic logic programs---that we apply to a transformed dataset produced\nby LIME. Our proposed approach is agnostic to the choice of the ILP algorithm.\nOur experiments with UCI standard benchmarks suggest a significant improvement\nin terms of classification evaluation metrics. Meanwhile, the number of induced\nrules dramatically decreases compared to ALEPH, a state-of-the-art ILP system.\n",
"title": "Induction of Non-Monotonic Logic Programs to Explain Boosted Tree Models Using LIME"
}
| null | null | null | null | true | null |
19292
| null |
Default
| null | null |
null |
{
"abstract": " Manipulation of deformable objects, such as ropes and cloth, is an important\nbut challenging problem in robotics. We present a learning-based system where a\nrobot takes as input a sequence of images of a human manipulating a rope from\nan initial to goal configuration, and outputs a sequence of actions that can\nreproduce the human demonstration, using only monocular images as input. To\nperform this task, the robot learns a pixel-level inverse dynamics model of\nrope manipulation directly from images in a self-supervised manner, using about\n60K interactions with the rope collected autonomously by the robot. The human\ndemonstration provides a high-level plan of what to do and the low-level\ninverse model is used to execute the plan. We show that by combining the high\nand low-level plans, the robot can successfully manipulate a rope into a\nvariety of target shapes using only a sequence of human-provided images for\ndirection.\n",
"title": "Combining Self-Supervised Learning and Imitation for Vision-Based Rope Manipulation"
}
| null | null | null | null | true | null |
19293
| null |
Default
| null | null |
null |
{
"abstract": " We derive upper and lower bounds for the policy regret of $T$-round online\nlearning problems with graph-structured feedback, where the adversary is\nnonoblivious but assumed to have a bounded memory. We obtain upper bounds of\n$\\widetilde O(T^{2/3})$ and $\\widetilde O(T^{3/4})$ for strongly-observable and\nweakly-observable graphs, respectively, based on analyzing a variant of the\nExp3 algorithm. When the adversary is allowed a bounded memory of size 1, we\nshow that a matching lower bound of $\\widetilde\\Omega(T^{2/3})$ is achieved in\nthe case of full-information feedback. We also study the particular loss\nstructure of an oblivious adversary with switching costs, and show that in such\na setting, non-revealing strongly-observable feedback graphs achieve a lower\nbound of $\\widetilde\\Omega(T^{2/3})$, as well.\n",
"title": "Online learning with graph-structured feedback against adaptive adversaries"
}
| null | null | null | null | true | null |
19294
| null |
Default
| null | null |
null |
{
"abstract": " The dynamical characteristics of electromagnetic fields include energy,\nmomentum, angular momentum (spin) and helicity. We analyze their spatial\ndistributions near the planar interface between two transparent and\nnon-dispersive media, when the incident monochromatic plane wave with arbitrary\npolarization is totally reflected, and an evanescent wave is formed in the\nmedium with lower optical density. Based on the recent arguments in favor of\nthe Minkowski definition of the electromagnetic momentum in a material medium\n[Phys. Rev. A 83, 013823 (2011); 86, 055802 (2012); Phys. Rev. Lett. 119,\n073901 (2017)], we derive the explicit expressions for the dynamical\ncharacteristics in both media, with special attention to their behavior at the\ninterface. Especially, the \"extraordinary\" spin and momentum components\northogonal to the plane of incidence are described, and the canonical (spin -\norbital) momentum decomposition is performed that contains no singular terms.\nThe field energy, helicity, the spin momentum and orbital momentum components\nare everywhere regular but experience discontinuities at the interface; the\nspin components parallel to the interface appear to be continuous, which\ntestifies for the consistency of the adopted Minkowski picture. The results\nsupply a meaningful example of the electromagnetic momentum decomposition, with\nseparation of spatial and polarization degrees of freedom, in inhomogeneous\nmedia, and can be used in engineering the structured fields designed for\noptical sorting, dispatching and micromanipulation.\n",
"title": "Dynamical characteristics of electromagnetic field under conditions of total reflection"
}
| null | null | null | null | true | null |
19295
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we give three functors $\\mathfrak{P}$, $[\\cdot]_K$ and\n$\\mathfrak{F}$ on the category of C$^\\ast$-algebras. The functor $\\mathfrak{P}$\nassigns to each C$^\\ast$-algebra $\\mathcal{A}$ a pre-C$^\\ast$-algebra\n$\\mathfrak{P}(\\mathcal{A})$ with completion $[\\mathcal{A}]_K$. The functor\n$[\\cdot]_K$ assigns to each C$^\\ast$-algebra $\\mathcal{A}$ the Cauchy extension\n$[\\mathcal{A}]_K$ of $\\mathcal{A}$ by a non-unital C$^\\ast$-algebra\n$\\mathfrak{F}(\\mathcal{A})$. Some properties of these functors are also given.\nIn particular, we show that the functors $[\\cdot]_K$ and $\\mathfrak{F}$ are\nexact and the functor $\\mathfrak{P}$ is normal exact.\n",
"title": "Functors induced by Cauchy extension of C*-algebras"
}
| null | null | null | null | true | null |
19296
| null |
Default
| null | null |
null |
{
"abstract": " We define the notion of a hierarchically cocompact classifying space for a\nfamily of subgroups of a group. Our main application is to show that the\nmapping class group $\\mbox{Mod}(S)$ of any connected oriented compact surface\n$S$, possibly with punctures and boundary components and with negative Euler\ncharacteristic has a hierarchically cocompact model for the family of virtually\ncyclic subgroups of dimension at most $\\mbox{vcd} \\mbox{Mod}(S)+1$. When the\nsurface is closed, we prove that this bound is optimal. In particular, this\nanswers a question of Lück for mapping class groups of surfaces.\n",
"title": "Hierarchically cocompact classifying spaces for mapping class groups of surfaces"
}
| null | null | null | null | true | null |
19297
| null |
Default
| null | null |
null |
{
"abstract": " An explosion of high-throughput DNA sequencing in the past decade has led to\na surge of interest in population-scale inference with whole-genome data.\nRecent work in population genetics has centered on designing inference methods\nfor relatively simple model classes, and few scalable general-purpose inference\ntechniques exist for more realistic, complex models. To achieve this, two\ninferential challenges need to be addressed: (1) population data are\nexchangeable, calling for methods that efficiently exploit the symmetries of\nthe data, and (2) computing likelihoods is intractable as it requires\nintegrating over a set of correlated, extremely high-dimensional latent\nvariables. These challenges are traditionally tackled by likelihood-free\nmethods that use scientific simulators to generate datasets and reduce them to\nhand-designed, permutation-invariant summary statistics, often leading to\ninaccurate inference. In this work, we develop an exchangeable neural network\nthat performs summary statistic-free, likelihood-free inference. Our framework\ncan be applied in a black-box fashion across a variety of simulation-based\ntasks, both within and outside biology. We demonstrate the power of our\napproach on the recombination hotspot testing problem, outperforming the\nstate-of-the-art.\n",
"title": "A Likelihood-Free Inference Framework for Population Genetic Data using Exchangeable Neural Networks"
}
| null | null | null | null | true | null |
19298
| null |
Default
| null | null |
null |
{
"abstract": " Quantum effects, prevalent in the microscopic scale, generally elusive in\nmacroscopic systems due to dissipation and decoherence. Quantum phenomena in\nlarge systems emerge only when particles are strongly correlated as in\nsuperconductors and superfluids. Cooperative interaction of correlated atoms\nwith electromagnetic fields leads to superradiance, the enhanced quantum\nradiation phenomenon, exhibiting novel physics such as quantum Dicke phase and\nultranarrow linewidth for optical clocks. Recent researches to imprint atomic\ncorrelation directly demonstrated controllable collective atom-field\ninteractions. Here, we report cavity-mediated coherent single-atom\nsuperradiance. Single atoms with predefined correlation traverse a high-Q\ncavity one by one, emitting photons cooperatively with the atoms already gone\nthrough the cavity. Such collective behavior of time-separated atoms is\nmediated by the long-lived cavity field. As a result, a coherent field is\ngenerated in the steady state, whose intensity varies as the square of the\nnumber of traversing atoms during the cavity decay time, exhibiting more than\nten-fold enhancement from noncollective cases. The correlation among single\natoms is prepared with the aligned atomic phase achieved by nanometer-precision\nposition control of atoms with a nanohole-array aperture. The present work\ndeepens our understanding of the collective matter-light interaction and\nprovides an advanced platform for phase-controlled atom-field interactions.\n",
"title": "Coherent single-atom superradiance"
}
| null | null | null | null | true | null |
19299
| null |
Default
| null | null |
null |
{
"abstract": " This paper presents a deep learning framework that is capable of solving\npartially observable locomotion tasks based on our novel interpretation of\nRecurrent Deterministic Policy Gradient (RDPG). We study on bias of sampled\nerror measure and its variance induced by the partial observability of\nenvironment and subtrajectory sampling, respectively. Three major improvements\nare introduced in our RDPG based learning framework: tail-step bootstrap of\ninterpolated temporal difference, initialisation of hidden state using past\ntrajectory scanning, and injection of external experiences learned by other\nagents. The proposed learning framework was implemented to solve the\nBipedal-Walker challenge in OpenAI's gym simulation environment where only\npartial state information is available. Our simulation study shows that the\nautonomous behaviors generated by the RDPG agent are highly adaptive to a\nvariety of obstacles and enables the agent to effectively traverse rugged\nterrains for long distance with higher success rate than leading contenders.\n",
"title": "Recurrent Deterministic Policy Gradient Method for Bipedal Locomotion on Rough Terrain Challenge"
}
| null | null | null | null | true | null |
19300
| null |
Default
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.