text
null | inputs
dict | prediction
null | prediction_agent
null | annotation
list | annotation_agent
null | multi_label
bool 1
class | explanation
null | id
stringlengths 1
5
| metadata
null | status
stringclasses 2
values | event_timestamp
null | metrics
null |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
null |
{
"abstract": " We introduce Deep-HiTS, a rotation invariant convolutional neural network\n(CNN) model for classifying images of transients candidates into artifacts or\nreal sources for the High cadence Transient Survey (HiTS). CNNs have the\nadvantage of learning the features automatically from the data while achieving\nhigh performance. We compare our CNN model against a feature engineering\napproach using random forests (RF). We show that our CNN significantly\noutperforms the RF model reducing the error by almost half. Furthermore, for a\nfixed number of approximately 2,000 allowed false transient candidates per\nnight we are able to reduce the miss-classified real transients by\napproximately 1/5. To the best of our knowledge, this is the first time CNNs\nhave been used to detect astronomical transient events. Our approach will be\nvery useful when processing images from next generation instruments such as the\nLarge Synoptic Survey Telescope (LSST). We have made all our code and data\navailable to the community for the sake of allowing further developments and\ncomparisons at this https URL.\n",
"title": "Deep-HiTS: Rotation Invariant Convolutional Neural Network for Transient Detection"
}
| null | null | null | null | true | null |
17301
| null |
Default
| null | null |
null |
{
"abstract": " In various approaches to learning, notably in domain adaptation, active\nlearning, learning under covariate shift, semi-supervised learning, learning\nwith concept drift, and the like, one often wants to compare a baseline\nclassifier to one or more advanced (or at least different) strategies. In this\nchapter, we basically argue that if such classifiers, in their respective\ntraining phases, optimize a so-called surrogate loss that it may also be\nvaluable to compare the behavior of this loss on the test set, next to the\nregular classification error rate. It can provide us with an additional view on\nthe classifiers' relative performances that error rates cannot capture. As an\nexample, limited but convincing empirical results demonstrates that we may be\nable to find semi-supervised learning strategies that can guarantee performance\nimprovements with increasing numbers of unlabeled data in terms of\nlog-likelihood. In contrast, the latter may be impossible to guarantee for the\nclassification error rate.\n",
"title": "On Measuring and Quantifying Performance: Error Rates, Surrogate Loss, and an Example in SSL"
}
| null | null | null | null | true | null |
17302
| null |
Default
| null | null |
null |
{
"abstract": " The principle of common cause asserts that positive correlations between\ncausally unrelated events ought to be explained through the action of some\nshared causal factors. Reichenbachian common cause systems are probabilistic\nstructures aimed at accounting for cases where correlations of the aforesaid\nsort cannot be explained through the action of a single common cause. The\nexistence of Reichenbachian common cause systems of arbitrary finite size for\neach pair of non-causally correlated events was allegedly demonstrated by\nHofer-Szabó and Rédei in 2006. This paper shows that their proof is\nlogically deficient, and we propose an improved proof.\n",
"title": "Do Reichenbachian Common Cause Systems of Arbitrary Finite Size Exist?"
}
| null | null | null | null | true | null |
17303
| null |
Default
| null | null |
null |
{
"abstract": " When three species compete cyclically in a well-mixed, stochastic system of\n$N$ individuals, extinction is known to typically occur at times scaling as the\nsystem size $N$. This happens, for example, in rock-paper-scissors games or\nconserved Lotka-Volterra models in which every pair of individuals can interact\non a complete graph. Here we show that if the competing individuals also have a\n\"social temperament\" to be either introverted or extroverted, leading them to\ncut or add links respectively, then long-living state in which all species\ncoexist can occur when both introverts and extroverts are present. These states\nare non-equilibrium quasi-steady states, maintained by a subtle balance between\nspecies competition and network dynamcis. Remarkably, much of the phenomena is\nembodied in a mean-field description. However, an intuitive understanding of\nwhy diversity stabilizes the co-evolving node and link dynamics remains an open\nissue.\n",
"title": "Co-evolution of nodes and links: diversity driven coexistence in cyclic competition of three species"
}
| null | null | null | null | true | null |
17304
| null |
Default
| null | null |
null |
{
"abstract": " We study the multiclass online learning problem where a forecaster makes a\nsequence of predictions using the advice of $n$ experts. Our main contribution\nis to analyze the regime where the best expert makes at most $b$ mistakes and\nto show that when $b = o(\\log_4{n})$, the expected number of mistakes made by\nthe optimal forecaster is at most $\\log_4{n} + o(\\log_4{n})$. We also describe\nan adversary strategy showing that this bound is tight and that the worst case\nis attained for binary prediction.\n",
"title": "Online Learning with an Almost Perfect Expert"
}
| null | null |
[
"Statistics"
] | null | true | null |
17305
| null |
Validated
| null | null |
null |
{
"abstract": " Deep learning techniques have been hugely successful for traditional\nsupervised and unsupervised machine learning problems. In large part, these\ntechniques solve continuous optimization problems. Recently however, discrete\ngenerative deep learning models have been successfully used to efficiently\nsearch high-dimensional discrete spaces. These methods work by representing\ndiscrete objects as sequences, for which powerful sequence-based deep models\ncan be employed. Unfortunately, these techniques are significantly hindered by\nthe fact that these generative models often produce invalid sequences. As a\nstep towards solving this problem, we propose to learn a deep recurrent\nvalidator model. Given a partial sequence, our model learns the probability of\nthat sequence occurring as the beginning of a full valid sequence. Thus this\nidentifies valid versus invalid sequences and crucially it also provides\ninsight about how individual sequence elements influence the validity of\ndiscrete objects. To learn this model we propose an approach inspired by\nseminal work in Bayesian active learning. On a synthetic dataset, we\ndemonstrate the ability of our model to distinguish valid and invalid\nsequences. We believe this is a key step toward learning generative models that\nfaithfully produce valid discrete objects.\n",
"title": "Actively Learning what makes a Discrete Sequence Valid"
}
| null | null | null | null | true | null |
17306
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we study the infinitesimal symmetries, Newtonoid vector fields,\ninfinitesimal Noether symmetries and conservation laws of Hamiltonian systems.\nUsing the dynamical covariant derivative and Jacobi endomorphism on the\ncotangent bundle we find the invariant equations of infinitesimal symmetries\nand Newtonoid vector fields and prove that the canonical nonlinear connection\ninduced by a regular Hamiltonian can be determined by these symmetries.\nFinally, an example from optimal control theory is given.\n",
"title": "Symmetries and conservation laws of Hamiltonian systems"
}
| null | null | null | null | true | null |
17307
| null |
Default
| null | null |
null |
{
"abstract": " The studying of anomalous diffusion by pulsed field gradient (PFG) diffusion\ntechnique still faces challenges. Two different research groups have proposed\nmodified Bloch equation for anomalous diffusion. However, these equations have\ndifferent forms and, therefore, yield inconsistent results. The discrepancy in\nthese reported modified Bloch equations may arise from different ways of\ncombining the fractional diffusion equation with the precession equation where\nthe time derivatives have different derivative orders and forms. Moreover, to\nthe best of my knowledge, the general PFG signal attenuation expression\nincluding finite gradient pulse width (FGPW) effect for time-space fractional\ndiffusion based on the fractional derivative has yet to be reported by other\nmethods. Here, based on different combination strategy, two new modified Bloch\nequations are proposed, which belong to two significantly different types: a\ndifferential type based on the fractal derivative and an integral type based on\nthe fractional derivative. The merit of the integral type modified Bloch\nequation is that the original properties of the contributions from linear or\nnonlinear processes remain unchanged at the instant of the combination. The\ngeneral solutions including the FGPW effect were derived from these two\nequations as well as from two other methods: a method observing the signal\nintensity at the origin and the recently reported effective phase shift\ndiffusion equation method. The relaxation effect was also considered. It is\nfound that the relaxation behavior influenced by fractional diffusion based on\nthe fractional derivative deviates from that of normal diffusion. The general\nsolution agrees perfectly with continuous-time random walk (CTRW) simulations\nas well as reported literature results. The new modified Bloch equations is a\nvaluable tool to describe PFG anomalous diffusion in NMR and MRI.\n",
"title": "Fractional differential and fractional integral modified-Bloch equations for PFG anomalous diffusion and their general solutions"
}
| null | null | null | null | true | null |
17308
| null |
Default
| null | null |
null |
{
"abstract": " We report a nontrivial transition in the core structure of vortices in\ntwo-band superconductors as a function of interband impurity scattering. We\ndemonstrate that, in addition to singular zeros of the order parameter, the\nvortices there can acquire a circular nodal line around the singular point in\none of the superconducting components. It results in the formation of the\npeculiar \"moat\"-like profile in one of the superconducting gaps. The moat-core\nvortices occur generically in the vicinity of the impurity-induced crossover\nbetween $s_{\\pm}$ and $s_{++}$ states.\n",
"title": "Change of the vortex core structure in two-band superconductors at impurity-scattering-driven $s_\\pm/s_{++}$ crossover"
}
| null | null |
[
"Physics"
] | null | true | null |
17309
| null |
Validated
| null | null |
null |
{
"abstract": " Multi-armed bandit (MAB) is a class of online learning problems where a\nlearning agent aims to maximize its expected cumulative reward while repeatedly\nselecting to pull arms with unknown reward distributions. We consider a\nscenario where the reward distributions may change in a piecewise-stationary\nfashion at unknown time steps. We show that by incorporating a simple\nchange-detection component with classic UCB algorithms to detect and adapt to\nchanges, our so-called M-UCB algorithm can achieve nearly optimal regret bound\non the order of $O(\\sqrt{MKT\\log T})$, where $T$ is the number of time steps,\n$K$ is the number of arms, and $M$ is the number of stationary segments.\nComparison with the best available lower bound shows that our M-UCB is nearly\noptimal in $T$ up to a logarithmic factor. We also compare M-UCB with the\nstate-of-the-art algorithms in numerical experiments using a public Yahoo!\ndataset to demonstrate its superior performance.\n",
"title": "Nearly Optimal Adaptive Procedure with Change Detection for Piecewise-Stationary Bandit"
}
| null | null | null | null | true | null |
17310
| null |
Default
| null | null |
null |
{
"abstract": " We investigate the initial-boundary value problem for the general\nthree-component nonlinear Schrodinger (gtc-NLS) equation with a 4x4 Lax pair on\na finite interval by extending the Fokas unified approach. The solutions of the\ngtc-NLS equation can be expressed in terms of the solutions of a 4x4 matrix\nRiemann-Hilbert (RH) problem formulated in the complex k-plane. Moreover, the\nrelevant jump matrices of the RH problem can be explicitly found via the three\nspectral functions arising from the initial data, the Dirichlet-Neumann\nboundary data. The global relation is also established to deduce two distinct\nbut equivalent types of representations (i.e., one by using the large k of\nasymptotics of the eigenfunctions and another one in terms of the\nGelfand-Levitan-Marchenko (GLM) method) for the Dirichlet and Neumann boundary\nvalue problems. Moreover, the relevant formulae for boundary value problems on\nthe finite interval can reduce to ones on the half-line as the length of the\ninterval approaches to infinity. Finally, we also give the linearizable\nboundary conditions for the GLM representation.\n",
"title": "An initial-boundary value problem of the general three-component nonlinear Schrodinger equation with a 4x4 Lax pair on a finite interval"
}
| null | null | null | null | true | null |
17311
| null |
Default
| null | null |
null |
{
"abstract": " We demonstrate that a deep neural network can significantly improve optical\nmicroscopy, enhancing its spatial resolution over a large field-of-view and\ndepth-of-field. After its training, the only input to this network is an image\nacquired using a regular optical microscope, without any changes to its design.\nWe blindly tested this deep learning approach using various tissue samples that\nare imaged with low-resolution and wide-field systems, where the network\nrapidly outputs an image with remarkably better resolution, matching the\nperformance of higher numerical aperture lenses, also significantly surpassing\ntheir limited field-of-view and depth-of-field. These results are\ntransformative for various fields that use microscopy tools, including e.g.,\nlife sciences, where optical microscopy is considered as one of the most widely\nused and deployed techniques. Beyond such applications, our presented approach\nis broadly applicable to other imaging modalities, also spanning different\nparts of the electromagnetic spectrum, and can be used to design computational\nimagers that get better and better as they continue to image specimen and\nestablish new transformations among different modes of imaging.\n",
"title": "Deep Learning Microscopy"
}
| null | null | null | null | true | null |
17312
| null |
Default
| null | null |
null |
{
"abstract": " The development of needle-free injection systems utilizing high-speed\nmicrojets is of great importance to world healthcare. It is thus crucial to\ncontrol the microjets, which are often induced by underwater shock waves. In\nthis contribution from fluid-mechanics point of view, we experimentally\ninvestigate the effect of a shock wave on the velocity of a free surface\n(microjet) and underwater cavitation onset in a microchannel, focusing on the\npressure impulse and peak pressure of the shock wave. The shock wave used had a\nnon-spherically-symmetric peak pressure distribution and a spherically\nsymmetric pressure impulse distribution [Tagawa et al., J. Fluid Mech., 2016,\n808, 5-18]. First, we investigate the effect of the shock wave on the jet\nvelocity by installing a narrow tube and a hydrophone in different\nconfigurations in a large water tank, and measuring the shock wave pressure and\nthe jet velocity simultaneously. The results suggest that the jet velocity\ndepends only on the pressure impulse of the shock wave. We then investigate the\neffect of the shock wave on the cavitation onset by taking measurements in an\nL-shaped microchannel. The results suggest that the probability of cavitation\nonset depends only on the peak pressure of the shock wave. In addition, the jet\nvelocity varies according to the presence or absence of cavitation. The above\nfindings provide new insights for advancing a control method for high-speed\nmicrojets.\n",
"title": "Effects of pressure impulse and peak pressure of a shock wave on microjet velocity and the onset of cavitation in a microchannel"
}
| null | null | null | null | true | null |
17313
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we initiate a rigorous theoretical study of clustering with\nnoisy queries (or a faulty oracle). Given a set of $n$ elements, our goal is to\nrecover the true clustering by asking minimum number of pairwise queries to an\noracle. Oracle can answer queries of the form : \"do elements $u$ and $v$ belong\nto the same cluster?\" -- the queries can be asked interactively (adaptive\nqueries), or non-adaptively up-front, but its answer can be erroneous with\nprobability $p$. In this paper, we provide the first information theoretic\nlower bound on the number of queries for clustering with noisy oracle in both\nsituations. We design novel algorithms that closely match this query complexity\nlower bound, even when the number of clusters is unknown. Moreover, we design\ncomputationally efficient algorithms both for the adaptive and non-adaptive\nsettings. The problem captures/generalizes multiple application scenarios. It\nis directly motivated by the growing body of work that use crowdsourcing for\n{\\em entity resolution}, a fundamental and challenging data mining task aimed\nto identify all records in a database referring to the same entity. Here crowd\nrepresents the noisy oracle, and the number of queries directly relates to the\ncost of crowdsourcing. Another application comes from the problem of {\\em sign\nedge prediction} in social network, where social interactions can be both\npositive and negative, and one must identify the sign of all pair-wise\ninteractions by querying a few pairs. Furthermore, clustering with noisy oracle\nis intimately connected to correlation clustering, leading to improvement\ntherein. Finally, it introduces a new direction of study in the popular {\\em\nstochastic block model} where one has an incomplete stochastic block model\nmatrix to recover the clusters.\n",
"title": "Clustering with Noisy Queries"
}
| null | null | null | null | true | null |
17314
| null |
Default
| null | null |
null |
{
"abstract": " Classical reverse-mode automatic differentiation (AD) imposes only a small\nconstant-factor overhead in operation count over the original computation, but\nhas storage requirements that grow, in the worst case, in proportion to the\ntime consumed by the original computation. This storage blowup can be\nameliorated by checkpointing, a process that reorders application of classical\nreverse-mode AD over an execution interval to tradeoff space \\vs\\ time.\nApplication of checkpointing in a divide-and-conquer fashion to strategically\nchosen nested execution intervals can break classical reverse-mode AD into\nstages which can reduce the worst-case growth in storage from linear to\nsublinear. Doing this has been fully automated only for computations of\nparticularly simple form, with checkpoints spanning execution intervals\nresulting from a limited set of program constructs. Here we show how the\ntechnique can be automated for arbitrary computations. The essential innovation\nis to apply the technique at the level of the language implementation itself,\nthus allowing checkpoints to span any execution interval.\n",
"title": "Divide-and-Conquer Checkpointing for Arbitrary Programs with No User Annotation"
}
| null | null |
[
"Computer Science"
] | null | true | null |
17315
| null |
Validated
| null | null |
null |
{
"abstract": " Many-degree-scale gamma-ray halos are expected to surround extragalactic\nhigh-energy gamma ray sources. These arise from the inverse Compton emission of\nan intergalactic population of relativistic electron/positron pairs generated\nby the annihilation of >100 GeV gamma rays on the extragalactic background\nlight. These are typically anisotropic due to the jetted structure from which\nthey originate or the presence of intergalactic magnetic fields. Here we\npropose a novel method for detecting these inverse-Compton gamma-ray halos\nbased upon this anisotropic structure. Specifically, we show that by stacking\nsuitably defined angular power spectra instead of images it is possible to\nrobustly detect gamma-ray halos with existing Fermi Large Area Telescope (LAT)\nobservations for a broad class of intergalactic magnetic fields. Importantly,\nthese are largely insensitive to systematic uncertainties within the LAT\ninstrumental response or associated with contaminating astronomical sources.\n",
"title": "Bow Ties in the Sky II: Searching for Gamma-ray Halos in the Fermi Sky Using Anisotropy"
}
| null | null |
[
"Physics"
] | null | true | null |
17316
| null |
Validated
| null | null |
null |
{
"abstract": " In this work we investigate a one-dimensional parity-time (PT)-symmetric\nmagnetic metamaterial consisting of split-ring dimers having gain or loss.\nEmploying a Melnikov analysis we study the existence of localized travelling\nwaves, i.e. homoclinic or heteroclinic solutions. We find conditions under\nwhich the homoclinic or heteroclinic orbits persist. Our analytical results are\nfound to be in good agreement with direct numerical computations. For the\nparticular nonlinearity admitting travelling kinks, numerically we observe\nhomoclinic snaking in the bifurcation diagram. The Melnikov analysis yields a\ngood approximation to one of the boundaries of the snaking profile.\n",
"title": "Gain-loss-driven travelling waves in PT-symmetric nonlinear metamaterials"
}
| null | null | null | null | true | null |
17317
| null |
Default
| null | null |
null |
{
"abstract": " We present Generative Adversarial Capsule Network (CapsuleGAN), a framework\nthat uses capsule networks (CapsNets) instead of the standard convolutional\nneural networks (CNNs) as discriminators within the generative adversarial\nnetwork (GAN) setting, while modeling image data. We provide guidelines for\ndesigning CapsNet discriminators and the updated GAN objective function, which\nincorporates the CapsNet margin loss, for training CapsuleGAN models. We show\nthat CapsuleGAN outperforms convolutional-GAN at modeling image data\ndistribution on MNIST and CIFAR-10 datasets, evaluated on the generative\nadversarial metric and at semi-supervised image classification.\n",
"title": "CapsuleGAN: Generative Adversarial Capsule Network"
}
| null | null | null | null | true | null |
17318
| null |
Default
| null | null |
null |
{
"abstract": " Zoonotic diseases are a major cause of morbidity, and productivity losses in\nboth humans and animal populations. Identifying the source of food-borne\nzoonoses (e.g. an animal reservoir or food product) is crucial for the\nidentification and prioritisation of food safety interventions. For many\nzoonotic diseases it is difficult to attribute human cases to sources of\ninfection because there is little epidemiological information on the cases.\nHowever, microbial strain typing allows zoonotic pathogens to be categorised,\nand the relative frequencies of the strain types among the sources and in human\ncases allows inference on the likely source of each infection. We introduce\nsourceR, an R package for quantitative source attribution, aimed at food-borne\ndiseases. It implements a fully joint Bayesian model using strain-typed\nsurveillance data from both human cases and source samples, capable of\nidentifying important sources of infection. The model measures the force of\ninfection from each source, allowing for varying survivability, pathogenicity\nand virulence of pathogen strains, and varying abilities of the sources to act\nas vehicles of infection. A Bayesian non-parametric (Dirichlet process)\napproach is used to cluster pathogen strain types by epidemiological behaviour,\navoiding model overfitting and allowing detection of strain types associated\nwith potentially high 'virulence'.\nsourceR is demonstrated using Campylobacter jejuni isolate data collected in\nNew Zealand between 2005 and 2008. It enables straightforward attribution of\ncases of zoonotic infection to putative sources of infection by epidemiologists\nand public health decision makers. As sourceR develops, we intend it to become\nan important and flexible resource for food-borne disease attribution studies.\n",
"title": "sourceR: Classification and Source Attribution of Infectious Agents among Heterogeneous Populations"
}
| null | null | null | null | true | null |
17319
| null |
Default
| null | null |
null |
{
"abstract": " The exploitation of the excellent intrinsic electronic properties of graphene\nfor device applications is hampered by a large contact resistance between the\nmetal and graphene. The formation of edge contacts rather than top contacts is\none of the most promising solutions for realizing low ohmic contacts. In this\npaper the fabrication and characterization of edge contacts to large area\nCVD-grown monolayer graphene by means of optical lithography using CMOS\ncompatible metals, i.e. Nickel and Aluminum is reported. Extraction of the\ncontact resistance by Transfer Line Method (TLM) as well as the direct\nmeasurement using Kelvin Probe Force Microscopy demonstrates a very low width\nspecific contact resistance.\n",
"title": "Low resistive edge contacts to CVD-grown graphene using a CMOS compatible metal"
}
| null | null | null | null | true | null |
17320
| null |
Default
| null | null |
null |
{
"abstract": " We investigate a steady planar flow of an ideal fluid in a bounded simple\nconnected domain and focus on the vortex patch problem with prescribed\nvorticity strength. There are two methods to deal with the existence of\nsolutions for this problem: the vorticity method and the stream function\nmethod. A long standing open problem is whether these two entirely different\nmethods result in the same solution. In this paper, we will give a positive\nanswer to this problem by studying the local uniqueness of the solutions.\nAnother result obtained in this paper is that if the domain is convex, then the\nvortex patch problem has a unique solution.\n",
"title": "Uniqueness of planar vortex patch in incompressible steady flow"
}
| null | null | null | null | true | null |
17321
| null |
Default
| null | null |
null |
{
"abstract": " This article demonstrates that convolutional operation can be converted to\nmatrix multiplication, which has the same calculation way with fully connected\nlayer. The article is helpful for the beginners of the neural network to\nunderstand how fully connected layer and the convolutional layer work in the\nbackend. To be concise and to make the article more readable, we only consider\nthe linear case. It can be extended to the non-linear case easily through\nplugging in a non-linear encapsulation to the values like this $\\sigma(x)$\ndenoted as $x^{\\prime}$.\n",
"title": "An Equivalence of Fully Connected Layer and Convolutional Layer"
}
| null | null | null | null | true | null |
17322
| null |
Default
| null | null |
null |
{
"abstract": " Due to the success of deep learning to solving a variety of challenging\nmachine learning tasks, there is a rising interest in understanding loss\nfunctions for training neural networks from a theoretical aspect. Particularly,\nthe properties of critical points and the landscape around them are of\nimportance to determine the convergence performance of optimization algorithms.\nIn this paper, we provide full (necessary and sufficient) characterization of\nthe analytical forms for the critical points (as well as global minimizers) of\nthe square loss functions for various neural networks. We show that the\nanalytical forms of the critical points characterize the values of the\ncorresponding loss functions as well as the necessary and sufficient conditions\nto achieve global minimum. Furthermore, we exploit the analytical forms of the\ncritical points to characterize the landscape properties for the loss functions\nof these neural networks. One particular conclusion is that: The loss function\nof linear networks has no spurious local minimum, while the loss function of\none-hidden-layer nonlinear networks with ReLU activation function does have\nlocal minimum that is not global minimum.\n",
"title": "Critical Points of Neural Networks: Analytical Forms and Landscape Properties"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
17323
| null |
Validated
| null | null |
null |
{
"abstract": " Let $R$ be a commutative ring with identity, and let $Z(R)$ be the set of\nzero-divisors of $R$. The annihilator graph of $R$ is defined as the undirected\ngraph $AG(R)$ with the vertex set $Z(R)^*=Z(R)\\setminus\\{0\\}$, and two distinct\nvertices $x$ and $y$ are adjacent if and only if $ann_R(xy)\\neq ann_R(x)\\cup\nann_R(y)$. In this paper, all rings whose annihilator graphs can be embed on\nthe plane or torus are classified.\n",
"title": "When the Annihilator Graph of a Commutative Ring Is Planar or Toroidal?"
}
| null | null | null | null | true | null |
17324
| null |
Default
| null | null |
null |
{
"abstract": " In the following paper we analyse the ID$_3$-Price on German Intraday\nContinuous Electricity Market using an econometric time series model. A\nmultivariate approach is conducted for hourly and quarter-hourly products\nseparately. We estimate the model using lasso and elastic net techniques and\nperform an out-of-sample very short-term forecasting study. The model's\nperformance is compared with benchmark models and is discussed in detail.\nForecasting results provide new insights to the German Intraday Continuous\nElectricity Market regarding its efficiency and to the ID$_3$-Price behaviour.\nThe supplementary materials are available online.\n",
"title": "Econometric modelling and forecasting of intraday electricity prices"
}
| null | null | null | null | true | null |
17325
| null |
Default
| null | null |
null |
{
"abstract": " Characterization of the uncertainty in robotic manipulators is the focus of\nthis paper. Based on the random matrix theory (RMT), we propose uncertainty\ncharacterization schemes in which the uncertainty is modeled at the macro\n(system) level. This is different from the traditional approaches that model\nthe uncertainty in the parametric space of micro (state) level. We show that\nperturbing the system matrices rather than the state of the system provides\nunique advantages especially for robotic manipulators. First, it requires only\nlimited statistical information that becomes effective when dealing with\ncomplex systems where detailed information on their variability is not\navailable. Second, the RMT-based models are aware of the system state and\nconfiguration that are significant factors affecting the level of uncertainty\nin system behavior. In this study, in addition to the motion uncertainty\nanalysis that was first proposed in our earlier work, we also develop an\nRMT-based model for the quantification of the static wrench uncertainty in\nmulti-agent cooperative systems. This model is aimed to be an alternative to\nthe elaborate parametric formulation when only rough bounds are available on\nthe system parameters. We discuss that how RMT-based model becomes advantageous\nwhen the complexity of the system increases. We perform experimental studies on\na KUKA youBot arm to demonstrate the superiority of the RMT-based motion\nuncertainty models. We show that how these models outperform the traditional\nmodels built upon Gaussianity assumption in capturing real-system uncertainty\nand providing accurate bounds on the state estimation errors. In addition, to\nexperimentally support our wrench uncertainty quantification model, we study\nthe behavior of a cooperative system of mobile robots. It is shown that one can\nrely on less demanding RMT-based formulation and yet meets the acceptable\naccuracy.\n",
"title": "Matrix-Based Characterization of the Motion and Wrench Uncertainties in Robotic Manipulators"
}
| null | null | null | null | true | null |
17326
| null |
Default
| null | null |
null |
{
"abstract": " Patch-based denoising algorithms like BM3D have achieved outstanding\nperformance. An important idea for the success of these methods is to exploit\nthe recurrence of similar patches in an input image to estimate the underlying\nimage structures. However, in these algorithms, the similar patches used for\ndenoising are obtained via Nearest Neighbour Search (NNS) and are sometimes not\noptimal. First, due to the existence of noise, NNS can select similar patches\nwith similar noise patterns to the reference patch. Second, the unreliable\nnoisy pixels in digital images can bring a bias to the patch searching process\nand result in a loss of color fidelity in the final denoising result. We\nobserve that given a set of good similar patches, their distribution is not\nnecessarily centered at the noisy reference patch and can be approximated by a\nGaussian component. Based on this observation, we present a patch searching\nmethod that clusters similar patch candidates into patch groups using Gaussian\nMixture Model-based clustering, and selects the patch group that contains the\nreference patch as the final patches for denoising. We also use an unreliable\npixel estimation algorithm to pre-process the input noisy images to further\nimprove the patch searching. Our experiments show that our approach can better\ncapture the underlying patch structures and can consistently enable the\nstate-of-the-art patch-based denoising algorithms, such as BM3D, LPCA and PLOW,\nto better denoise images by providing them with patches found by our approach\nwhile without modifying these algorithms.\n",
"title": "Good Similar Patches for Image Denoising"
}
| null | null | null | null | true | null |
17327
| null |
Default
| null | null |
null |
{
"abstract": " We have studied disordering effects on the coefficients of Ginzburg - Landau\nexpansion in powers of superconducting order - parameter in attractive Anderson\n- Hubbard model within the generalized $DMFT+\\Sigma$ approximation. We consider\nthe wide region of attractive potentials $U$ from the weak coupling region,\nwhere superconductivity is described by BCS model, to the strong coupling\nregion, where superconducting transition is related with Bose - Einstein\ncondensation (BEC) of compact Cooper pairs formed at temperatures essentially\nlarger than the temperature of superconducting transition, and the wide range\nof disorder - from weak to strong, where the system is in the vicinity of\nAnderson transition. In case of semi - elliptic bare density of states disorder\ninfluence upon the coefficients $A$ and $B$ before the square and the fourth\npower of the order - parameter is universal for any value of electron\ncorrelation and is related only to the general disorder widening of the bare\nband (generalized Anderson theorem). Such universality is absent for the\ngradient term expansion coefficient $C$. In the usual theory of \"dirty\"\nsuperconductors the $C$ coefficient drops with the growth of disorder. In the\nlimit of strong disorder in BCS limit the coefficient $C$ is very sensitive to\nthe effects of Anderson localization, which lead to its further drop with\ndisorder growth up to the region of Anderson insulator. In the region of BCS -\nBEC crossover and in BEC limit the coefficient $C$ and all related physical\nproperties are weakly dependent on disorder. In particular, this leads to\nrelatively weak disorder dependence of both penetration depth and coherence\nlengths, as well as of related slope of the upper critical magnetic field at\nsuperconducting transition, in the region of very strong coupling.\n",
"title": "Ginzburg - Landau expansion in strongly disordered attractive Anderson - Hubbard model"
}
| null | null | null | null | true | null |
17328
| null |
Default
| null | null |
null |
{
"abstract": " Simulation-based inference plays a major role in modern statistics, and often\nemploys either reallocating (as in a randomization test) or resampling (as in\nbootstrapping). Reallocating mimics random allocation to treatment groups,\nwhile resampling mimics random sampling from a larger population; does it\nmatter whether the simulation method matches the data collection method?\nMoreover, do the results differ for testing versus estimation? Here we answer\nthese questions in a simple setting by exploring the distribution of a sample\ndifference in means under a basic two group design and four different\nscenarios: true random allocation, true random sampling, reallocating, and\nresampling. For testing a sharp null hypothesis, reallocating is superior in\nsmall samples, but reallocating and resampling are asymptotically equivalent.\nFor estimation, resampling is generally superior, unless the effect is truly\nadditive. Moreover, these results hold regardless of whether the data were\ncollected by random sampling or random allocation.\n",
"title": "Reallocating and Resampling: A Comparison for Inference"
}
| null | null |
[
"Mathematics",
"Statistics"
] | null | true | null |
17329
| null |
Validated
| null | null |
null |
{
"abstract": " K-Nearest Neighbours (k-NN) is a popular classification and regression\nalgorithm, yet one of its main limitations is the difficulty in choosing the\nnumber of neighbours. We present a Bayesian algorithm to compute the posterior\nprobability distribution for k given a target point within a data-set,\nefficiently and without the use of Markov Chain Monte Carlo (MCMC) methods or\nsimulation - alongside an exact solution for distributions within the\nexponential family. The central idea is that data points around our target are\ngenerated by the same probability distribution, extending outwards over the\nappropriate, though unknown, number of neighbours. Once the data is projected\nonto a distance metric of choice, we can transform the choice of k into a\nchange-point detection problem, for which there is an efficient solution: we\nrecursively compute the probability of the last change-point as we move towards\nour target, and thus de facto compute the posterior probability distribution\nover k. Applying this approach to both a classification and a regression UCI\ndata-sets, we compare favourably and, most importantly, by removing the need\nfor simulation, we are able to compute the posterior probability of k exactly\nand rapidly. As an example, the computational time for the Ripley data-set is a\nfew milliseconds compared to a few hours when using a MCMC approach.\n",
"title": "An Efficient Algorithm for Bayesian Nearest Neighbours"
}
| null | null | null | null | true | null |
17330
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we extend the work by Ryuzo Sato devoted to the development of\neconomic growth models within the framework of the Lie group theory. We propose\na new growth model based on the assumption of logistic growth in factors. It is\nemployed to derive new production functions and introduce a new notion of wage\nshare. In the process it is shown that the new functions compare reasonably\nwell against relevant economic data. The corresponding problem of maximization\nof profit under conditions of perfect competition is solved with the aid of one\nof these functions. In addition, it is explained in reasonably rigorous\nmathematical terms why Bowley's law no longer holds true in post-1960 data.\n",
"title": "In search of a new economic model determined by logistic growth"
}
| null | null | null | null | true | null |
17331
| null |
Default
| null | null |
null |
{
"abstract": " We report results of a search for light weakly interacting massive particle\n(WIMP) dark matter from the CDEX-1 experiment at the China Jinping Underground\nLaboratory (CJPL). Constraints on WIMP-nucleon spin-independent (SI) and\nspin-dependent (SD) couplings are derived with a physics threshold of 160 eVee,\nfrom an exposure of 737.1 kg-days. The SI and SD limits extend the lower reach\nof light WIMPs to 2 GeV and improve over our earlier bounds at WIMP mass less\nthan 6 GeV.\n",
"title": "Limits on light WIMPs with a 1 kg-scale germanium detector at 160 eVee physics threshold at the China Jinping Underground Laboratory"
}
| null | null | null | null | true | null |
17332
| null |
Default
| null | null |
null |
{
"abstract": " The 32 Orionis group was discovered almost a decade ago and despite the fact\nthat it represents the first northern, young (age ~ 25 Myr) stellar aggregate\nwithin 100 pc of the Sun ($d \\simeq 93$ pc), a comprehensive survey for members\nand detailed characterisation of the group has yet to be performed. We present\nthe first large-scale spectroscopic survey for new (predominantly M-type)\nmembers of the group after combining kinematic and photometric data to select\ncandidates with Galactic space motion and positions in colour-magnitude space\nconsistent with membership. We identify 30 new members, increasing the number\nof known 32 Ori group members by a factor of three and bringing the total\nnumber of identified members to 46, spanning spectral types B5 to L1. We also\nidentify the lithium depletion boundary (LDB) of the group, i.e. the luminosity\nat which lithium remains unburnt in a coeval population. We estimate the age of\nthe 32 Ori group independently using both isochronal fitting and LDB analyses\nand find it is essentially coeval with the {\\beta} Pictoris moving group, with\nan age of $24\\pm4$ Myr. Finally, we have also searched for circumstellar disc\nhosts utilising the AllWISE catalogue. Although we find no evidence for warm,\ndusty discs, we identify several stars with excess emission in the WISE W4-band\nat 22 {\\mu}m. Based on the limited number of W4 detections we estimate a debris\ndisc fraction of $32^{+12}_{-8}$ per cent for the 32 Ori group.\n",
"title": "A stellar census of the nearby, young 32 Orionis group"
}
| null | null | null | null | true | null |
17333
| null |
Default
| null | null |
null |
{
"abstract": " This paper proposes XML-Defined Network policies (XDNP), a new high-level\nlanguage based on XML notation, to describe network control rules in Software\nDefined Network environments. We rely on existing OpenFlow controllers\nspecifically Floodlight but the novelty of this project is to separate\ncomplicated language- and framework-specific APIs from policy descriptions.\nThis separation makes it possible to extend the current work as a northbound\nhigher level abstraction that can support a wide range of controllers who are\nbased on different programming languages. By this approach, we believe that\nnetwork administrators can develop and deploy network control policies easier\nand faster.\n",
"title": "A High-Level Rule-based Language for Software Defined Network Programming based on OpenFlow"
}
| null | null | null | null | true | null |
17334
| null |
Default
| null | null |
null |
{
"abstract": " This paper discusses a roadmap to investigate Domain Objects being an\nadequate formalism to capture the peculiarity of microservice architecture, and\nto support Software development since the early stages. It provides a survey of\nboth Microservices and Domain Objects, and it discusses plans and reflections\non how to investigate whether a modeling approach suited to adaptable\nservice-based components can also be applied with success to the microservice\nscenario.\n",
"title": "Domain Objects and Microservices for Systems Development: a roadmap"
}
| null | null |
[
"Computer Science"
] | null | true | null |
17335
| null |
Validated
| null | null |
null |
{
"abstract": " We discuss the effect of dissipation on heating which occurs in periodically\ndriven quantum many body systems. We especially focus on a periodically driven\nBose-Hubbard model coupled to an energy and particle reservoir. Without\ndissipation, this model is known to undergo parametric instabilities which can\nbe considered as an initial stage of heating. By taking the weak on-site\ninteraction limit as well as the weak system-reservoir coupling limit, we find\nthat parametric instabilities are suppressed if the dissipation is stronger\nthan the on-site interaction strength and stable steady states appear. Our\nresults demonstrate that periodically-driven systems can emit energy, which is\nabsorbed from external drivings, to the reservoir so that they can avoid\nheating.\n",
"title": "Stabilization of prethermal Floquet steady states in a periodically driven dissipative Bose-Hubbard model"
}
| null | null | null | null | true | null |
17336
| null |
Default
| null | null |
null |
{
"abstract": " The goal of compressed sensing is to estimate a vector from an\nunderdetermined system of noisy linear measurements, by making use of prior\nknowledge on the structure of vectors in the relevant domain. For almost all\nresults in this literature, the structure is represented by sparsity in a\nwell-chosen basis. We show how to achieve guarantees similar to standard\ncompressed sensing but without employing sparsity at all. Instead, we suppose\nthat vectors lie near the range of a generative model $G: \\mathbb{R}^k \\to\n\\mathbb{R}^n$. Our main theorem is that, if $G$ is $L$-Lipschitz, then roughly\n$O(k \\log L)$ random Gaussian measurements suffice for an $\\ell_2/\\ell_2$\nrecovery guarantee. We demonstrate our results using generative models from\npublished variational autoencoder and generative adversarial networks. Our\nmethod can use $5$-$10$x fewer measurements than Lasso for the same accuracy.\n",
"title": "Compressed Sensing using Generative Models"
}
| null | null | null | null | true | null |
17337
| null |
Default
| null | null |
null |
{
"abstract": " Several researchers have described two-part models with patient-specific\nstochastic processes for analysing longitudinal semicontinuous data. In theory,\nsuch models can offer greater flexibility than the standard two-part model with\npatient-specific random effects. However, in practice the high dimensional\nintegrations involved in the marginal likelihood (i.e. integrated over the\nstochastic processes) significantly complicates model fitting. Thus\nnon-standard computationally intensive procedures based on simulating the\nmarginal likelihood have so far only been proposed. In this paper, we describe\nan efficient method of implementation by demonstrating how the high dimensional\nintegrations involved in the marginal likelihood can be computed efficiently.\nSpecifically, by using a property of the multivariate normal distribution and\nthe standard marginal cumulative distribution function identity, we transform\nthe marginal likelihood so that the high dimensional integrations are contained\nin the cumulative distribution function of a multivariate normal distribution,\nwhich can then be efficiently evaluated. Hence maximum likelihood estimation\ncan be used to obtain parameter estimates and asymptotic standard errors (from\nthe observed information matrix) of model parameters. We describe our proposed\nefficient implementation procedure for the standard two-part model\nparameterisation and when it is of interest to directly model the overall\nmarginal mean. The methodology is applied on a psoriatic arthritis data set\nconcerning functional disability.\n",
"title": "Two-part models with stochastic processes for modelling longitudinal semicontinuous data: computationally efficient inference and modelling the overall marginal mean"
}
| null | null | null | null | true | null |
17338
| null |
Default
| null | null |
null |
{
"abstract": " Along with the deraining performance improvement of deep networks, their\nstructures and learning become more and more complicated and diverse, making it\ndifficult to analyze the contribution of various network modules when\ndeveloping new deraining networks. To handle this issue, this paper provides a\nbetter and simpler baseline deraining network by considering network\narchitecture, input and output, and loss functions. Specifically, by repeatedly\nunfolding a shallow ResNet, progressive ResNet (PRN) is proposed to take\nadvantage of recursive computation. A recurrent layer is further introduced to\nexploit the dependencies of deep features across stages, forming our\nprogressive recurrent network (PReNet). Furthermore, intra-stage recursive\ncomputation of ResNet can be adopted in PRN and PReNet to notably reduce\nnetwork parameters with graceful degradation in deraining performance. For\nnetwork input and output, we take both stage-wise result and original rainy\nimage as input to each ResNet and finally output the prediction of {residual\nimage}. As for loss functions, single MSE or negative SSIM losses are\nsufficient to train PRN and PReNet. Experiments show that PRN and PReNet\nperform favorably on both synthetic and real rainy images. Considering its\nsimplicity, efficiency and effectiveness, our models are expected to serve as a\nsuitable baseline in future deraining research. The source codes are available\nat this https URL.\n",
"title": "Progressive Image Deraining Networks: A Better and Simpler Baseline"
}
| null | null | null | null | true | null |
17339
| null |
Default
| null | null |
null |
{
"abstract": " Statistical inference based on lossy or incomplete samples is of fundamental\nimportance in research areas such as signal/image processing, medical image\nstorage, remote sensing, signal transmission. In this paper, we propose a\nnonparametric testing procedure based on quantized samples. In contrast to the\nclassic nonparametric approach, our method lives on a coarse grid of sample\ninformation and are simple-to-use. Under mild technical conditions, we\nestablish the asymptotic properties of the proposed procedures including\nasymptotic null distribution of the quantization test statistic as well as its\nminimax power optimality. Concrete quantizers are constructed for achieving the\nminimax optimality in practical use. Simulation results and a real data\nanalysis are provided to demonstrate the validity and effectiveness of the\nproposed test. Our work bridges the classical nonparametric inference to modern\nlossy data setting.\n",
"title": "Optimal Nonparametric Inference under Quantization"
}
| null | null | null | null | true | null |
17340
| null |
Default
| null | null |
null |
{
"abstract": " Nearest neighbor imputation is popular for handling item nonresponse in\nsurvey sampling. In this article, we study the asymptotic properties of the\nnearest neighbor imputation estimator for general population parameters,\nincluding population means, proportions and quantiles. For variance estimation,\nthe conventional bootstrap inference for matching estimators with fixed number\nof matches has been shown to be invalid due to the nonsmoothness nature of the\nmatching estimator. We propose asymptotically valid replication variance\nestimation. The key strategy is to construct replicates of the estimator\ndirectly based on linear terms, instead of individual records of variables. A\nsimulation study confirms that the new procedure provides valid variance\nestimation.\n",
"title": "Nearest neighbor imputation for general parameter estimation in survey sampling"
}
| null | null |
[
"Statistics"
] | null | true | null |
17341
| null |
Validated
| null | null |
null |
{
"abstract": " We demonstrate that a semiconductor laser perturbed by the distributed\nfeedback from a fiber random grating can emit light chaotically without the\ntime delay signature. A theoretical model is developed based on the\nLang-Kobayashi model in order to numerically explore the chaotic dynamics of\nthe laser diode subjected to the random distributed feedback. It is predicted\nthat the random distributed feedback is superior to the single reflection\nfeedback in suppressing the time-delay signature. In experiments, a massive\nnumber of feedbacks with randomly varied time delays induced by a fiber random\ngrating introduce large numbers of external cavity modes into the semiconductor\nlaser, leading to the high dimension of chaotic dynamics and thus the\nconcealment of the time delay signature. The obtained time delay signature with\nthe maximum suppression is 0.0088, which is the smallest to date.\n",
"title": "Time-delay signature suppression in a chaotic semiconductor laser by fiber random grating induced distributed feedback"
}
| null | null | null | null | true | null |
17342
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we propose a new deep feature selection method based on deep\narchitecture. Our method uses stacked auto-encoders for feature representation\nin higher-level abstraction. We developed and applied a novel feature learning\napproach to a specific precision medicine problem, which focuses on assessing\nand prioritizing risk factors for hypertension (HTN) in a vulnerable\ndemographic subgroup (African-American). Our approach is to use deep learning\nto identify significant risk factors affecting left ventricular mass indexed to\nbody surface area (LVMI) as an indicator of heart damage risk. The results show\nthat our feature learning and representation approach leads to better results\nin comparison with others.\n",
"title": "SAFS: A Deep Feature Selection Approach for Precision Medicine"
}
| null | null | null | null | true | null |
17343
| null |
Default
| null | null |
null |
{
"abstract": " To detect and segment salient objects accurately, existing methods are\nusually devoted to designing complex network architectures to fuse powerful\nfeatures from the backbone networks. However, they put much less efforts on the\nsaliency inference module and only use a few fully convolutional layers to\nperform saliency reasoning from the fused features. However, should feature\nfusion strategies receive much attention but saliency reasoning be ignored a\nlot? In this paper, we find that weakness of the saliency reasoning unit limits\nsalient object detection performance, and claim that saliency reasoning after\nmulti-scale convolutional features fusion is critical. To verify our findings,\nwe first extract multi-scale features with a fully convolutional network, and\nthen directly reason from these comprehensive features using a deep yet\nlight-weighted network, modified from ShuffleNet, to fast and precisely predict\nsalient objects. Such simple design is shown to be capable of reasoning from\nmulti-scale saliency features as well as giving superior saliency detection\nperformance with less computation cost. Experimental results show that our\nsimple framework outperforms the best existing method with 2.3\\% and 3.6\\%\npromotion for F-measure scores, 2.8\\% reduction for MAE score on PASCAL-S,\nDUT-OMRON and SOD datasets respectively.\n",
"title": "Deep Reasoning with Multi-scale Context for Salient Object Detection"
}
| null | null | null | null | true | null |
17344
| null |
Default
| null | null |
null |
{
"abstract": " We provide a complete picture of asymptotically minimax estimation of\n$L_r$-norms (for any $r\\ge 1$) of the mean in Gaussian white noise model over\nNikolskii-Besov spaces. In this regard, we complement the work of Lepski,\nNemirovski and Spokoiny (1999), who considered the cases of $r=1$ (with\npoly-logarithmic gap between upper and lower bounds) and $r$ even (with\nasymptotically sharp upper and lower bounds) over Hölder spaces. We\nadditionally consider the case of asymptotically adaptive minimax estimation\nand demonstrate a difference between even and non-even $r$ in terms of an\ninvestigator's ability to produce asymptotically adaptive minimax estimators\nwithout paying a penalty.\n",
"title": "On Estimation of $L_{r}$-Norms in Gaussian White Noise Models"
}
| null | null | null | null | true | null |
17345
| null |
Default
| null | null |
null |
{
"abstract": " This paper studies the secrecy rate maximization problem of a secure wireless\ncommunication system, in the presence of multiple eavesdroppers. The security\nof the communication link is enhanced through cooperative jamming, with the\nhelp of multiple jammers. First, a feasibility condition is derived to achieve\na positive secrecy rate at the destination. Then, we solve the original secrecy\nrate maximization problem, which is not convex in terms of power allocation at\nthe jammers. To circumvent this non-convexity, the achievable secrecy rate is\napproximated for a given power allocation at the jammers and the approximated\nproblem is formulated into a geometric programming one. Based on this\napproximation, an iterative algorithm has been developed to obtain the optimal\npower allocation at the jammers. Next, we provide a bisection approach, based\non one-dimensional search, to validate the optimality of the proposed\nalgorithm. In addition, by assuming Rayleigh fading, the secrecy outage\nprobability (SOP) of the proposed cooperative jamming scheme is analyzed. More\nspecifically, a single-integral form expression for SOP is derived for the most\ngeneral case as well as a closed-form expression for the special case of two\ncooperative jammers and one eavesdropper. Simulation results have been provided\nto validate the convergence and the optimality of the proposed algorithm as\nwell as the theoretical derivations of the presented SOP analysis.\n",
"title": "Secure communications with cooperative jamming: Optimal power allocation and secrecy outage analysis"
}
| null | null | null | null | true | null |
17346
| null |
Default
| null | null |
null |
{
"abstract": " Stochastic integration \\textit{wrt} Gaussian processes has raised strong\ninterest in recent years, motivated in particular by its applications in\nInternet traffic modeling, biomedicine and finance. The aim of this work is to\ndefine and develop a White Noise Theory-based anticipative stochastic calculus\nwith respect to all Gaussian processes that have an integral representation\nover a real (maybe infinite) interval. Very rich, this class of Gaussian\nprocesses contains, among many others, Volterra processes (and thus fractional\nBrownian motion) as well as processes the regularity of which varies along the\ntime (such as multifractional Brownian motion).A systematic comparison of the\nstochastic calculus (including It{ô} formula) we provide here, to the ones\ngiven by Malliavin calculus in\n\\cite{nualart,MV05,NuTa06,KRT07,KrRu10,LN12,SoVi14,LN12}, and by It{ô}\nstochastic calculus is also made. Not only our stochastic calculus fully\ngeneralizes and extends the ones originally proposed in \\cite{MV05} and in\n\\cite{NuTa06} for Gaussian processes, but also the ones proposed in\n\\cite{ell,bosw,ben1} for fractional Brownian motion (\\textit{resp.} in\n\\cite{JLJLV1,JL13,LLVH} for multifractional Brownian motion).\n",
"title": "Stochastic Calculus with respect to Gaussian Processes: Part I"
}
| null | null | null | null | true | null |
17347
| null |
Default
| null | null |
null |
{
"abstract": " We naturally associate a measurable space of paths to a couple of orthogonal\nvector fields over a surface and we integrate the length function over it. This\nintegral is interpreted as a natural continuous generalization of indirect\ninfluences on finite graphs and can be thought as a tool to capture geometric\ninformation of the surface. As a byproduct we calculate volumes in different\nexamples of spaces of paths.\n",
"title": "Path-like integrals of lenght on surfaces of constant curvature"
}
| null | null | null | null | true | null |
17348
| null |
Default
| null | null |
null |
{
"abstract": " This paper focuses on automated synthesis of divide-and-conquer parallelism,\nwhich is a common parallel programming skeleton supported by many\ncross-platform multithreaded libraries. The challenges of producing (manually\nor automatically) a correct divide-and-conquer parallel program from a given\nsequential code are two-fold: (1) assuming that individual worker threads\nexecute a code identical to the sequential code, the programmer has to provide\nthe extra code for dividing the tasks and combining the computation results,\nand (2) sometimes, the sequential code may not be usable as is, and may need to\nbe modified by the programmer. We address both challenges in this paper. We\npresent an automated synthesis technique for the case where no modifications to\nthe sequential code are required, and we propose an algorithm for modifying the\nsequential code to make it suitable for parallelization when some modification\nis necessary. The paper presents theoretical results for when this {\\em\nmodification} is efficiently possible, and experimental evaluation of the\ntechnique and the quality of the produced parallel programs.\n",
"title": "Automated Synthesis of Divide and Conquer Parallelism"
}
| null | null | null | null | true | null |
17349
| null |
Default
| null | null |
null |
{
"abstract": " In the present work we prove a Nikol'ski inequality for trigonometric\npolynomials and Ul'yanov type inequalities for functions in Lebesgue spaces\nwith Muckenhoupt weights. Realization result and Jackson inequalities are\nobtained. Simultaneous approximation by polynomials is considered. Some uniform\nnorm inequalities are transferred to weighted Lebesgue space.\n",
"title": "Nikol'ski\\uı, Jackson and Ul'yanov type inequalities with Muckenhoupt weights"
}
| null | null | null | null | true | null |
17350
| null |
Default
| null | null |
null |
{
"abstract": " Inferring model parameters from experimental data is a grand challenge in\nmany sciences, including cosmology. This often relies critically on high\nfidelity numerical simulations, which are prohibitively computationally\nexpensive. The application of deep learning techniques to generative modeling\nis renewing interest in using high dimensional density estimators as\ncomputationally inexpensive emulators of fully-fledged simulations. These\ngenerative models have the potential to make a dramatic shift in the field of\nscientific simulations, but for that shift to happen we need to study the\nperformance of such generators in the precision regime needed for science\napplications. To this end, in this work we apply Generative Adversarial\nNetworks to the problem of generating weak lensing convergence maps. We show\nthat our generator network produces maps that are described by, with high\nstatistical confidence, the same summary statistics as the fully simulated\nmaps.\n",
"title": "CosmoGAN: creating high-fidelity weak lensing convergence maps using Generative Adversarial Networks"
}
| null | null | null | null | true | null |
17351
| null |
Default
| null | null |
null |
{
"abstract": " This paper establishes an upper bound for the Kolmogorov distance between the\nmaximum of a high-dimensional vector of smooth Wiener functionals and the\nmaximum of a Gaussian random vector. As a special case, we show that the\nmaximum of multiple Wiener-Itô integrals with common orders is\nwell-approximated by its Gaussian analog in terms of the Kolmogorov distance if\ntheir covariance matrices are close to each other and the maximum of the fourth\ncumulants of the multiple Wiener-Itô integrals is close to zero. This may be\nviewed as a new kind of fourth moment phenomenon, which has attracted\nconsiderable attention in the recent studies of probability. This type of\nGaussian approximation result has many potential applications to statistics. To\nillustrate this point, we present two statistical applications in\nhigh-frequency financial econometrics: One is the hypothesis testing problem\nfor the absence of lead-lag effects and the other is the construction of\nuniform confidence bands for spot volatility.\n",
"title": "Gaussian approximation of maxima of Wiener functionals and its application to high-frequency data"
}
| null | null | null | null | true | null |
17352
| null |
Default
| null | null |
null |
{
"abstract": " By considering a limiting case of a Kronecker-type identity, we obtain an\nidentity found by both Andrews and Crandall. We then use the Andrews-Crandall\nidentity to give a new proof of a formula of Gauss for the representations of a\nnumber as a sum of three squares. From the Kronecker-type identity, we also\ndeduce Gauss's theorem that every positive integer is representable as a sum of\nthree triangular numbers.\n",
"title": "A Kronecker-type identity and the representations of a number as a sum of three squares"
}
| null | null | null | null | true | null |
17353
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we consider the temporal pattern in traffic flow time series,\nand implement a deep learning model for traffic flow prediction. Detrending\nbased methods decompose original flow series into trend and residual series, in\nwhich trend describes the fixed temporal pattern in traffic flow and residual\nseries is used for prediction. Inspired by the detrending method, we propose\nDeepTrend, a deep hierarchical neural network used for traffic flow prediction\nwhich considers and extracts the time-variant trend. DeepTrend has two stacked\nlayers: extraction layer and prediction layer. Extraction layer, a fully\nconnected layer, is used to extract the time-variant trend in traffic flow by\nfeeding the original flow series concatenated with corresponding simple average\ntrend series. Prediction layer, an LSTM layer, is used to make flow prediction\nby feeding the obtained trend from the output of extraction layer and\ncalculated residual series. To make the model more effective, DeepTrend needs\nfirst pre-trained layer-by-layer and then fine-tuned in the entire network.\nExperiments show that DeepTrend can noticeably boost the prediction performance\ncompared with some traditional prediction models and LSTM with detrending based\nmethods.\n",
"title": "DeepTrend: A Deep Hierarchical Neural Network for Traffic Flow Prediction"
}
| null | null | null | null | true | null |
17354
| null |
Default
| null | null |
null |
{
"abstract": " We propose in this paper a new approach to the Kaluza-Klein idea of a five\ndimensional space-time unifying gravitation and electromagnetism, and extension\nto higher-dimensional space-time. By considering a natural geometric definition\nof a matter fluid and abandoning the usual requirement of a Ricci-flat five\ndimensional space-time, we show that a unified geometrical frame can be set for\ngravitation and electromagnetism, giving, by projection on the classical\n4-dimensional space-time, the known Einstein-Maxwell-Lorentz equations for\ncharged fluids. Thus, although not introducing new physics, we get a very\naesthetic presentation of classical physics in the spirit of general\nrelativity. The usual physical concepts, such as mass, energy, charge,\ntrajectory, Maxwell-Lorentz law, are shown to be only various aspects of the\ngeometry, for example curvature, of space-time considered as a Lorentzian\nmanifold; that is no physical objects are introduced in space-time, no laws are\ngiven, everything is only geometry.\nWe then extend these ideas to more than 5 dimensions, by considering\nspacetime as a generalization of a $(S^1\\times W)$-fiber bundle, that we named\nmulti-fibers bundle, where $S^1$ is the circle and $W$ a compact manifold. We\nwill use this geometric structure as a possible way to model or encode\ndeviations from standard 4-dimensional General Relativity, or \"dark\" effects\nsuch as dark matter or energy.\n",
"title": "A new approach to Kaluza-Klein Theory"
}
| null | null |
[
"Mathematics"
] | null | true | null |
17355
| null |
Validated
| null | null |
null |
{
"abstract": " We prove a conjecture of Medvedev and Scanlon in the case of regular\nmorphisms of semiabelian varieties. That is, if $G$ is a semiabelian variety\ndefined over an algebraically closed field $K$ of characteristic $0$, and\n$\\varphi\\colon G\\to G$ is a dominant regular self-map of $G$ which is not\nnecessarily a group homomorphism, we prove that one of the following holds:\neither there exists a non-constant rational fibration preserved by $\\varphi$,\nor there exists a point $x\\in G(K)$ whose $\\varphi$-orbit is Zariski dense in\n$G$.\n",
"title": "Density of orbits of dominant regular self-maps of semiabelian varieties"
}
| null | null | null | null | true | null |
17356
| null |
Default
| null | null |
null |
{
"abstract": " The asymptotic behaviour of the commonly used bootstrap percentile confidence\ninterval is investigated when the parameters are subject to linear inequality\nconstraints. We concentrate on the important one- and two-sample problems with\ndata generated from general parametric distributions in the natural exponential\nfamily. The focus of this paper is on quantifying the coverage probabilities of\nthe parametric bootstrap percentile confidence intervals, in particular their\nlimiting behaviour near boundaries. We propose a local asymptotic framework to\nstudy this subtle coverage behaviour. Under this framework, we discover that\nwhen the true parameters are on, or close to, the restriction boundary, the\nasymptotic coverage probabilities can always exceed the nominal level in the\none-sample case; however, they can be, remarkably, both under and over the\nnominal level in the two-sample case. Using illustrative examples, we show that\nthe results provide theoretical justification and guidance on applying the\nbootstrap percentile method to constrained inference problems.\n",
"title": "Asymptotic coverage probabilities of bootstrap percentile confidence intervals for constrained parameters"
}
| null | null | null | null | true | null |
17357
| null |
Default
| null | null |
null |
{
"abstract": " We compute physical properties across the phase diagram of the $t$-$J_\\perp$\nchain with long-range dipolar interactions, which describe ultracold polar\nmolecules on optical lattices. Our results obtained by the density-matrix\nrenormalization group (DMRG) indicate that superconductivity is enhanced when\nthe Ising component $J_z$ of the spin-spin interaction and the charge component\n$V$ are tuned to zero, and even further by the long-range dipolar interactions.\nAt low densities, a substantially larger spin gap is obtained. We provide\nevidence that long-range interactions lead to algebraically decaying\ncorrelation functions despite the presence of a gap. Although this has recently\nbeen observed in other long-range interacting spin and fermion models, the\ncorrelations in our case have the peculiar property of having a small and\ncontinuously varying exponent. We construct simple analytic models and\narguments to understand the most salient features.\n",
"title": "Correlations and enlarged superconducting phase of $t$-$J_\\perp$ chains of ultracold molecules on optical lattices"
}
| null | null | null | null | true | null |
17358
| null |
Default
| null | null |
null |
{
"abstract": " We introduce MinimalRNN, a new recurrent neural network architecture that\nachieves comparable performance as the popular gated RNNs with a simplified\nstructure. It employs minimal updates within RNN, which not only leads to\nefficient learning and testing but more importantly better interpretability and\ntrainability. We demonstrate that by endorsing the more restrictive update\nrule, MinimalRNN learns disentangled RNN states. We further examine the\nlearning dynamics of different RNN structures using input-output Jacobians, and\nshow that MinimalRNN is able to capture longer range dependencies than existing\nRNN architectures.\n",
"title": "MinimalRNN: Toward More Interpretable and Trainable Recurrent Neural Networks"
}
| null | null | null | null | true | null |
17359
| null |
Default
| null | null |
null |
{
"abstract": " Let $BQP(n)$ be a boolean quadric polytope, $LOP(m)$ be a linear ordering\npolytope. It is shown that $BQP(n)$ is linearly isomorphic to a face of\n$LOP(2n)$.\n",
"title": "Boolean quadric polytopes are faces of linear ordering polytopes"
}
| null | null | null | null | true | null |
17360
| null |
Default
| null | null |
null |
{
"abstract": " Analyzing array-based computations to determine data dependences is useful\nfor many applications including automatic parallelization, race detection,\ncomputation and communication overlap, verification, and shape analysis. For\nsparse matrix codes, array data dependence analysis is made more difficult by\nthe use of index arrays that make it possible to store only the nonzero entries\nof the matrix (e.g., in A[B[i]], B is an index array). Here, dependence\nanalysis is often stymied by such indirect array accesses due to the values of\nthe index array not being available at compile time. Consequently, many\ndependences cannot be proven unsatisfiable or determined until runtime.\nNonetheless, index arrays in sparse matrix codes often have properties such as\nmonotonicity of index array elements that can be exploited to reduce the amount\nof runtime analysis needed. In this paper, we contribute a formulation of array\ndata dependence analysis that includes encoding index array properties as\nuniversally quantified constraints. This makes it possible to leverage existing\nSMT solvers to determine whether such dependences are unsatisfiable and\nsignificantly reduces the number of dependences that require runtime analysis\nin a set of eight sparse matrix kernels. Another contribution is an algorithm\nfor simplifying the remaining satisfiable data dependences by discovering\nequalities and/or subset relationships. These simplifications are essential to\nmake a runtime-inspection-based approach feasible.\n",
"title": "Sparse Matrix Code Dependence Analysis Simplification at Compile Time"
}
| null | null | null | null | true | null |
17361
| null |
Default
| null | null |
null |
{
"abstract": " Independent Component Analysis (ICA) - one of the basic tools in data\nanalysis - aims to find a coordinate system in which the components of the data\nare independent. Most of existing methods are based on the minimization of the\nfunction of fourth-order moment (kurtosis). Skewness (third-order moment) has\nreceived much less attention.\nIn this paper we present a competitive approach to ICA based on the Split\nGaussian distribution, which is well adapted to asymmetric data. Consequently,\nwe obtain a method which works better than the classical approaches, especially\nin the case when the underlying density is not symmetric, which is a typical\nsituation in the color distribution in images.\n",
"title": "ICA based on the data asymmetry"
}
| null | null | null | null | true | null |
17362
| null |
Default
| null | null |
null |
{
"abstract": " We study weighted $H^\\infty$ spaces of analytic functions on the open unit\ndisc in the case of non-doubling weights, which decrease rapidly with respect\nto the boundary distance. We characterize the solid hulls of such spaces and\ngive quite explicit representations of them in the case of the most natural\nexponentially decreasing weights.\n",
"title": "Solid hulls of weighted Banach spaces of analytic functions on the unit disc with exponential weights"
}
| null | null | null | null | true | null |
17363
| null |
Default
| null | null |
null |
{
"abstract": " Cauchy and exponential transforms are characterized, and constructed, as\ncanonical holomorphic sections of certain line bundles on the Riemann sphere\ndefined in terms of the Schwarz function. A well known natural connection\nbetween Schwarz reflection and line bundles defined on the Schottky double of a\nplanar domain is briefly discussed in the same context.\n",
"title": "Line bundles defined by the Schwarz function"
}
| null | null | null | null | true | null |
17364
| null |
Default
| null | null |
null |
{
"abstract": " We report extensive theoretical calculations on the rotation-inversion\nexcitation of interstellar ammonia (NH3) due to collisions with atomic and\nmolecular hydrogen (both para- and ortho-H2). Close-coupling calculations are\nperformed for total energies in the range 1-2000 cm-1 and rotational cross\nsections are obtained for all transitions among the lowest 17 and 34\nrotation-inversion levels of ortho- and para-NH3, respectively. Rate\ncoefficients are deduced for kinetic temperatures up to 200 K. Propensity rules\nfor the three colliding partners are discussed and we also compare the new\nresults to previous calculations for the spherically symmetrical He and para-H2\nprojectiles. Significant differences are found between the different sets of\ncalculations. Finally, we test the impact of the new rate coefficients on the\ncalibration of the ammonia thermometer. We find that the calibration curve is\nonly weakly sensitive to the colliding partner and we confirm that the ammonia\nthermometer is robust.\n",
"title": "Collisional excitation of NH3 by atomic and molecular hydrogen"
}
| null | null |
[
"Physics"
] | null | true | null |
17365
| null |
Validated
| null | null |
null |
{
"abstract": " We consider the multi-view data completion problem, i.e., to complete a\nmatrix $\\mathbf{U}=[\\mathbf{U}_1|\\mathbf{U}_2]$ where the ranks of\n$\\mathbf{U},\\mathbf{U}_1$, and $\\mathbf{U}_2$ are given. In particular, we\ninvestigate the fundamental conditions on the sampling pattern, i.e., locations\nof the sampled entries for finite completability of such a multi-view data\ngiven the corresponding rank constraints. In contrast with the existing\nanalysis on Grassmannian manifold for a single-view matrix, i.e., conventional\nmatrix completion, we propose a geometric analysis on the manifold structure\nfor multi-view data to incorporate more than one rank constraint. We provide a\ndeterministic necessary and sufficient condition on the sampling pattern for\nfinite completability. We also give a probabilistic condition in terms of the\nnumber of samples per column that guarantees finite completability with high\nprobability. Finally, using the developed tools, we derive the deterministic\nand probabilistic guarantees for unique completability.\n",
"title": "Deterministic and Probabilistic Conditions for Finite Completability of Low-rank Multi-View Data"
}
| null | null |
[
"Computer Science",
"Mathematics"
] | null | true | null |
17366
| null |
Validated
| null | null |
null |
{
"abstract": " We consider the problem of grid-forming control of power converters in\nlow-inertia power systems. Starting from an average-switch three-phase inverter\nmodel, we draw parallels to a synchronous machine (SM) model and propose a\nnovel grid-forming converter control strategy which dwells upon the main\ncharacteristic of a SM: the presence of an internal rotating magnetic field. In\nparticular, we augment the converter system with a virtual oscillator whose\nfrequency is driven by the DC-side voltage measurement and which sets the\nconverter pulse-width-modulation signal, thereby achieving exact matching\nbetween the converter in closed-loop and the SM dynamics. We then provide a\nsufficient condition assuring existence, uniqueness, and global asymptotic\nstability of equilibria in a coordinate frame attached to the virtual\noscillator angle. By actuating the DC-side input of the converter we are able\nto enforce this sufficient condition. In the same setting, we highlight strict\nincremental passivity, droop, and power-sharing properties of the proposed\nframework, which are compatible with conventional requirements of power system\noperation. We subsequently adopt disturbance decoupling techniques to design\nadditional control loops that regulate the DC-side voltage, as well as AC-side\nfrequency and amplitude, while in the end validating them with numerical\nexperiments.\n",
"title": "Grid-forming Control for Power Converters based on Matching of Synchronous Machines"
}
| null | null | null | null | true | null |
17367
| null |
Default
| null | null |
null |
{
"abstract": " We characterize the near-infrared (NIR) dust attenuation for a sample of\n~5500 local (z<0.1) star-forming galaxies and obtain an estimate of their\naverage total-to-selective attenuation $k(\\lambda)$. We utilize data from the\nUnited Kingdom Infrared Telescope (UKIRT) and the Two Micron All-Sky Survey\n(2MASS), which is combined with previously measured UV-optical data for these\ngalaxies. The average attenuation curve is slightly lower in the far-UV than\nlocal starburst galaxies, by roughly 15%, but appears similar at longer\nwavelengths with a total-to-selective normalization at V-band of\n$R_V=3.67\\substack{+0.44 \\\\ -0.35}$. Under the assumption of energy balance,\nthe total attenuated energy inferred from this curve is found to be broadly\nconsistent with the observed infrared dust emission ($L_{\\rm{TIR}}$) in a small\nsample of local galaxies for which far-IR measurements are available. However,\nthe significant scatter in this quantity among the sample may reflect large\nvariations in the attenuation properties of individual galaxies. We also derive\nthe attenuation curve for sub-populations of the main sample, separated\naccording to mean stellar population age (via $D_n4000$), specific star\nformation rate, stellar mass, and metallicity, and find that they show only\ntentative trends with low significance, at least over the range which is probed\nby our sample. These results indicate that a single curve is reasonable for\napplications seeking to broadly characterize large samples of galaxies in the\nlocal Universe, while applications to individual galaxies would yield large\nuncertainties and is not recommended.\n",
"title": "Characterizing Dust Attenuation in Local Star-Forming Galaxies: Near-Infrared Reddening and Normalization"
}
| null | null | null | null | true | null |
17368
| null |
Default
| null | null |
null |
{
"abstract": " Using tape or optical devices for scale-out storage is one option for storing\na vast amount of data. However, it is impossible or almost impossible to\nrewrite data with such devices. Thus, scale-out storage using such devices\ncannot use standard data-distribution algorithms because they rewrite data for\nmoving between servers constituting the scale-out storage when the server\nconfiguration is changed. Although using rewritable devices for scale-out\nstorage, when server capacity is huge, rewriting data is very hard when server\nconstitution is changed. In this paper, a data-distribution algorithm called\nSequential Checking is proposed, which can be used for scale-out storage\ncomposed of devices that are hardly able to rewrite data. Sequential Checking\n1) does not need to move data between servers when the server configuration is\nchanged, 2) distribute data, the amount of which depends on the server's\nvolume, 3) select a unique server when datum is written, and 4) select servers\nwhen datum is read (there are few such server(s) in most cases) and find out a\nunique server that stores the newest datum from them. These basic\ncharacteristics were confirmed through proofs and simulations. Data can be read\nby accessing 1.98 servers on average from a storage comprising 256 servers\nunder a realistic condition. And it is confirmed by evaluations in real\nenvironment that access time is acceptable. Sequential Checking makes selecting\nscale-out storage using tape or optical devices or using huge capacity servers\nrealistic.\n",
"title": "Sequential Checking: Reallocation-Free Data-Distribution Algorithm for Scale-out Storage"
}
| null | null | null | null | true | null |
17369
| null |
Default
| null | null |
null |
{
"abstract": " Noisy PN learning is the problem of binary classification when training\nexamples may be mislabeled (flipped) uniformly with noise rate rho1 for\npositive examples and rho0 for negative examples. We propose Rank Pruning (RP)\nto solve noisy PN learning and the open problem of estimating the noise rates,\ni.e. the fraction of wrong positive and negative labels. Unlike prior\nsolutions, RP is time-efficient and general, requiring O(T) for any\nunrestricted choice of probabilistic classifier with T fitting time. We prove\nRP has consistent noise estimation and equivalent expected risk as learning\nwith uncorrupted labels in ideal conditions, and derive closed-form solutions\nwhen conditions are non-ideal. RP achieves state-of-the-art noise estimation\nand F1, error, and AUC-PR for both MNIST and CIFAR datasets, regardless of the\namount of noise and performs similarly impressively when a large portion of\ntraining examples are noise drawn from a third distribution. To highlight, RP\nwith a CNN classifier can predict if an MNIST digit is a \"one\"or \"not\" with\nonly 0.25% error, and 0.46 error across all digits, even when 50% of positive\nexamples are mislabeled and 50% of observed positive labels are mislabeled\nnegative examples.\n",
"title": "Learning with Confident Examples: Rank Pruning for Robust Classification with Noisy Labels"
}
| null | null | null | null | true | null |
17370
| null |
Default
| null | null |
null |
{
"abstract": " We present a neural model for representing snippets of code as continuous\ndistributed vectors (\"code embeddings\"). The main idea is to represent a code\nsnippet as a single fixed-length $\\textit{code vector}$, which can be used to\npredict semantic properties of the snippet. This is performed by decomposing\ncode to a collection of paths in its abstract syntax tree, and learning the\natomic representation of each path $\\textit{simultaneously}$ with learning how\nto aggregate a set of them. We demonstrate the effectiveness of our approach by\nusing it to predict a method's name from the vector representation of its body.\nWe evaluate our approach by training a model on a dataset of 14M methods. We\nshow that code vectors trained on this dataset can predict method names from\nfiles that were completely unobserved during training. Furthermore, we show\nthat our model learns useful method name vectors that capture semantic\nsimilarities, combinations, and analogies. Comparing previous techniques over\nthe same data set, our approach obtains a relative improvement of over 75%,\nbeing the first to successfully predict method names based on a large,\ncross-project, corpus. Our trained model, visualizations and vector\nsimilarities are available as an interactive online demo at\nthis http URL. The code, data, and trained models are available at\nthis https URL.\n",
"title": "code2vec: Learning Distributed Representations of Code"
}
| null | null | null | null | true | null |
17371
| null |
Default
| null | null |
null |
{
"abstract": " Robust data association is necessary for virtually every SLAM system and\nfinding corresponding points is typically a preprocessing step for scan\nalignment algorithms. Traditionally, handcrafted feature descriptors were used\nfor these problems but recently learned descriptors have been shown to perform\nmore robustly. In this work, we propose a local feature descriptor for 3D LiDAR\nscans. The descriptor is learned using a Convolutional Neural Network (CNN).\nOur proposed architecture consists of a Siamese network for learning a feature\ndescriptor and a metric learning network for matching the descriptors. We also\npresent a method for estimating local surface patches and obtaining\nground-truth correspondences. In extensive experiments, we compare our learned\nfeature descriptor with existing 3D local descriptors and report highly\ncompetitive results for multiple experiments in terms of matching accuracy and\ncomputation time. \\end{abstract}\n",
"title": "Learning a Local Feature Descriptor for 3D LiDAR Scans"
}
| null | null | null | null | true | null |
17372
| null |
Default
| null | null |
null |
{
"abstract": " We study the effect of dynamical tides associated with the excitation of\ngravity waves in an interior radiative region of the central star on orbital\nevolution in observed systems containing Hot Jupiters. We consider WASP-43,\nOgle-tr-113, WASP-12, and WASP-18 which contain stars on the main sequence\n(MS). For these systems there are observational estimates regarding the rate of\nchange of the orbital period. We also investigate Kepler-91 which contains an\nevolved giant star. We adopt the formalism of Ivanov et al. for calculating the\norbital evolution.\nFor the MS stars we determine expected rates of orbital evolution under\ndifferent assumptions about the amount of dissipation acting on the tides,\nestimate the effect of stellar rotation for the two most rapidly rotating stars\nand compare results with observations. All cases apart from possibly WASP-43\nare consistent with a regime in which gravity waves are damped during their\npropagation over the star. However, at present this is not definitive as\nobservational errors are large. We find that although it is expected to apply\nto Kepler-91, linear radiative damping cannot explain this dis- sipation regime\napplying to MS stars. Thus, a nonlinear mechanism may be needed.\nKepler-91 is found to be such that the time scale for evolution of the star\nis comparable to that for the orbit. This implies that significant orbital\ncircularisation may have occurred through tides acting on the star.\nQuasi-static tides, stellar winds, hydrodynamic drag and tides acting on the\nplanet have likely played a minor role.\n",
"title": "Dynamical tides in exoplanetary systems containing Hot Jupiters: confronting theory and observations"
}
| null | null | null | null | true | null |
17373
| null |
Default
| null | null |
null |
{
"abstract": " We consider a Bose-Einstein condensate (BEC) with attractive two-body\ninteractions in a cigar-shaped trap, initially prepared in its ground state for\na given negative scattering length, which is quenched to a larger absolute\nvalue of the scattering length. Using the mean-field approximation, we compute\nnumerically, for an experimentally relevant range of aspect ratios and initial\nstrengths of the coupling, two critical values of quench: one corresponds to\nthe weakest attraction strength the quench to which causes the system to\ncollapse before completing even a single return from the narrow configuration\n(\"perihelion\") in its breathing cycle. The other is a similar critical point\nfor the occurrence of collapse before completing two returns. In the latter\ncase, we also compute the limiting value, as we keep increasing the strength of\nthe post-quench attraction towards its critical value, of the time interval\nbetween the first two perihelia. We also use a Gaussian variational model to\nestimate the critical quenched attraction strength below which the system is\nstable against the collapse for long times. These time intervals and critical\nattraction strengths---apart from being fundamental properties of nonlinear\ndynamics of self-attractive BECs---may provide clues to the design of upcoming\nexperiments that are trying to create robust BEC breathers.\n",
"title": "Metastability versus collapse following a quench in attractive Bose-Einstein condensates"
}
| null | null | null | null | true | null |
17374
| null |
Default
| null | null |
null |
{
"abstract": " The execution of sequential programs allows them to be represented using\nmathematical functions formed by the composition of statements following one\nafter the other. Each such statement is in itself a partial function, which\nallows only inputs satisfying a particular Boolean condition to carry forward\nthe execution and hence, the composition of such functions (as a result of\nsequential execution of the statements) strengthens the valid set of input\nstate variables for the program to complete its execution and halt succesfully.\nWith this thought in mind, this paper tries to study a particular class of\npartial functions, which tend to preserve the truth of two given Boolean\nconditions whenever the state variables satisfying one are mapped through such\nfunctions into a domain of state variables satisfying the other. The existence\nof such maps allows us to study isomorphism between different programs, based\nnot only on their structural characteristics (e.g. the kind of programming\nconstructs used and the overall input-output transformation), but also the\nnature of computation performed on seemingly different inputs. Consequently, we\ncan now relate programs which perform a given type of computation, like a loop\ncounting down indefinitely, without caring about the input sets they work on\nindividually or the set of statements each program contains.\n",
"title": "A similarity criterion for sequential programs using truth-preserving partial functions"
}
| null | null | null | null | true | null |
17375
| null |
Default
| null | null |
null |
{
"abstract": " Specify a randomized algorithm that, given a very large graph or network,\nextracts a random subgraph. What can we learn about the input graph from a\nsingle subsample? We derive laws of large numbers for the sampler output, by\nrelating randomized subsampling to distributional invariance: Assuming an\ninvariance holds is tantamount to assuming the sample has been generated by a\nspecific algorithm. That in turn yields a notion of ergodicity. Sampling\nalgorithms induce model classes---graphon models, sparse generalizations of\nexchangeable graphs, and random multigraphs with exchangeable edges can all be\nobtained in this manner, and we specialize our results to a number of examples.\nOne class of sampling algorithms emerges as special: Roughly speaking, those\ndefined as limits of random transformations drawn uniformly from certain\nsequences of groups. Some known pathologies of network models based on graphons\nare explained as a form of selection bias.\n",
"title": "Subsampling large graphs and invariance in networks"
}
| null | null | null | null | true | null |
17376
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we prove modularity results of Taylor coefficients of certain\nnon-holomorphic Jacobi forms. It is well-known that Taylor coefficients of\nholomorphic Jacobi forms are quasimoular forms. However recently there has been\na wide interest for Taylor coefficients of non-holomorphic Jacobi forms for\nexample arising in combinatorics. In this paper, we show that such coefficients\nstill inherit modular properties. We then work out the precise spaces in which\nthese coefficients lie for two examples.\n",
"title": "Taylor coefficients of non-holomorphic Jacobi forms and applications"
}
| null | null | null | null | true | null |
17377
| null |
Default
| null | null |
null |
{
"abstract": " For future networks (i.e., the fifth generation (5G) wireless networks and\nbeyond), millimeter-wave (mmWave) communication with large available unlicensed\nspectrum is a promising technology that enables gigabit multimedia\napplications. Thanks to the short wavelength of mmWave radio, massive antenna\narrays can be packed into the limited dimensions of mmWave transceivers.\nTherefore, with directional beamforming (BF), both mmWave transmitters (MTXs)\nand mmWave receivers (MRXs) are capable of supporting multiple beams in 5G\nnetworks. However, for the transmission between an MTX and an MRX, most works\nhave only considered a single beam, which means that they do not make full\npotential use of mmWave. Furthermore, the connectivity of single beam\ntransmission can easily be blocked. In this context, we propose a single-user\nmulti-beam concurrent transmission scheme for future mmWave networks with\nmultiple reflected paths. Based on spatial spectrum reuse, the scheme can be\ndescribed as a multiple-input multiple-output (MIMO) technique in beamspace\n(i.e., in the beam-number domain). Moreover, this study investigates the\nchallenges and potential solutions for implementing this scheme, including\nmultibeam selection, cooperative beam tracking, multi-beam power allocation and\nsynchronization. The theoretical and numerical results show that the proposed\nbeamspace SU-MIMO can largely improve the achievable rate of the transmission\nbetween an MTX and an MRX and, meanwhile, can maintain the connectivity.\n",
"title": "Beamspace SU-MIMO for Future Millimeter Wave Wireless Communications"
}
| null | null | null | null | true | null |
17378
| null |
Default
| null | null |
null |
{
"abstract": " Many of the existing methods for learning joint embedding of images and text\nuse only supervised information from paired images and its textual attributes.\nTaking advantage of the recent success of unsupervised learning in deep neural\nnetworks, we propose an end-to-end learning framework that is able to extract\nmore robust multi-modal representations across domains. The proposed method\ncombines representation learning models (i.e., auto-encoders) together with\ncross-domain learning criteria (i.e., Maximum Mean Discrepancy loss) to learn\njoint embeddings for semantic and visual features. A novel technique of\nunsupervised-data adaptation inference is introduced to construct more\ncomprehensive embeddings for both labeled and unlabeled data. We evaluate our\nmethod on Animals with Attributes and Caltech-UCSD Birds 200-2011 dataset with\na wide range of applications, including zero and few-shot image recognition and\nretrieval, from inductive to transductive settings. Empirically, we show that\nour framework improves over the current state of the art on many of the\nconsidered tasks.\n",
"title": "Learning Robust Visual-Semantic Embeddings"
}
| null | null | null | null | true | null |
17379
| null |
Default
| null | null |
null |
{
"abstract": " Kepler-452b is currently the best example of an Earth-size planet in the\nhabitable zone of a sun-like star, a type of planet whose number of detections\nis expected to increase in the future. Searching for biosignatures in the\nsupposedly thin atmospheres of these planets is a challenging goal that\nrequires a careful selection of the targets. Under the assumption of a\nrocky-dominated nature for Kepler-452b, we considered it as a test case to\ncalculate a temperature-dependent habitability index, $h_{050}$, designed to\nmaximize the potential presence of biosignature-producing activity (Silva et\nal.\\ 2016). The surface temperature has been computed for a broad range of\nclimate factors using a climate model designed for terrestrial-type exoplanets\n(Vladilo et al.\\ 2015). After fixing the planetary data according to the\nexperimental results (Jenkins et al.\\ 2015), we changed the surface gravity,\nCO$_2$ abundance, surface pressure, orbital eccentricity, rotation period, axis\nobliquity and ocean fraction within the range of validity of our model. For\nmost choices of parameters we find habitable solutions with $h_{050}>0.2$ only\nfor CO$_2$ partial pressure $p_\\mathrm{CO_2} \\lesssim 0.04$\\,bar. At this\nlimiting value of CO$_2$ abundance the planet is still habitable if the total\npressure is $p \\lesssim 2$\\,bar. In all cases the habitability drops for\neccentricity $e \\gtrsim 0.3$. Changes of rotation period and obliquity affect\nthe habitability through their impact on the equator-pole temperature\ndifference rather than on the mean global temperature. We calculated the\nvariation of $h_{050}$ resulting from the luminosity evolution of the host star\nfor a wide range of input parameters. Only a small combination of parameters\nyield habitability-weighted lifetimes $\\gtrsim 2$\\,Gyr, sufficiently long to\ndevelop atmospheric biosignatures still detectable at the present time.\n",
"title": "Quantitative estimates of the surface habitability of Kepler-452b"
}
| null | null | null | null | true | null |
17380
| null |
Default
| null | null |
null |
{
"abstract": " We report the propagation of a square wave signal in a quasi-periodically\ndriven Murali-Lakshmanan-Chua (QPDMLC) circuit system. It is observed that\nsignal propagation is possible only above a certain threshold strength of the\nsquare wave or digital signal and all the values above the threshold amplitude\nare termed as 'region of signal propagation'. Then, we extend this region of\nsignal propagation to perform various logical operations like AND/NAND/OR/NOR\nand hence it is also designated as the 'region of logical operation'. Based on\nthis region, we propose implementing the dynamic logic gates, namely\nAND/NAND/OR/NOR, which can be decided by the asymmetrical input square waves\nwithout altering the system parameters. Further, we show that a single QPDMLC\nsystem will produce simultaneously two outputs which are complementary to each\nother. As a result, a single QPDMLC system yields either AND as well as NAND or\nOR as well as NOR gates simultaneously. Then we combine the corresponding two\nQPDMLC systems in a cross-coupled way and report that its dynamics mimics that\nof fundamental R-S flip-flop circuit. All these phenomena have been explained\nwith analytical solutions of the circuit equations characterizing the system\nand finally the results are compared with the corresponding numerical and\nexperimental analysis.\n",
"title": "Design and implementation of dynamic logic gates and R-S flip-flop using quasiperiodically driven Murali-Lakshmanan-Chua circuit"
}
| null | null | null | null | true | null |
17381
| null |
Default
| null | null |
null |
{
"abstract": " We present Sequential Attend, Infer, Repeat (SQAIR), an interpretable deep\ngenerative model for videos of moving objects. It can reliably discover and\ntrack objects throughout the sequence of frames, and can also generate future\nframes conditioning on the current frame, thereby simulating expected motion of\nobjects. This is achieved by explicitly encoding object presence, locations and\nappearances in the latent variables of the model. SQAIR retains all strengths\nof its predecessor, Attend, Infer, Repeat (AIR, Eslami et. al., 2016),\nincluding learning in an unsupervised manner, and addresses its shortcomings.\nWe use a moving multi-MNIST dataset to show limitations of AIR in detecting\noverlapping or partially occluded objects, and show how SQAIR overcomes them by\nleveraging temporal consistency of objects. Finally, we also apply SQAIR to\nreal-world pedestrian CCTV data, where it learns to reliably detect, track and\ngenerate walking pedestrians with no supervision.\n",
"title": "Sequential Attend, Infer, Repeat: Generative Modelling of Moving Objects"
}
| null | null | null | null | true | null |
17382
| null |
Default
| null | null |
null |
{
"abstract": " We present a new local descriptor for 3D shapes, directly applicable to a\nwide range of shape analysis problems such as point correspondences, semantic\nsegmentation, affordance prediction, and shape-to-scan matching. The descriptor\nis produced by a convolutional network that is trained to embed geometrically\nand semantically similar points close to one another in descriptor space. The\nnetwork processes surface neighborhoods around points on a shape that are\ncaptured at multiple scales by a succession of progressively zoomed out views,\ntaken from carefully selected camera positions. We leverage two extremely large\nsources of data to train our network. First, since our network processes\nrendered views in the form of 2D images, we repurpose architectures pre-trained\non massive image datasets. Second, we automatically generate a synthetic dense\npoint correspondence dataset by non-rigid alignment of corresponding shape\nparts in a large collection of segmented 3D models. As a result of these design\nchoices, our network effectively encodes multi-scale local context and\nfine-grained surface detail. Our network can be trained to produce either\ncategory-specific descriptors or more generic descriptors by learning from\nmultiple shape categories. Once trained, at test time, the network extracts\nlocal descriptors for shapes without requiring any part segmentation as input.\nOur method can produce effective local descriptors even for shapes whose\ncategory is unknown or different from the ones used while training. We\ndemonstrate through several experiments that our learned local descriptors are\nmore discriminative compared to state of the art alternatives, and are\neffective in a variety of shape analysis applications.\n",
"title": "Learning Local Shape Descriptors from Part Correspondences With Multi-view Convolutional Networks"
}
| null | null | null | null | true | null |
17383
| null |
Default
| null | null |
null |
{
"abstract": " We present theoretical guarantees for an alternating minimization algorithm\nfor the dictionary learning/sparse coding problem. The dictionary learning\nproblem is to factorize vector samples $y^{1},y^{2},\\ldots, y^{n}$ into an\nappropriate basis (dictionary) $A^*$ and sparse vectors $x^{1*},\\ldots,x^{n*}$.\nOur algorithm is a simple alternating minimization procedure that switches\nbetween $\\ell_1$ minimization and gradient descent in alternate steps.\nDictionary learning and specifically alternating minimization algorithms for\ndictionary learning are well studied both theoretically and empirically.\nHowever, in contrast to previous theoretical analyses for this problem, we\nreplace the condition on the operator norm (that is, the largest magnitude\nsingular value) of the true underlying dictionary $A^*$ with a condition on the\nmatrix infinity norm (that is, the largest magnitude term). This not only\nallows us to get convergence rates for the error of the estimated dictionary\nmeasured in the matrix infinity norm, but also ensures that a random\ninitialization will provably converge to the global optimum. Our guarantees are\nunder a reasonable generative model that allows for dictionaries with growing\noperator norms, and can handle an arbitrary level of overcompleteness, while\nhaving sparsity that is information theoretically optimal. We also establish\nupper bounds on the sample complexity of our algorithm.\n",
"title": "Alternating minimization for dictionary learning with random initialization"
}
| null | null | null | null | true | null |
17384
| null |
Default
| null | null |
null |
{
"abstract": " In recent years, there have been increasing concerns about how geomagnetic\ndisturbances (GMDs) impact electrical power systems. Geomagnetically-induced\ncurrents (GICs) can saturate transformers, induce hot spot heating and increase\nreactive power losses. These effects can potentially cause catastrophic damage\nto transformers and severely impact the ability of a power system to deliver\npower. To address this problem, we develop a model of GIC impacts to power\nsystems that includes 1) GIC thermal capacity of transformers as a function of\nnormal Alternating Current (AC) and 2) reactive power losses as a function of\nGIC. We use this model to derive an optimization problem that protects power\nsystems from GIC impacts through line switching, generator redispatch, and load\nshedding. We employ state-of-the-art convex relaxations of AC power flow\nequations to lower bound the objective. We demonstrate the approach on a\nmodified RTS96 system and the UIUC 150-bus system and show that line switching\nis an effective means to mitigate GIC impacts. We also provide a sensitivity\nanalysis of optimal switching decisions with respect to GMD direction.\n",
"title": "Optimal Transmission Line Switching under Geomagnetic Disturbances"
}
| null | null | null | null | true | null |
17385
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we propose to utilize Convolutional Neural Networks (CNNs) and\nthe segmentation-based multi-scale analysis to locate tampered areas in digital\nimages. First, to deal with color input sliding windows of different scales, a\nunified CNN architecture is designed. Then, we elaborately design the training\nprocedures of CNNs on sampled training patches. With a set of robust\nmulti-scale tampering detectors based on CNNs, complementary tampering\npossibility maps can be generated. Last but not least, a segmentation-based\nmethod is proposed to fuse the maps and generate the final decision map. By\nexploiting the benefits of both the small-scale and large-scale analyses, the\nsegmentation-based multi-scale analysis can lead to a performance leap in\nforgery localization of CNNs. Numerous experiments are conducted to demonstrate\nthe effectiveness and efficiency of our method.\n",
"title": "Image Forgery Localization Based on Multi-Scale Convolutional Neural Networks"
}
| null | null | null | null | true | null |
17386
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we consider the derivation of the Kadomtsev-Petviashvili (KP)\nequation for cold ion-acoustic wave in the long wavelength limit of the\ntwo-dimensional quantum Euler-Poisson system, under different scalings for\nvarying directions in the Gardner-Morikawa transform. It is shown that the\ntypes of the KP equation depend on the scaled quantum parameter $H>0$. The\nQKP-I is derived for $H>2$, QKP-II for $0<H<2$ and the dispersive-less KP (dKP)\nequation for the critical case $H=2$. The rigorous proof for these limits is\ngiven in the well-prepared initial data case, and the norm that is chosen to\nclose the proof is anisotropic in the two directions, in accordance with the\nanisotropic structure of the KP equation as well as the Gardner-Morikawa\ntransform. The results can be generalized in several directions.\n",
"title": "The QKP limit of the quantum Euler-Poisson equation"
}
| null | null | null | null | true | null |
17387
| null |
Default
| null | null |
null |
{
"abstract": " This paper introduces the variational implicit processes (VIPs), a Bayesian\nnonparametric method based on a class of highly flexible priors over functions.\nSimilar to Gaussian processes (GPs), in implicit processes (IPs), an implicit\nmultivariate prior (data simulators, Bayesian neural networks, etc.) is placed\nover any finite collections of random variables. A novel and efficient\nvariational inference algorithm for IPs is derived using wake-sleep updates,\nwhich gives analytic solutions and allows scalable hyper-parameter learning\nwith stochastic optimization. Experiments on real-world regression datasets\ndemonstrate that VIPs return better uncertainty estimates and superior\nperformance over existing inference methods for GPs and Bayesian neural\nnetworks. With a Bayesian LSTM as the implicit prior, the proposed approach\nachieves state-of-the-art results on predicting power conversion efficiency of\nmolecules based on raw chemical formulas.\n",
"title": "Variational Implicit Processes"
}
| null | null | null | null | true | null |
17388
| null |
Default
| null | null |
null |
{
"abstract": " Transformation optics methods and gradient index electromagnetic structures\nrely upon spatially varied arbitrary permittivity. This, along with recent\ninterest in millimeter-wave lens-based antennas demands high spatial resolution\ndielectric variation. Perforated media have been used to fabricate gradient\nindex structures from microwaves to THz but are often limited in contrast. We\nshow that by employing regular polygon unit-cells (hexagon, square, and\ntriangle) on matched lattices we can realize very high contrast permittivity\nranging from 0.1-1.0 of the background permittivity. Silicon micromachining\n(Bosch process) is performed on high resistivity Silicon wafers to achieve a\nminimum permittivity of 1.25 (10% of Silicon) in the WR28 waveguide band,\nspecifically targeting the proposed 39 GHz 5G communications band. The method\nis valid into the THz band.\n",
"title": "Silicon Micromachined High-contrast Artificial Dielectrics for Millimeter-wave Transformation Optics Antennas"
}
| null | null | null | null | true | null |
17389
| null |
Default
| null | null |
null |
{
"abstract": " Lanthanum family of high-temperature cuprate superconductors is known to\nexhibit both spin and charge electronic modulations around doping level 1/8. We\nassume that these modulations have the character of two-dimensional spin-vortex\ncheckerboard and investigate whether this assumption is consistent with the\nFermi surface and the pseudogap measured by angle-resolved photo-emission\nspectroscopy. We also explore the possibility of observing quantum oscillations\nof transport coefficients in such a background. These investigations are based\non a model of non-interacting spin-1/2 fermions hopping on a square lattice and\ncoupled through spins to a magnetic field imitating spin-vortex checkerboard.\nThe main results of this article include (i) calculation of Fermi surface\ncontaining Fermi arcs at the positions in the Brillouin zone largely consistent\nwith experiments; (ii) identification of factors complicating the observations\nof quantum oscillations in the presence of spin modulations; and (iii)\ninvestigation of the symmetries of the resulting electronic energy bands,\nwhich, in particular, indicates that each band is double-degenerate and, in\naddition, has at least one conical point, where it touches another\ndouble-degenerate band. We discuss possible implications these cones may have\nfor the transport properties and the pseudogap.\n",
"title": "Pseudogap and Fermi surface in the presence of spin-vortex checkerboard for 1/8-doped lanthanum cuprates"
}
| null | null | null | null | true | null |
17390
| null |
Default
| null | null |
null |
{
"abstract": " Capable of reaching similar magnitudes to large megathrust earthquakes\n($M_w>7$), slow slip events play a major role in accommodating tectonic motion\non plate boundaries. These slip transients are the slow release of built-up\ntectonic stress that are geodetically imaged as a predominantly aseismic\nrupture, which is smooth in both time and space. We demonstrate here that large\nslow slip events are in fact a cluster of short-duration slow transients. Using\na dense catalog of low-frequency earthquakes as a guide, we investigate the\n$M_w7.5$ slow slip event that occurred in 2006 along the subduction interface\n40~km beneath Guerrero, Mexico. We show that while the long-period surface\ndisplacement as recorded by GPS suggests a six month duration, motion in the\ndirection of tectonic release only sporadically occurs over 55 days and its\nsurface signature is attenuated by rapid relocking of the plate interface.These\nresults demonstrate that our current conceptual model of slow and continuous\nrupture is an artifact of low-resolution geodetic observations of a\nsuperposition of small, clustered slip events. Our proposed description of slow\nslip as a cluster of slow transients implies that we systematically\noverestimate the duration $T$ and underestimate the moment magnitude $M$ of\nlarge slow slip events.\n",
"title": "Revealing the cluster of slow transients behind a large slow slip event"
}
| null | null |
[
"Physics"
] | null | true | null |
17391
| null |
Validated
| null | null |
null |
{
"abstract": " General $N$-solitons in three recently-proposed nonlocal nonlinear\nSchrödinger equations are presented. These nonlocal equations include the\nreverse-space, reverse-time, and reverse-space-time nonlinear Schrödinger\nequations, which are nonlocal reductions of the Ablowitz-Kaup-Newell-Segur\n(AKNS) hierarchy. It is shown that general $N$-solitons in these different\nequations can be derived from the same Riemann-Hilbert solutions of the AKNS\nhierarchy, except that symmetry relations on the scattering data are different\nfor these equations. This Riemann-Hilbert framework allows us to identify new\ntypes of solitons with novel eigenvalue configurations in the spectral plane.\nDynamics of $N$-solitons in these equations is also explored. In all the three\nnonlocal equations, a generic feature of their solutions is repeated\ncollapsing. In addition, multi-solitons can behave very differently from\nfundamental solitons and may not correspond to a nonlinear superposition of\nfundamental solitons.\n",
"title": "General $N$-solitons and their dynamics in several nonlocal nonlinear Schrödinger equations"
}
| null | null | null | null | true | null |
17392
| null |
Default
| null | null |
null |
{
"abstract": " We revisit the mathematical models for wireless network jamming introduced by\nCommander et al.: we first point out the strong connections with classical\nwireless network design and then we propose a new model based on the explicit\nuse of signal-to-interference quantities. Moreover, to address the intrinsic\nuncertain nature of the jamming problem and tackle the peculiar right-hand-side\n(RHS) uncertainty of the problem, we propose an original robust cutting-plane\nalgorithm drawing inspiration from Multiband Robust Optimization. Finally, we\nassess the performance of the proposed cutting plane algorithm by experiments\non realistic network instances.\n",
"title": "Revisiting wireless network jamming by SIR-based considerations and Multiband Robust Optimization"
}
| null | null | null | null | true | null |
17393
| null |
Default
| null | null |
null |
{
"abstract": " Symbolic data analysis (SDA) is an emerging area of statistics based on\naggregating individual level data into group-based distributional summaries\n(symbols), and then developing statistical methods to analyse them. It is ideal\nfor analysing large and complex datasets, and has immense potential to become a\nstandard inferential technique in the near future. However, existing SDA\ntechniques are either non-inferential, do not easily permit meaningful\nstatistical models, are unable to distinguish between competing models, and are\nbased on simplifying assumptions that are known to be false. Further, the\nprocedure for constructing symbols from the underlying data is erroneously not\nconsidered relevant to the resulting statistical analysis. In this paper we\nintroduce a new general method for constructing likelihood functions for\nsymbolic data based on a desired probability model for the underlying classical\ndata, while only observing the distributional summaries. This approach resolves\nmany of the conceptual and practical issues with current SDA methods, opens the\ndoor for new classes of symbol design and construction, in addition to\ndeveloping SDA as a viable tool to enable and improve upon classical data\nanalyses, particularly for very large and complex datasets. This work creates a\nnew direction for SDA research, which we illustrate through several real and\nsimulated data analyses.\n",
"title": "New models for symbolic data analysis"
}
| null | null | null | null | true | null |
17394
| null |
Default
| null | null |
null |
{
"abstract": " Many real-world data mining applications need varying cost for different\ntypes of classification errors and thus call for cost-sensitive classification\nalgorithms. Existing algorithms for cost-sensitive classification are\nsuccessful in terms of minimizing the cost, but can result in a high error rate\nas the trade-off. The high error rate holds back the practical use of those\nalgorithms. In this paper, we propose a novel cost-sensitive classification\nmethodology that takes both the cost and the error rate into account. The\nmethodology, called soft cost-sensitive classification, is established from a\nmulticriteria optimization problem of the cost and the error rate, and can be\nviewed as regularizing cost-sensitive classification with the error rate. The\nsimple methodology allows immediate improvements of existing cost-sensitive\nclassification algorithms. Experiments on the benchmark and the real-world data\nsets show that our proposed methodology indeed achieves lower test error rates\nand similar (sometimes lower) test costs than existing cost-sensitive\nclassification algorithms. We also demonstrate that the methodology can be\nextended for considering the weighted error rate instead of the original error\nrate. This extension is useful for tackling unbalanced classification problems.\n",
"title": "Soft Methodology for Cost-and-error Sensitive Classification"
}
| null | null |
[
"Computer Science"
] | null | true | null |
17395
| null |
Validated
| null | null |
null |
{
"abstract": " The Cherenkov Telescope Array (CTA) is the next generation of Imaging\nAtmospheric Cherenkov Telescopes. It will reach a sensitivity and energy\nresolution never obtained until now by any other high energy gamma-ray\nexperiment. Understanding the systematic uncertainties in general will be a\ncrucial issue for the performance of CTA. It is well known that atmospheric\nconditions contribute particularly in this aspect.Within the CTA consortium\nseveral groups are currently building Raman LIDARs to be installed on the two\nsites. Raman LIDARs are devices composed of a powerful laser that shoots into\nthe atmosphere, a collector that gathers the backscattered light from molecules\nand aerosols, a photo-sensor, an optical module that spectrally selects\nwavelengths of interest, and a read--out system.Unlike currently used elastic\nLIDARs, they can help reduce the systematic uncertainties of the molecular and\naerosol components of the atmosphere to <5% so that CTA can achieve its energy\nresolution requirements of<10% uncertainty at 1 TeV.All the Raman LIDARs in\nthis work have design features that make them different than typical Raman\nLIDARs used in atmospheric science and are characterized by large collecting\nmirrors (2.5m2) and reduced acquisition time.They provide both multiple elastic\nand Raman read-out channels and custom made optics design.In this paper, the\nmotivation for Raman LIDARs, the design and the status of advance of these\ntechnologies are described.\n",
"title": "Raman LIDARs and atmospheric calibration for the Cherenkov Telescope Array"
}
| null | null | null | null | true | null |
17396
| null |
Default
| null | null |
null |
{
"abstract": " The restricted isometry property (RIP) is a universal tool for data recovery.\nWe explore the implication of the RIP in the framework of generalized sparsity\nand group measurements introduced in the Part I paper. It turns out that for a\ngiven measurement instrument the number of measurements for RIP can be improved\nby optimizing over families of Banach spaces. Second, we investigate the\npreservation of difference of two sparse vectors, which is not trivial in\ngeneralized models. Third, we extend the RIP of partial Fourier measurements at\noptimal scaling of number of measurements with random sign to far more general\ngroup structured measurements. Lastly, we also obtain RIP in infinite dimension\nin the context of Fourier measurement concepts with sparsity naturally replaced\nby smoothness assumptions.\n",
"title": "Generalized notions of sparsity and restricted isometry property. Part II: Applications"
}
| null | null | null | null | true | null |
17397
| null |
Default
| null | null |
null |
{
"abstract": " Nonparametric kernel density estimation is a very natural procedure which\nsimply makes use of the smoothing power of the convolution operation. Yet, it\nperforms poorly when the density of a positive variable is to be estimated\n(boundary issues, spurious bumps in the tail). So various extensions of the\nbasic kernel estimator allegedly suitable for $\\mathbb{R}^+$-supported\ndensities, such as those using Gamma or other asymmetric kernels, abound in the\nliterature. Those, however, are not based on any valid smoothing operation\nanalogous to the convolution, which typically leads to inconsistencies. By\ncontrast, in this paper a kernel estimator for $\\mathbb{R}^+$-supported\ndensities is defined by making use of the Mellin convolution, the natural\nanalogue of the usual convolution on $\\mathbb{R}^+$. From there, a very\ntransparent theory flows and leads to new type of asymmetric kernels strongly\nrelated to Meijer's $G$-functions. The numerous pleasant properties of this\n`Mellin-Meijer-kernel density estimator' are demonstrated in the paper. Its\npointwise and $L_2$-consistency (with optimal rate of convergence) is\nestablished for a large class of densities, including densities unbounded at 0\nand showing power-law decay in their right tail. Its practical behaviour is\ninvestigated further through simulations and some real data analyses.\n",
"title": "Mellin-Meijer-kernel density estimation on $\\mathbb{R}^+$"
}
| null | null | null | null | true | null |
17398
| null |
Default
| null | null |
null |
{
"abstract": " We applied machine learning to predict whether a gene is involved in axon\nregeneration. We extracted 31 features from different databases and trained\nfive machine learning models. Our optimal model, a Random Forest Classifier\nwith 50 submodels, yielded a test score of 85.71%, which is 4.1% higher than\nthe baseline score. We concluded that our models have some predictive\ncapability. Similar methodology and features could be applied to predict other\nGene Ontology (GO) terms.\n",
"title": "Gene Ontology (GO) Prediction using Machine Learning Methods"
}
| null | null | null | null | true | null |
17399
| null |
Default
| null | null |
null |
{
"abstract": " This paper investigates the algorithmic dimension spectra of lines in the\nEuclidean plane. Given any line L with slope a and vertical intercept b, the\ndimension spectrum sp(L) is the set of all effective Hausdorff dimensions of\nindividual points on L. We draw on Kolmogorov complexity and geometrical\narguments to show that if the effective Hausdorff dimension dim(a, b) is equal\nto the effective packing dimension Dim(a, b), then sp(L) contains a unit\ninterval. We also show that, if the dimension dim(a, b) is at least one, then\nsp(L) is infinite. Together with previous work, this implies that the dimension\nspectrum of any line is infinite.\n",
"title": "Dimension Spectra of Lines"
}
| null | null |
[
"Computer Science"
] | null | true | null |
17400
| null |
Validated
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.