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 describe the road which led to the construction and exploitation of\nelectron positron colliders, hightlighting how the young physics student Bruno\nTouschek met the Norwegian engineer Rolf Wideroe in Germany, during WWII, and\ncollaborated in building the 15 MeV betatron, a secret project directed by\nWideroe and financed by the Ministry of Aviation of the Reich. This is how\nBruno Touschek learnt the science of making particle accelerators and was\nready, many years later, to propose and build AdA, the first electron positron\ncollider, in Frascati, Italy, in 1960. We shall then see how AdA was brought\nfrom Frascati to Orsay, in France. Taking advantage of the Orsay Linear\nAccelerator as injector, the Franco-Italian team was able to prove that\ncollisions had taken place, opening the way to the use of particle colliders as\na mean to explore high energy physics.\n",
"title": "The path to high-energy electron-positron colliders: from Wideroe's betatron to Touschek's AdA and to LEP"
}
| null | null | null | null | true | null |
3801
| null |
Default
| null | null |
null |
{
"abstract": " Consider an ample and globally generated line bundle $L$ on a smooth\nprojective variety $X$ of dimension $N\\geq 2$ over $\\mathbb{C}$. Let $D$ be a\nsmooth divisor in the complete linear system of $L$. We construct reflexive\nsheaves on $X$ by an elementary transformation of a trivial bundle on $X$ along\ncertain globally generated torsion-free sheaves on $D$. The dual reflexive\nsheaves are called the Lazarsfeld-Mukai reflexive sheaves. We prove the\n$\\mu_L$-(semi)stability of such reflexive sheaves under certain conditions.\n",
"title": "Lazarsfeld-Mukai Reflexive Sheaves and their Stability"
}
| null | null | null | null | true | null |
3802
| null |
Default
| null | null |
null |
{
"abstract": " We propose a method for recognizing moving vehicles, using data from roadside\naudio sensors. This problem has applications ranging widely, from traffic\nanalysis to surveillance. We extract a frequency signature from the audio\nsignal using a short-time Fourier transform, and treat each time window as an\nindividual data point to be classified. By applying a spectral embedding, we\ndecrease the dimensionality of the data sufficiently for K-nearest neighbors to\nprovide accurate vehicle identification.\n",
"title": "Dimensionality reduction for acoustic vehicle classification with spectral embedding"
}
| null | null | null | null | true | null |
3803
| null |
Default
| null | null |
null |
{
"abstract": " We study principal component analysis (PCA) for mean zero i.i.d. Gaussian\nobservations $X_1,\\dots, X_n$ in a separable Hilbert space $\\mathbb{H}$ with\nunknown covariance operator $\\Sigma.$ The complexity of the problem is\ncharacterized by its effective rank ${\\bf r}(\\Sigma):= \\frac{{\\rm\ntr}(\\Sigma)}{\\|\\Sigma\\|},$ where ${\\rm tr}(\\Sigma)$ denotes the trace of\n$\\Sigma$ and $\\|\\Sigma\\|$ denotes its operator norm. We develop a method of\nbias reduction in the problem of estimation of linear functionals of\neigenvectors of $\\Sigma.$ Under the assumption that ${\\bf r}(\\Sigma)=o(n),$ we\nestablish the asymptotic normality and asymptotic properties of the risk of the\nresulting estimators and prove matching minimax lower bounds, showing their\nsemi-parametric optimality.\n",
"title": "Efficient Estimation of Linear Functionals of Principal Components"
}
| null | null | null | null | true | null |
3804
| null |
Default
| null | null |
null |
{
"abstract": " The aim of this paper is to give a short overview on error bounds and to\nprovide the first bricks of a unified theory. Inspired by the works of [8, 15,\n13, 16, 10], we show indeed the centrality of the Lojasiewicz gradient\ninequality. For this, we review some necessary and sufficient conditions for\nglobal/local error bounds, both in the convex and nonconvex case. We also\nrecall some results on quantitative error bounds which play a major role in\nconvergence rate analysis and complexity theory of many optimization methods.\n",
"title": "A stroll in the jungle of error bounds"
}
| null | null | null | null | true | null |
3805
| null |
Default
| null | null |
null |
{
"abstract": " We introduce variational obstacle avoidance problems on Riemannian manifolds\nand derive necessary conditions for the existence of their normal extremals.\nThe problem consists of minimizing an energy functional depending on the\nvelocity and covariant acceleration, among a set of admissible curves, and also\ndepending on a navigation function used to avoid an obstacle on the workspace,\na Riemannian manifold.\nWe study two different scenarios, a general one on a Riemannian manifold and,\na sub-Riemannian problem. By introducing a left-invariant metric on a Lie\ngroup, we also study the variational obstacle avoidance problem on a Lie group.\nWe apply the results to the obstacle avoidance problem of a planar rigid body\nand an unicycle.\n",
"title": "Variational obstacle avoidance problem on Riemannian manifolds"
}
| null | null |
[
"Computer Science",
"Mathematics"
] | null | true | null |
3806
| null |
Validated
| null | null |
null |
{
"abstract": " This paper introduces an extension of Heron's formula to approximate area of\ncyclic n-gons where the error never exceeds $\\frac{\\pi}{e}-1$\n",
"title": "An Extension of Heron's Formula"
}
| null | null | null | null | true | null |
3807
| null |
Default
| null | null |
null |
{
"abstract": " In representation learning (RL), how to make the learned representations easy\nto interpret and less overfitted to training data are two important but\nchallenging issues. To address these problems, we study a new type of\nregulariza- tion approach that encourages the supports of weight vectors in RL\nmodels to have small overlap, by simultaneously promoting near-orthogonality\namong vectors and sparsity of each vector. We apply the proposed regularizer to\ntwo models: neural networks (NNs) and sparse coding (SC), and develop an\nefficient ADMM-based algorithm for regu- larized SC. Experiments on various\ndatasets demonstrate that weight vectors learned under our regularizer are more\ninterpretable and have better generalization performance.\n",
"title": "Learning Less-Overlapping Representations"
}
| null | null | null | null | true | null |
3808
| null |
Default
| null | null |
null |
{
"abstract": " Distributed storage systems suffer from significant repair traffic generated\ndue to frequent storage node failures. This paper shows that properly designed\nlow-density parity-check (LDPC) codes can substantially reduce the amount of\nrequired block downloads for repair thanks to the sparse nature of their factor\ngraph representation. In particular, with a careful construction of the factor\ngraph, both low repair-bandwidth and high reliability can be achieved for a\ngiven code rate. First, a formula for the average repair bandwidth of LDPC\ncodes is developed. This formula is then used to establish that the minimum\nrepair bandwidth can be achieved by forcing a regular check node degree in the\nfactor graph. Moreover, it is shown that given a fixed code rate, the variable\nnode degree should also be regular to yield minimum repair bandwidth, under\nsome reasonable minimum variable node degree constraint. It is also shown that\nfor a given repair-bandwidth requirement, LDPC codes can yield substantially\nhigher reliability than currently utilized Reed-Solomon (RS) codes. Our\nreliability analysis is based on a formulation of the general equation for the\nmean-time-to-data-loss (MTTDL) associated with LDPC codes. The formulation\nreveals that the stopping number is closely related to the MTTDL. It is further\nshown that LDPC codes can be designed such that a small loss of\nrepair-bandwidth optimality may be traded for a large improvement in\nerasure-correction capability and thus the MTTDL.\n",
"title": "LDPC Code Design for Distributed Storage: Balancing Repair Bandwidth, Reliability and Storage Overhead"
}
| null | null | null | null | true | null |
3809
| null |
Default
| null | null |
null |
{
"abstract": " We report on the optimization process to synthesize epitaxial thin films of\nGdTiO3 on SrLaGaO4 substrates by pulsed laser deposition. Optimized films are\nfree of impurity phases and are fully strained. They possess a magnetic Curie\ntemperature TC = 31.8 K with a saturation magnetization of 4.2 muB per formula\nunit at 10 K. Transport measurements reveal an insulating response, as\nexpected. Optical spectroscopy indicates a band gap of 0.7 eV, comparable to\nthe bulk value. Our work adds ferrimagnetic orthotitanates to the palette of\nperovskite materials for the design of emergent strongly correlated states at\noxide interfaces using a versatile growth technique such as pulsed laser\ndeposition.\n",
"title": "Structural, magnetic, and electronic properties of GdTiO3 Mott insulator thin films grown by pulsed laser deposition"
}
| null | null | null | null | true | null |
3810
| null |
Default
| null | null |
null |
{
"abstract": " As a dedicated solar radio interferometer, the MingantU SpEctral\nRadioHeliograph (MUSER) generates massive observational data in the frequency\nrange of 400 MHz -- 15 GHz. High-performance imaging forms a significantly\nimportant aspect of MUSER's massive data processing requirements. In this\nstudy, we implement a practical high-performance imaging pipeline for MUSER\ndata processing. At first, the specifications of the MUSER are introduced and\nits imaging requirements are analyzed. Referring to the most commonly used\nradio astronomy software such as CASA and MIRIAD, we then implement a\nhigh-performance imaging pipeline based on the Graphics Processing Unit (GPU)\ntechnology with respect to the current operational status of the MUSER. A\nseries of critical algorithms and their pseudo codes, i.e., detection of the\nsolar disk and sky brightness, automatic centering of the solar disk and\nestimation of the number of iterations for clean algorithms, are proposed in\ndetail. The preliminary experimental results indicate that the proposed imaging\napproach significantly increases the processing performance of MUSER and\ngenerates images with high-quality, which can meet the requirements of the\nMUSER data processing.\n",
"title": "GPU-Based High-Performance Imaging for Mingantu Spectral RadioHeliograph"
}
| null | null | null | null | true | null |
3811
| null |
Default
| null | null |
null |
{
"abstract": " We present Deep Voice 3, a fully-convolutional attention-based neural\ntext-to-speech (TTS) system. Deep Voice 3 matches state-of-the-art neural\nspeech synthesis systems in naturalness while training ten times faster. We\nscale Deep Voice 3 to data set sizes unprecedented for TTS, training on more\nthan eight hundred hours of audio from over two thousand speakers. In addition,\nwe identify common error modes of attention-based speech synthesis networks,\ndemonstrate how to mitigate them, and compare several different waveform\nsynthesis methods. We also describe how to scale inference to ten million\nqueries per day on one single-GPU server.\n",
"title": "Deep Voice 3: Scaling Text-to-Speech with Convolutional Sequence Learning"
}
| null | null | null | null | true | null |
3812
| null |
Default
| null | null |
null |
{
"abstract": " Every year, 3 million newborns die within the first month of life. Birth\nasphyxia and other breathing-related conditions are a leading cause of\nmortality during the neonatal phase. Current diagnostic methods are too\nsophisticated in terms of equipment, required expertise, and general logistics.\nConsequently, early detection of asphyxia in newborns is very difficult in many\nparts of the world, especially in resource-poor settings. We are developing a\nmachine learning system, dubbed Ubenwa, which enables diagnosis of asphyxia\nthrough automated analysis of the infant cry. Deployed via smartphone and\nwearable technology, Ubenwa will drastically reduce the time, cost and skill\nrequired to make accurate and potentially life-saving diagnoses.\n",
"title": "Ubenwa: Cry-based Diagnosis of Birth Asphyxia"
}
| null | null | null | null | true | null |
3813
| null |
Default
| null | null |
null |
{
"abstract": " We propose a novel fluid-structure interaction (FSI) scheme using the\nentropic multi-relaxation time lattice Boltzmann (KBC) model for the fluid\ndomain in combination with a nonlinear finite element solver for the structural\npart. We show validity of the proposed scheme for various challenging set-ups\nby comparison to literature data. Beyond validation, we extend the KBC model to\nmultiphase flows and couple it with FEM solver. Robustness and viability of the\nentropic multi-relaxation time model for complex FSI applications is shown by\nsimulations of droplet impact on elastic superhydrophobic surfaces.\n",
"title": "Fluid-Structure Interaction with the Entropic Lattice Boltzmann Method"
}
| null | null | null | null | true | null |
3814
| null |
Default
| null | null |
null |
{
"abstract": " In this article, we advance divide-and-conquer strategies for solving the\ncommunity detection problem in networks. We propose two algorithms which\nperform clustering on a number of small subgraphs and finally patches the\nresults into a single clustering. The main advantage of these algorithms is\nthat they bring down significantly the computational cost of traditional\nalgorithms, including spectral clustering, semi-definite programs, modularity\nbased methods, likelihood based methods etc., without losing on accuracy and\neven improving accuracy at times. These algorithms are also, by nature,\nparallelizable. Thus, exploiting the facts that most traditional algorithms are\naccurate and the corresponding optimization problems are much simpler in small\nproblems, our divide-and-conquer methods provide an omnibus recipe for scaling\ntraditional algorithms up to large networks. We prove consistency of these\nalgorithms under various subgraph selection procedures and perform extensive\nsimulations and real-data analysis to understand the advantages of the\ndivide-and-conquer approach in various settings.\n",
"title": "Two provably consistent divide and conquer clustering algorithms for large networks"
}
| null | null | null | null | true | null |
3815
| null |
Default
| null | null |
null |
{
"abstract": " Robust and fast motion estimation and mapping is a key prerequisite for\nautonomous operation of mobile robots. The goal of performing this task solely\non a stereo pair of video cameras is highly demanding and bears conflicting\nobjectives: on one hand, the motion has to be tracked fast and reliably, on the\nother hand, high-level functions like navigation and obstacle avoidance depend\ncrucially on a complete and accurate environment representation. In this work,\nwe propose a two-layer approach for visual odometry and SLAM with stereo\ncameras that runs in real-time and combines feature-based matching with\nsemi-dense direct image alignment. Our method initializes semi-dense depth\nestimation, which is computationally expensive, from motion that is tracked by\na fast but robust keypoint-based method. Experiments on public benchmark and\nproprietary datasets show that our approach is faster than state-of-the-art\nmethods without losing accuracy and yields comparable map building\ncapabilities. Moreover, our approach is shown to handle large inter-frame\nmotion and illumination changes much more robustly than its direct\ncounterparts.\n",
"title": "Feature-based visual odometry prior for real-time semi-dense stereo SLAM"
}
| null | null | null | null | true | null |
3816
| null |
Default
| null | null |
null |
{
"abstract": " Most existing theories of dark energy and/or modified gravity, involving a\nscalar degree of freedom, can be conveniently described within the framework of\nthe Effective Theory of Dark Energy, based on the unitary gauge where the\nscalar field is uniform. We extend this effective approach by allowing the\nLagrangian in unitary gauge to depend on the time derivative of the lapse\nfunction. Although this dependence generically signals the presence of an extra\nscalar degree of freedom, theories that contain only one propagating scalar\ndegree of freedom, in addition to the usual tensor modes, can be constructed by\nrequiring the initial Lagrangian to be degenerate. Starting from a general\nquadratic action, we derive the dispersion relations for the linear\nperturbations around Minkowski and a cosmological background. Our analysis\ndirectly applies to the recently introduced Degenerate Higher-Order\nScalar-Tensor (DHOST) theories. For these theories, we find that one cannot\nrecover a Poisson-like equation in the static linear regime except for the\nsubclass that includes the Horndeski and so-called \"beyond Horndeski\" theories.\nWe also discuss Lorentz-breaking models inspired by Horava gravity.\n",
"title": "Effective Description of Higher-Order Scalar-Tensor Theories"
}
| null | null |
[
"Physics"
] | null | true | null |
3817
| null |
Validated
| null | null |
null |
{
"abstract": " Video analytics requires operating with large amounts of data. Compressive\nsensing allows to reduce the number of measurements required to represent the\nvideo using the prior knowledge of sparsity of the original signal, but it\nimposes certain conditions on the design matrix. The Bayesian compressive\nsensing approach relaxes the limitations of the conventional approach using the\nprobabilistic reasoning and allows to include different prior knowledge about\nthe signal structure. This paper presents two Bayesian compressive sensing\nmethods for autonomous object detection in a video sequence from a static\ncamera. Their performance is compared on the real datasets with the\nnon-Bayesian greedy algorithm. It is shown that the Bayesian methods can\nprovide the same accuracy as the greedy algorithm but much faster; or if the\ncomputational time is not critical they can provide more accurate results.\n",
"title": "Compressive Sensing Approaches for Autonomous Object Detection in Video Sequences"
}
| null | null | null | null | true | null |
3818
| null |
Default
| null | null |
null |
{
"abstract": " Recent development of large-scale question answering (QA) datasets triggered\na substantial amount of research into end-to-end neural architectures for QA.\nIncreasingly complex systems have been conceived without comparison to simpler\nneural baseline systems that would justify their complexity. In this work, we\npropose a simple heuristic that guides the development of neural baseline\nsystems for the extractive QA task. We find that there are two ingredients\nnecessary for building a high-performing neural QA system: first, the awareness\nof question words while processing the context and second, a composition\nfunction that goes beyond simple bag-of-words modeling, such as recurrent\nneural networks. Our results show that FastQA, a system that meets these two\nrequirements, can achieve very competitive performance compared with existing\nmodels. We argue that this surprising finding puts results of previous systems\nand the complexity of recent QA datasets into perspective.\n",
"title": "Making Neural QA as Simple as Possible but not Simpler"
}
| null | null | null | null | true | null |
3819
| null |
Default
| null | null |
null |
{
"abstract": " Introducing inequality constraints in Gaussian process (GP) models can lead\nto more realistic uncertainties in learning a great variety of real-world\nproblems. We consider the finite-dimensional Gaussian approach from Maatouk and\nBay (2017) which can satisfy inequality conditions everywhere (either\nboundedness, monotonicity or convexity). Our contributions are threefold.\nFirst, we extend their approach in order to deal with general sets of linear\ninequalities. Second, we explore several Markov Chain Monte Carlo (MCMC)\ntechniques to approximate the posterior distribution. Third, we investigate\ntheoretical and numerical properties of the constrained likelihood for\ncovariance parameter estimation. According to experiments on both artificial\nand real data, our full framework together with a Hamiltonian Monte Carlo-based\nsampler provides efficient results on both data fitting and uncertainty\nquantification.\n",
"title": "Finite-dimensional Gaussian approximation with linear inequality constraints"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
3820
| null |
Validated
| null | null |
null |
{
"abstract": " Cyclic codes and their various generalizations, such as quasi-twisted (QT)\ncodes, have a special place in algebraic coding theory. Among other things,\nmany of the best-known or optimal codes have been obtained from these classes.\nIn this work we introduce a new generalization of QT codes that we call\nmulti-twisted (MT) codes and study some of their basic properties. Presenting\nseveral methods of constructing codes in this class and obtaining bounds on the\nminimum distances, we show that there exist codes with good parameters in this\nclass that cannot be obtained as QT or constacyclic codes. This suggests that\nconsidering this larger class in computer searches is promising for\nconstructing codes with better parameters than currently best-known linear\ncodes. Working with this new class of codes motivated us to consider a problem\nabout binomials over finite fields and to discover a result that is interesting\nin its own right.\n",
"title": "A Generalization of Quasi-twisted Codes: Multi-twisted codes"
}
| null | null |
[
"Computer Science",
"Mathematics"
] | null | true | null |
3821
| null |
Validated
| null | null |
null |
{
"abstract": " A method for efficiently successive cancellation (SC) decoding of polar codes\nwith high-dimensional linear binary kernels (HDLBK) is presented and analyzed.\nWe devise a $l$-expressions method which can obtain simplified recursive\nformulas of SC decoder in likelihood ratio form for arbitrary linear binary\nkernels to reduce the complexity of corresponding SC decoder. By considering\nthe bit-channel transition probabilities $W_{G}^{(\\cdot)}(\\cdot|0)$ and\n$W_{G}^{(\\cdot)}(\\cdot|1)$ separately, a $W$-expressions method is proposed to\nfurther reduce the complexity of HDLBK based SC decoder. For a $m\\times m$\nbinary kernel, the complexity of straightforward SC decoder is $O(2^{m}N\\log\nN)$. With $W$-expressions, we reduce the complexity of straightforward SC\ndecoder to $O(m^{2}N\\log N)$ when $m\\leq 16$. Simulation results show that\n$16\\times16$ kernel polar codes offer significant advantages in terms of error\nperformances compared with $2\\times2$ kernel polar codes under SC and list SC\ndecoders.\n",
"title": "On the Successive Cancellation Decoding of Polar Codes with Arbitrary Linear Binary Kernels"
}
| null | null | null | null | true | null |
3822
| null |
Default
| null | null |
null |
{
"abstract": " OpenML is an online machine learning platform where researchers can easily\nshare data, machine learning tasks and experiments as well as organize them\nonline to work and collaborate more efficiently. In this paper, we present an R\npackage to interface with the OpenML platform and illustrate its usage in\ncombination with the machine learning R package mlr. We show how the OpenML\npackage allows R users to easily search, download and upload data sets and\nmachine learning tasks. Furthermore, we also show how to upload results of\nexperiments, share them with others and download results from other users.\nBeyond ensuring reproducibility of results, the OpenML platform automates much\nof the drudge work, speeds up research, facilitates collaboration and increases\nthe users' visibility online.\n",
"title": "OpenML: An R Package to Connect to the Machine Learning Platform OpenML"
}
| null | null | null | null | true | null |
3823
| null |
Default
| null | null |
null |
{
"abstract": " Cultural activity is an inherent aspect of urban life and the success of a\nmodern city is largely determined by its capacity to offer generous cultural\nentertainment to its citizens. To this end, the optimal allocation of cultural\nestablishments and related resources across urban regions becomes of vital\nimportance, as it can reduce financial costs in terms of planning and improve\nquality of life in the city, more generally. In this paper, we make use of a\nlarge longitudinal dataset of user location check-ins from the online social\nnetwork WeChat to develop a data-driven framework for cultural planning in the\ncity of Beijing. We exploit rich spatio-temporal representations on user\nactivity at cultural venues and use a novel extended version of the traditional\nlatent Dirichlet allocation model that incorporates temporal information to\nidentify latent patterns of urban cultural interactions. Using the\ncharacteristic typologies of mobile user cultural activities emitted by the\nmodel, we determine the levels of demand for different types of cultural\nresources across urban areas. We then compare those with the corresponding\nlevels of supply as driven by the presence and spatial reach of cultural venues\nin local areas to obtain high resolution maps that indicate urban regions with\nlack of cultural resources, and thus give suggestions for further urban\ncultural planning and investment optimisation.\n",
"title": "Discovering Latent Patterns of Urban Cultural Interactions in WeChat for Modern City Planning"
}
| null | null | null | null | true | null |
3824
| null |
Default
| null | null |
null |
{
"abstract": " Accurate measurements of the physical structure of protoplanetary discs are\ncritical inputs for planet formation models. These constraints are\ntraditionally established via complex modelling of continuum and line\nobservations. Instead, we present an empirical framework to locate the CO\nisotopologue emitting surfaces from high spectral and spatial resolution ALMA\nobservations. We apply this framework to the disc surrounding IM Lupi, where we\nreport the first direct, i.e. model independent, measurements of the radial and\nvertical gradients of temperature and velocity in a protoplanetary disc. The\nmeasured disc structure is consistent with an irradiated self-similar disc\nstructure, where the temperature increases and the velocity decreases towards\nthe disc surface. We also directly map the vertical CO snow line, which is\nlocated at about one gas scale height at radii between 150 and 300 au, with a\nCO freeze-out temperature of $21\\pm2$ K. In the outer disc ($> 300$ au), where\nthe gas surface density transitions from a power law to an exponential taper,\nthe velocity rotation field becomes significantly sub-Keplerian, in agreement\nwith the expected steeper pressure gradient. The sub-Keplerian velocities\nshould result in a very efficient inward migration of large dust grains,\nexplaining the lack of millimetre continuum emission outside of 300 au. The\nsub-Keplerian motions may also be the signature of the base of an externally\nirradiated photo-evaporative wind. In the same outer region, the measured CO\ntemperature above the snow line decreases to $\\approx$ 15 K because of the\nreduced gas density, which can result in a lower CO freeze-out temperature,\nphoto-desorption, or deviations from local thermodynamic equilibrium.\n",
"title": "Direct mapping of the temperature and velocity gradients in discs. Imaging the vertical CO snow line around IM Lupi"
}
| null | null | null | null | true | null |
3825
| null |
Default
| null | null |
null |
{
"abstract": " Conditional expectiles are becoming an increasingly important tool in finance\nas well as in other areas of applications. We analyse a support vector machine\ntype approach for estimating conditional expectiles and establish learning\nrates that are minimax optimal modulo a logarithmic factor if Gaussian RBF\nkernels are used and the desired expectile is smooth in a Besov sense. As a\nspecial case, our learning rates improve the best known rates for kernel-based\nleast squares regression in this scenario. Key ingredients of our statistical\nanalysis are a general calibration inequality for the asymmetric least squares\nloss, a corresponding variance bound as well as an improved entropy number\nbound for Gaussian RBF kernels.\n",
"title": "Learning Rates for Kernel-Based Expectile Regression"
}
| null | null |
[
"Statistics"
] | null | true | null |
3826
| null |
Validated
| null | null |
null |
{
"abstract": " The choice that a solid system \"makes\" when adopting a crystal structure\n(stable or metastable) is ultimately governed by the interactions between\nelectrons forming chemical bonds. By analyzing 6 prototypical binary\ntransition-metal compounds we demonstrate here that the orbitally-selective\nstrong $d$-electron correlations influence dramatically the behavior of the\nenergy as a function of the spatial arrangements of the atoms. Remarkably, we\nfind that the main qualitative features of this complex behavior can be traced\nback to simple electrostatics, i.e., to the fact that the strong $d$-electron\ncorrelations influence substantially the charge transfer mechanism, which, in\nturn, controls the electrostatic interactions. This result advances our\nunderstanding of the influence of strong correlations on the crystal structure,\nopens a new avenue for extending structure prediction methodologies to strongly\ncorrelated materials, and paves the way for predicting and studying\nmetastability and polymorphism in these systems.\n",
"title": "Critical role of electronic correlations in determining crystal structure of transition metal compounds"
}
| null | null | null | null | true | null |
3827
| null |
Default
| null | null |
null |
{
"abstract": " This paper presents an alternative approach to p-values in regression\nsettings. This approach, whose origins can be traced to machine learning, is\nbased on the leave-one-out bootstrap for prediction error. In machine learning\nthis is called the out-of-bag (OOB) error. To obtain the OOB error for a model,\none draws a bootstrap sample and fits the model to the in-sample data. The\nout-of-sample prediction error for the model is obtained by calculating the\nprediction error for the model using the out-of-sample data. Repeating and\naveraging yields the OOB error, which represents a robust cross-validated\nestimate of the accuracy of the underlying model. By a simple modification to\nthe bootstrap data involving \"noising up\" a variable, the OOB method yields a\nvariable importance (VIMP) index, which directly measures how much a specific\nvariable contributes to the prediction precision of a model. VIMP provides a\nscientifically interpretable measure of the effect size of a variable, we call\nthe \"predictive effect size\", that holds whether the researcher's model is\ncorrect or not, unlike the p-value whose calculation is based on the assumed\ncorrectness of the model. We also discuss a marginal VIMP index, also easily\ncalculated, which measures the marginal effect of a variable, or what we call\n\"the discovery effect\". The OOB procedure can be applied to both parametric and\nnonparametric regression models and requires only that the researcher can\nrepeatedly fit their model to bootstrap and modified bootstrap data. We\nillustrate this approach on a survival data set involving patients with\nsystolic heart failure and to a simulated survival data set where the model is\nincorrectly specified to illustrate its robustness to model misspecification.\n",
"title": "A Machine Learning Alternative to P-values"
}
| null | null | null | null | true | null |
3828
| null |
Default
| null | null |
null |
{
"abstract": " Image Registration is the process of aligning two or more images of the same\nscene with reference to a particular image. The images are captured from\nvarious sensors at different times and at multiple view-points. Thus to get a\nbetter picture of any change of a scene or object over a considerable period of\ntime image registration is important. Image registration finds application in\nmedical sciences, remote sensing and in computer vision. This paper presents a\ndetailed review of several approaches which are classified accordingly along\nwith their contributions and drawbacks. The main steps of an image registration\nprocedure are also discussed. Different performance measures are presented that\ndetermine the registration quality and accuracy. The scope for the future\nresearch are presented as well.\n",
"title": "Image Registration Techniques: A Survey"
}
| null | null | null | null | true | null |
3829
| null |
Default
| null | null |
null |
{
"abstract": " This work initiates a general study of learning and generalization without\nthe i.i.d. assumption, starting from first principles. While the standard\napproach to statistical learning theory is based on assumptions chosen largely\nfor their convenience (e.g., i.i.d. or stationary ergodic), in this work we are\ninterested in developing a theory of learning based only on the most\nfundamental and natural assumptions implicit in the requirements of the\nlearning problem itself. We specifically study universally consistent function\nlearning, where the objective is to obtain low long-run average loss for any\ntarget function, when the data follow a given stochastic process. We are then\ninterested in the question of whether there exist learning rules guaranteed to\nbe universally consistent given only the assumption that universally consistent\nlearning is possible for the given data process. The reasoning that motivates\nthis criterion emanates from a kind of optimist's decision theory, and so we\nrefer to such learning rules as being optimistically universal. We study this\nquestion in three natural learning settings: inductive, self-adaptive, and\nonline. Remarkably, as our strongest positive result, we find that\noptimistically universal learning rules do indeed exist in the self-adaptive\nlearning setting. Establishing this fact requires us to develop new approaches\nto the design of learning algorithms. Along the way, we also identify concise\ncharacterizations of the family of processes under which universally consistent\nlearning is possible in the inductive and self-adaptive settings. We\nadditionally pose a number of enticing open problems, particularly for the\nonline learning setting.\n",
"title": "Learning Whenever Learning is Possible: Universal Learning under General Stochastic Processes"
}
| null | null | null | null | true | null |
3830
| null |
Default
| null | null |
null |
{
"abstract": " A graph is $H$-free if it has no induced subgraph isomorphic to $H$. We\ncharacterize all graphs $H$ for which there are only finitely many minimal\nnon-three-colorable $H$-free graphs. Such a characterization was previously\nknown only in the case when $H$ is connected. This solves a problem posed by\nGolovach et al. As a second result, we characterize all graphs $H$ for which\nthere are only finitely many $H$-free minimal obstructions for list\n3-colorability.\n",
"title": "Obstructions for three-coloring and list three-coloring $H$-free graphs"
}
| null | null | null | null | true | null |
3831
| null |
Default
| null | null |
null |
{
"abstract": " The design of electrically driven quantum dot devices for quantum optical\napplications asks for modeling approaches combining classical device physics\nwith quantum mechanics. We connect the well-established fields of\nsemi-classical semiconductor transport theory and the theory of open quantum\nsystems to meet this requirement. By coupling the van Roosbroeck system with a\nquantum master equation in Lindblad form, we introduce a new hybrid\nquantum-classical modeling approach, which provides a comprehensive description\nof quantum dot devices on multiple scales: It enables the calculation of\nquantum optical figures of merit and the spatially resolved simulation of the\ncurrent flow in realistic semiconductor device geometries in a unified way. We\nconstruct the interface between both theories in such a way, that the resulting\nhybrid system obeys the fundamental axioms of (non-)equilibrium thermodynamics.\nWe show that our approach guarantees the conservation of charge, consistency\nwith the thermodynamic equilibrium and the second law of thermodynamics. The\nfeasibility of the approach is demonstrated by numerical simulations of an\nelectrically driven single-photon source based on a single quantum dot in the\nstationary and transient operation regime.\n",
"title": "Hybrid quantum-classical modeling of quantum dot devices"
}
| null | null | null | null | true | null |
3832
| null |
Default
| null | null |
null |
{
"abstract": " This paper studies directed exploration for reinforcement learning agents by\ntracking uncertainty about the value of each available action. We identify two\nsources of uncertainty that are relevant for exploration. The first originates\nfrom limited data (parametric uncertainty), while the second originates from\nthe distribution of the returns (return uncertainty). We identify methods to\nlearn these distributions with deep neural networks, where we estimate\nparametric uncertainty with Bayesian drop-out, while return uncertainty is\npropagated through the Bellman equation as a Gaussian distribution. Then, we\nidentify that both can be jointly estimated in one network, which we call the\nDouble Uncertain Value Network. The policy is directly derived from the learned\ndistributions based on Thompson sampling. Experimental results show that both\ntypes of uncertainty may vastly improve learning in domains with a strong\nexploration challenge.\n",
"title": "Efficient exploration with Double Uncertain Value Networks"
}
| null | null | null | null | true | null |
3833
| null |
Default
| null | null |
null |
{
"abstract": " Bayesian models that mix multiple Dirichlet prior parameters, called\nMulti-Dirichlet priors (MD) in this paper, are gaining popularity. Inferring\nmixing weights and parameters of mixed prior distributions seems tricky, as\nsums over Dirichlet parameters complicate the joint distribution of model\nparameters.\nThis paper shows a novel auxiliary variable scheme which helps to simplify\nthe inference for models involving hierarchical MDs and MDPs. Using this\nscheme, it is easy to derive fully collapsed inference schemes which allow for\nan efficient inference.\n",
"title": "Auxiliary Variables for Multi-Dirichlet Priors"
}
| null | null | null | null | true | null |
3834
| null |
Default
| null | null |
null |
{
"abstract": " We present a new matched filter algorithm for direct detection of point\nsources in the immediate vicinity of bright stars. The stellar Point Spread\nFunction (PSF) is first subtracted using a Karhunen-Loéve Image Processing\n(KLIP) algorithm with Angular and Spectral Differential Imaging (ADI and SDI).\nThe KLIP-induced distortion of the astrophysical signal is included in the\nmatched filter template by computing a forward model of the PSF at every\nposition in the image. To optimize the performance of the algorithm, we conduct\nextensive planet injection and recovery tests and tune the exoplanet spectra\ntemplate and KLIP reduction aggressiveness to maximize the Signal-to-Noise\nRatio (SNR) of the recovered planets. We show that only two spectral templates\nare necessary to recover any young Jovian exoplanets with minimal SNR loss. We\nalso developed a complete pipeline for the automated detection of point source\ncandidates, the calculation of Receiver Operating Characteristics (ROC), false\npositives based contrast curves, and completeness contours. We process in a\nuniform manner more than 330 datasets from the Gemini Planet Imager Exoplanet\nSurvey (GPIES) and assess GPI typical sensitivity as a function of the star and\nthe hypothetical companion spectral type. This work allows for the first time a\ncomparison of different detection algorithms at a survey scale accounting for\nboth planet completeness and false positive rate. We show that the new forward\nmodel matched filter allows the detection of $50\\%$ fainter objects than a\nconventional cross-correlation technique with a Gaussian PSF template for the\nsame false positive rate.\n",
"title": "Improving and Assessing Planet Sensitivity of the GPI Exoplanet Survey with a Forward Model Matched Filter"
}
| null | null | null | null | true | null |
3835
| null |
Default
| null | null |
null |
{
"abstract": " 3D non-LTE radiative transfer problems are computationally demanding, and\nthis sets limits on the size of the problems that can be solved. So far\nMultilevel Accelerated Lambda Iteration (MALI) has been to the method of choice\nto perform high-resolution computations in multidimensional problems. The\ndisadvantage of MALI is that its computing time scales as $\\mathcal{O}(n^2)$,\nwith $n$ the number of grid points. When the grid gets finer, the computational\ncost increases quadratically. We aim to develop a 3D non-LTE radiative transfer\ncode that is more efficient than MALI. We implement a non-linear multigrid,\nfast approximation storage scheme, into the existing Multi3D radiative transfer\ncode. We verify our multigrid implementation by comparing with MALI\ncomputations. We show that multigrid can be employed in realistic problems with\nsnapshots from 3D radiative-MHD simulations as input atmospheres. With\nmultigrid, we obtain a factor 3.3-4.5 speedup compared to MALI. With\nfull-multigrid the speed-up increases to a factor 6. The speedup is expected to\nincrease for input atmospheres with more grid points and finer grid spacing.\nSolving 3D non-LTE radiative transfer problems using non-linear multigrid\nmethods can be applied to realistic atmospheres with a substantial speed-up.\n",
"title": "Numerical non-LTE 3D radiative transfer using a multigrid method"
}
| null | null |
[
"Physics"
] | null | true | null |
3836
| null |
Validated
| null | null |
null |
{
"abstract": " Recently, several Test Case Prioritization (TCP) techniques have been\nproposed to order test cases for achieving a goal during test execution,\nparticularly, revealing faults sooner. In the Model-Based Testing (MBT)\ncontext, such techniques are usually based on heuristics related to structural\nelements of the model and derived test cases. In this sense, techniques'\nperformance may vary due to a number of factors. While empirical studies\ncomparing the performance of TCP techniques have already been presented in\nliterature, there is still little knowledge, particularly in the MBT context,\nabout which factors may influence the outcomes suggested by a TCP technique. In\na previous family of empirical studies focusing on labeled transition systems,\nwe identified that the model layout, i.e. amount of branches, joins, and loops\nin the model, alone may have little influence on the performance of TCP\ntechniques investigated, whereas characteristics of test cases that actually\nfail definitely influences their performance. However, we considered only\nsynthetic artifacts in the study, which reduced the ability of representing\nproperly the reality. In this paper, we present a replication of one of these\nstudies, now with a larger and more representative selection of techniques and\nconsidering test suites from industrial applications as experimental objects.\nOur objective is to find out whether the results remain while increasing the\nvalidity in comparison to the original study. Results reinforce that there is\nno best performer among the investigated techniques and characteristics of test\ncases that fail represent an important factor, although adaptive random based\ntechniques are less affected by it.\n",
"title": "Test Case Prioritization Techniques for Model-Based Testing: A Replicated Study"
}
| null | null | null | null | true | null |
3837
| null |
Default
| null | null |
null |
{
"abstract": " The magnetism in Mn$_3$Si$_2$Te$_6$ has been investigated using thermodynamic\nmeasurements, first principles calculations, neutron diffraction and diffuse\nneutron scattering on single crystals. These data confirm that\nMn$_3$Si$_2$Te$_6$ is a ferrimagnet below a Curie temperature of $T_C$\napproximately 78K. The magnetism is anisotropic, with magnetization and neutron\ndiffraction demonstrating that the moments lie within the basal plane of the\ntrigonal structure. The saturation magnetization of approximately 1.6$\\mu_B$/Mn\nat 5K originates from the different multiplicities of the two\nantiferromagnetically-aligned Mn sites. First principles calculations reveal\nantiferromagnetic exchange for the three nearest Mn-Mn pairs, which leads to a\ncompetition between the ferrimagnetic ground state and three other magnetic\nconfigurations. The ferrimagnetic state results from the energy associated with\nthe third-nearest neighbor interaction, and thus long-range interactions are\nessential for the observed behavior. Diffuse magnetic scattering is observed\naround the 002 Bragg reflection at 120K, which indicates the presence of strong\nspin correlations well above $T_C$. These are promoted by the competing ground\nstates that result in a relative suppression of $T_C$, and may be associated\nwith a small ferromagnetic component that produces anisotropic magnetism below\n$\\approx$330K.\n",
"title": "Magnetic order and interactions in ferrimagnetic Mn3Si2Te6"
}
| null | null | null | null | true | null |
3838
| null |
Default
| null | null |
null |
{
"abstract": " The Big Data phenomenon has spawned large-scale linear programming problems.\nIn many cases, these problems are non-stationary. In this paper, we describe a\nnew scalable algorithm called NSLP for solving high-dimensional, non-stationary\nlinear programming problems on modern cluster computing systems. The algorithm\nconsists of two phases: Quest and Targeting. The Quest phase calculates a\nsolution of the system of inequalities defining the constraint system of the\nlinear programming problem under the condition of dynamic changes in input\ndata. To this end, the apparatus of Fejer mappings is used. The Targeting phase\nforms a special system of points having the shape of an n-dimensional\naxisymmetric cross. The cross moves in the n-dimensional space in such a way\nthat the solution of the linear programming problem is located all the time in\nan \"-vicinity of the central point of the cross.\n",
"title": "On the Solution of Linear Programming Problems in the Age of Big Data"
}
| null | null | null | null | true | null |
3839
| null |
Default
| null | null |
null |
{
"abstract": " This work is concerned with a unique combination of high order local\nabsorbing boundary conditions (ABC) with a general curvilinear Finite Element\nMethod (FEM) and its implementation in Isogeometric Analysis (IGA) for\ntime-harmonic acoustic waves. The ABC employed were recently devised by\nVillamizar, Acosta and Dastrup [J. Comput. Phys. 333 (2017) 331] . They are\nderived from exact Farfield Expansions representations of the outgoing waves in\nthe exterior of the regions enclosed by the artificial boundary. As a\nconsequence, the error due to the ABC on the artificial boundary can be reduced\nconveniently such that the dominant error comes from the volume discretization\nmethod used in the interior of the computational domain. Reciprocally, the\nerror in the interior can be made as small as the error at the artificial\nboundary by appropriate implementation of {\\it p-} and {\\it h}- refinement. We\napply this novel method to cylindrical, spherical and arbitrary shape\nscatterers including a prototype submarine. Our numerical results exhibits\nspectral-like approximation and high order convergence rate. Additionally, they\nshow that the proposed method can reduce both the pollution and artificial\nboundary errors to negligible levels even in very low- and high- frequency\nregimes with rather coarse discretization densities in the IGA. As a result, we\nhave developed a highly accurate computational platform to numerically solve\ntime-harmonic acoustic wave scattering in two- and three-dimensions.\n",
"title": "Highly accurate acoustic scattering: Isogeometric Analysis coupled with local high order Farfield Expansion ABC"
}
| null | null | null | null | true | null |
3840
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, a deep domain adaptation based method for video smoke\ndetection is proposed to extract a powerful feature representation of smoke.\nDue to the smoke image samples limited in scale and diversity for deep CNN\ntraining, we systematically produced adequate synthetic smoke images with a\nwide variation in the smoke shape, background and lighting conditions.\nConsidering that the appearance gap (dataset bias) between synthetic and real\nsmoke images degrades significantly the performance of the trained model on the\ntest set composed fully of real images, we build deep architectures based on\ndomain adaptation to confuse the distributions of features extracted from\nsynthetic and real smoke images. This approach expands the domain-invariant\nfeature space for smoke image samples. With their approximate feature\ndistribution off non-smoke images, the recognition rate of the trained model is\nimproved significantly compared to the model trained directly on mixed dataset\nof synthetic and real images. Experimentally, several deep architectures with\ndifferent design choices are applied to the smoke detector. The ultimate\nframework can get a satisfactory result on the test set. We believe that our\napproach is a start in the direction of utilizing deep neural networks enhanced\nwith synthetic smoke images for video smoke detection.\n",
"title": "Deep Domain Adaptation Based Video Smoke Detection using Synthetic Smoke Images"
}
| null | null | null | null | true | null |
3841
| null |
Default
| null | null |
null |
{
"abstract": " Under ambient conditions, we directly observed NaCl crystals experimentally\nin the rGO membranes soaked in the salt solution with concentration below and\nfar below the saturated concentration. Moreover, in most probability, the NaCl\ncrystals show stoichiometries behavior. We attribute this unexpected\ncrystallization to the cation-{\\pi} interactions between the ions and the\naromatic rings of the rGO.\n",
"title": "NaCl crystal from salt solution with far below saturated concentration under ambient condition"
}
| null | null | null | null | true | null |
3842
| null |
Default
| null | null |
null |
{
"abstract": " The development of plasmonic and metamaterial devices requires the research\nof high-performance materials, alternative to standard noble metals. Renewed as\nrefractory stable compound for durable coatings, titanium nitride has been\nrecently proposed as an efficient plasmonic material. Here, by using a first\nprinciples approach, we investigate the plasmon dispersion relations of TiN\nbulk and we predict the effect of pressure on its optoelectronic properties.\nOur results explain the main features of TiN in the visible range and prove a\nuniversal scaling law which relates its mechanical and plasmonic properties as\na function of pressure. Finally, we address the formation and stability of\nsurface-plasmon polaritons at different TiN/dielectric interfaces proposed by\nrecent experiments. The unusual combination of plasmonics and refractory\nfeatures paves the way for the realization of plasmonic devices able to work at\nconditions not sustainable by usual noble metals.\n",
"title": "Plasmonic properties of refractory titanium nitride"
}
| null | null |
[
"Physics"
] | null | true | null |
3843
| null |
Validated
| null | null |
null |
{
"abstract": " We study the Anderson-like localization transition in the spectrum of the\nDirac operator of quenched QCD. Above the deconfining transition we determine\nthe temperature dependence of the mobility edge separating localized and\ndelocalized eigenmodes in the spectrum. We show that the temperature where the\nmobility edge vanishes and localized modes disappear from the spectrum,\ncoincides with the critical temperature of the deconfining transition. We also\nidentify topological charge related close to zero modes in the Dirac spectrum\nand show that they account for only a small fraction of localized modes, a\nfraction that is rapidly falling as the temperature increases.\n",
"title": "The localization transition in SU(3) gauge theory"
}
| null | null |
[
"Physics"
] | null | true | null |
3844
| null |
Validated
| null | null |
null |
{
"abstract": " Recent developments in specialized computer hardware have greatly accelerated\natomic level Molecular Dynamics (MD) simulations. A single GPU-attached cluster\nis capable of producing microsecond-length trajectories in reasonable amounts\nof time. Multiple protein states and a large number of microstates associated\nwith folding and with the function of the protein can be observed as\nconformations sampled in the trajectories. Clustering those conformations,\nhowever, is needed for identifying protein states, evaluating transition rates\nand understanding protein behavior. In this paper, we propose a novel\ndata-driven generative conformation clustering method based on the adversarial\nautoencoder (AAE) and provide the associated software implementation Cong. The\nmethod was tested using a 208 microseconds MD simulation of the fast-folding\npeptide Trp-Cage (20 residues) obtained from the D.E. Shaw Research Group. The\nproposed clustering algorithm identifies many of the salient features of the\nfolding process by grouping a large number of conformations that share common\nfeatures not easily identifiable in the trajectory.\n",
"title": "Conformation Clustering of Long MD Protein Dynamics with an Adversarial Autoencoder"
}
| null | null | null | null | true | null |
3845
| null |
Default
| null | null |
null |
{
"abstract": " Dynamic Mode Decomposition (DMD) has emerged as a powerful tool for analyzing\nthe dynamics of non-linear systems from experimental datasets. Recently,\nseveral attempts have extended DMD to the context of low-rank approximations.\nThis extension is of particular interest for reduced-order modeling in various\napplicative domains, e.g. for climate prediction, to study molecular dynamics\nor micro-electromechanical devices. This low-rank extension takes the form of a\nnon-convex optimization problem. To the best of our knowledge, only sub-optimal\nalgorithms have been proposed in the literature to compute the solution of this\nproblem. In this paper, we prove that there exists a closed-form optimal\nsolution to this problem and design an effective algorithm to compute it based\non Singular Value Decomposition (SVD). A toy-example illustrates the gain in\nperformance of the proposed algorithm compared to state-of-the-art techniques.\n",
"title": "Optimal Low-Rank Dynamic Mode Decomposition"
}
| null | null | null | null | true | null |
3846
| null |
Default
| null | null |
null |
{
"abstract": " Social relationships can be divided into different classes based on the\nregularity with which they occur and the similarity among them. Thus, rare and\nsomewhat similar relationships are random and cause noise in a social network,\nthus hiding the actual structure of the network and preventing an accurate\nanalysis of it. In this context, in this paper we propose a process to handle\nsocial network data that exploits temporal features to improve the detection of\ncommunities by existing algorithms. By removing random interactions, we observe\nthat social networks converge to a topology with more purely social\nrelationships and more modular communities.\n",
"title": "Improving Community Detection by Mining Social Interactions"
}
| null | null |
[
"Computer Science"
] | null | true | null |
3847
| null |
Validated
| null | null |
null |
{
"abstract": " We construct a linear system non-local game which can be played perfectly\nusing a limit of finite-dimensional quantum strategies, but which cannot be\nplayed perfectly on any finite-dimensional Hilbert space, or even with any\ntensor-product strategy. In particular, this shows that the set of\n(tensor-product) quantum correlations is not closed. The constructed non-local\ngame provides another counterexample to the \"middle\" Tsirelson problem, with a\nshorter proof than our previous paper (though at the loss of the universal\nembedding theorem). We also show that it is undecidable to determine if a\nlinear system game can be played perfectly with a finite-dimensional strategy,\nor a limit of finite-dimensional quantum strategies.\n",
"title": "The set of quantum correlations is not closed"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3848
| null |
Validated
| null | null |
null |
{
"abstract": " This note corrects the mistakes in the splicing formulas of the paper \"Floer\nhomology and splicing knot complements\". The mistakes are the result of the\nincorrect assumption that for a knot $K$ inside a homology sphere $Y$, the\ninvolution on the knot Floer homology of $K$ which corresponds to moving the\nbasepoints by one full twist around $K$ is trivial. The correction implicitly\ninvolves considering the contribution from this (possibly non-trivial)\ninvolution in a number of places.\n",
"title": "Correction to the article: Floer homology and splicing knot complements"
}
| null | null | null | null | true | null |
3849
| null |
Default
| null | null |
null |
{
"abstract": " We show tight upper and lower bounds for switching lemmas obtained by the\naction of random $p$-restrictions on boolean functions that can be expressed as\ndecision trees in which every vertex is at a distance of at most $t$ from some\nleaf, also called $t$-clipped decision trees. More specifically, we show the\nfollowing:\n$\\bullet$ If a boolean function $f$ can be expressed as a $t$-clipped\ndecision tree, then under the action of a random $p$-restriction $\\rho$, the\nprobability that the smallest depth decision tree for $f|_{\\rho}$ has depth\ngreater than $d$ is upper bounded by $(4p2^{t})^{d}$.\n$\\bullet$ For every $t$, there exists a function $g_{t}$ that can be\nexpressed as a $t$-clipped decision tree, such that under the action of a\nrandom $p$-restriction $\\rho$, the probability that the smallest depth decision\ntree for $g_{t}|_{\\rho}$ has depth greater than $d$ is lower bounded by\n$(c_{0}p2^{t})^{d}$, for $0\\leq p\\leq c_{p}2^{-t}$ and $0\\leq d\\leq\nc_{d}\\frac{\\log n}{2^{t}\\log t}$, where $c_{0},c_{p},c_{d}$ are universal\nconstants.\n",
"title": "Tree tribes and lower bounds for switching lemmas"
}
| null | null | null | null | true | null |
3850
| null |
Default
| null | null |
null |
{
"abstract": " The kinetic effects of electrons are important to long wavelength\nmagnetohydrodynamic(MHD)instabilities and short wavelength drift-Alfvenic\ninstabilities responsible for turbulence transport in magnetized plasmas, since\nthe non-adiabatic electron can interact with, modify and drive the low\nfrequency instabilities. A novel conservative split weight scheme is proposed\nfor the electromagnetic simulation with drift kinetic electrons in tokamak\nplasmas, which shows great computational advantages that there is no numerical\nconstrain of electron skin depth on the perpendicular grid size without\nsacrificing any physics. Both kinetic Alfven wave and collision-less tearing\nmode are verified by using this model, which has already been implemented into\nthe gyrokinetic toroidal code(GTC). This model will be used for the micro\ntearing mode and neoclassical tearing mode simulation based on the first\nprinciple in the future.\n",
"title": "Gyrokinetic ion and drift kinetic electron model for electromagnetic simulation in the toroidal geometry"
}
| null | null | null | null | true | null |
3851
| null |
Default
| null | null |
null |
{
"abstract": " A major bottleneck for developing general reinforcement learning agents is\ndetermining rewards that will yield desirable behaviors under various\ncircumstances. We introduce a general mechanism for automatically specifying\nmeaningful behaviors from raw pixels. In particular, we train a generative\nadversarial network to produce short sub-goals represented through motion\ntemplates. We demonstrate that this approach generates visually meaningful\nbehaviors in unknown environments with novel agents and describe how these\nmotions can be used to train reinforcement learning agents.\n",
"title": "Transferring Agent Behaviors from Videos via Motion GANs"
}
| null | null | null | null | true | null |
3852
| null |
Default
| null | null |
null |
{
"abstract": " Developing an intelligent vehicle which can perform human-like actions\nrequires the ability to learn basic driving skills from a large amount of\nnaturalistic driving data. The algorithms will become efficient if we could\ndecompose the complex driving tasks into motion primitives which represent the\nelementary compositions of driving skills. Therefore, the purpose of this paper\nis to segment unlabeled trajectory data into a library of motion primitives. By\napplying a probabilistic inference based on an iterative\nExpectation-Maximization algorithm, our method segments the collected\ntrajectories while learning a set of motion primitives represented by the\ndynamic movement primitives. The proposed method utilizes the mutual\ndependencies between the segmentation and representation of motion primitives\nand the driving-specific based initial segmentation. By utilizing this mutual\ndependency and the initial condition, this paper presents how we can enhance\nthe performance of both the segmentation and the motion primitive library\nestablishment. We also evaluate the applicability of the primitive\nrepresentation method to imitation learning and motion planning algorithms. The\nmodel is trained and validated by using the driving data collected from the\nBeijing Institute of Technology intelligent vehicle platform. The results show\nthat the proposed approach can find the proper segmentation and establish the\nmotion primitive library simultaneously.\n",
"title": "Learning to Segment and Represent Motion Primitives from Driving Data for Motion Planning Applications"
}
| null | null | null | null | true | null |
3853
| null |
Default
| null | null |
null |
{
"abstract": " \"How much energy is consumed for an inference made by a convolutional neural\nnetwork (CNN)?\" With the increased popularity of CNNs deployed on the\nwide-spectrum of platforms (from mobile devices to workstations), the answer to\nthis question has drawn significant attention. From lengthening battery life of\nmobile devices to reducing the energy bill of a datacenter, it is important to\nunderstand the energy efficiency of CNNs during serving for making an\ninference, before actually training the model. In this work, we propose\nNeuralPower: a layer-wise predictive framework based on sparse polynomial\nregression, for predicting the serving energy consumption of a CNN deployed on\nany GPU platform. Given the architecture of a CNN, NeuralPower provides an\naccurate prediction and breakdown for power and runtime across all layers in\nthe whole network, helping machine learners quickly identify the power,\nruntime, or energy bottlenecks. We also propose the \"energy-precision ratio\"\n(EPR) metric to guide machine learners in selecting an energy-efficient CNN\narchitecture that better trades off the energy consumption and prediction\naccuracy. The experimental results show that the prediction accuracy of the\nproposed NeuralPower outperforms the best published model to date, yielding an\nimprovement in accuracy of up to 68.5%. We also assess the accuracy of\npredictions at the network level, by predicting the runtime, power, and energy\nof state-of-the-art CNN architectures, achieving an average accuracy of 88.24%\nin runtime, 88.34% in power, and 97.21% in energy. We comprehensively\ncorroborate the effectiveness of NeuralPower as a powerful framework for\nmachine learners by testing it on different GPU platforms and Deep Learning\nsoftware tools.\n",
"title": "NeuralPower: Predict and Deploy Energy-Efficient Convolutional Neural Networks"
}
| null | null | null | null | true | null |
3854
| null |
Default
| null | null |
null |
{
"abstract": " We investigate the association between musical chords and lyrics by analyzing\na large dataset of user-contributed guitar tablatures. Motivated by the idea\nthat the emotional content of chords is reflected in the words used in\ncorresponding lyrics, we analyze associations between lyrics and chord\ncategories. We also examine the usage patterns of chords and lyrics in\ndifferent musical genres, historical eras, and geographical regions. Our\noverall results confirms a previously known association between Major chords\nand positive valence. We also report a wide variation in this association\nacross regions, genres, and eras. Our results suggest possible existence of\ndifferent emotional associations for other types of chords.\n",
"title": "The Minor Fall, the Major Lift: Inferring Emotional Valence of Musical Chords through Lyrics"
}
| null | null | null | null | true | null |
3855
| null |
Default
| null | null |
null |
{
"abstract": " We propose a general algorithm to compute all the symmetry classes of any\n$\\mathrm{SO}(3)$ or $\\mathrm{O}(3)$ linear representation. This method relies\non the introduction of a binary operator between sets of conjugacy classes of\nclosed subgroups, called the clips. We compute explicit tables for this\noperation which allows to solve definitively the problem.\n",
"title": "Effective computation of $\\mathrm{SO}(3)$ and $\\mathrm{O}(3)$ linear representations symmetry classes"
}
| null | null | null | null | true | null |
3856
| null |
Default
| null | null |
null |
{
"abstract": " Using cohomological methods, we prove a criterion for the embedding of a\ngroup extension with abelian kernel into the split extension of a co-induced\nmodule. This generalises some earlier similar results. We also prove an\nassertion about the conjugacy of complements in split extensions of co-induced\nmodules. Both results follow from a relation between homomorphisms of certain\ncohomology groups.\n",
"title": "Subextensions for co-induced modules"
}
| null | null | null | null | true | null |
3857
| null |
Default
| null | null |
null |
{
"abstract": " The generators of the classical Specht module satisfy intricate relations. We\nintroduce the Specht matroid, which keeps track of these relations, and the\nSpecht polytope, which also keeps track of convexity relations. We establish\nbasic facts about the Specht polytope, for example, that the symmetric group\nacts transitively on its vertices and irreducibly on its ambient real vector\nspace. A similar construction builds a matroid and polytope for a tensor\nproduct of Specht modules, giving \"Kronecker matroids\" and \"Kronecker\npolytopes\" instead of the usual Kronecker coefficients. We dub this process of\nupgrading numbers to matroids and polytopes \"matroidification,\" giving two more\nexamples. In the course of describing these objects, we also give an elementary\naccount of the construction of Specht modules different from the standard one.\nFinally, we provide code to compute with Specht matroids and their Chow rings.\n",
"title": "Specht Polytopes and Specht Matroids"
}
| null | null | null | null | true | null |
3858
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we investigate the Cauchy problem of the nonhomogeneous\nincompressible non-resistive MHD on $\\mathbb{R}^2$ with vacuum as far field\ndensity and prove that the 2D Cauchy problem has a unique local strong solution\nprovided that the initial density and magnetic field decay not too slow at\ninfinity. Furthermore, if the initial data satisfy some additional regularity\nand compatibility conditions, the strong solution becomes a classical one.\n",
"title": "Existence theorems for the Cauchy problem of 2D nonhomogeneous incompressible non-resistive MHD equations with vacuum"
}
| null | null | null | null | true | null |
3859
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we describe two fully mass conservative, energy stable, finite\ndifference methods on a staggered grid for the quasi-incompressible\nNavier-Stokes-Cahn-Hilliard (q-NSCH) system governing a binary incompressible\nfluid flow with variable density and viscosity. Both methods, namely the\nprimitive method (finite difference method in the primitive variable\nformulation) and the projection method (finite difference method in a\nprojection-type formulation), are so designed that the mass of the binary fluid\nis preserved, and the energy of the system equations is always non-increasing\nin time at the fully discrete level. We also present an efficient, practical\nnonlinear multigrid method - comprised of a standard FAS method for the\nCahn-Hilliard equation, and a method based on the Vanka-type smoothing strategy\nfor the Navier-Stokes equation - for solving these equations. We test the\nscheme in the context of Capillary Waves, rising droplets and Rayleigh-Taylor\ninstability. Quantitative comparisons are made with existing analytical\nsolutions or previous numerical results that validate the accuracy of our\nnumerical schemes. Moreover, in all cases, mass of the single component and the\nbinary fluid was conserved up to 10 to -8 and energy decreases in time.\n",
"title": "Mass Conservative and Energy Stable Finite Difference Methods for the Quasi-incompressible Navier-Stokes-Cahn-Hilliard system: Primitive Variable and Projection-Type Schemes"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3860
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper, we are concerned with the problem of creating flattening maps\nof simply-connected open surfaces in $\\mathbb{R}^3$. Using a natural principle\nof density diffusion in physics, we propose an effective algorithm for\ncomputing density-equalizing flattening maps with any prescribed density\ndistribution. By varying the initial density distribution, a large variety of\nmappings with different properties can be achieved. For instance,\narea-preserving parameterizations of simply-connected open surfaces can be\neasily computed. Experimental results are presented to demonstrate the\neffectiveness of our proposed method. Applications to data visualization and\nsurface remeshing are explored.\n",
"title": "Density-equalizing maps for simply-connected open surfaces"
}
| null | null | null | null | true | null |
3861
| null |
Default
| null | null |
null |
{
"abstract": " Batygin and Brown (2016) have suggested the existence of a new Solar System\nplanet supposed to be responsible for the perturbation of eccentric orbits of\nsmall outer bodies. The main challenge is now to detect and characterize this\nputative body. Here we investigate the principles of the determination of its\nphysical parameters, mainly its mass and radius. For that purpose we\nconcentrate on two methods, stellar occultations and gravitational microlensing\neffects (amplification, deflection and time delay). We estimate the main\ncharacteristics of a possible occultation or gravitational effects: flux\nvariation of a background star, duration and probability of occurence. We\ninvestigate also additional benefits of direct imaging and of an occultation.\n",
"title": "Measuring the radius and mass of Planet Nine"
}
| null | null | null | null | true | null |
3862
| null |
Default
| null | null |
null |
{
"abstract": " We propose a fast method with statistical guarantees for learning an\nexponential family density model where the natural parameter is in a\nreproducing kernel Hilbert space, and may be infinite-dimensional. The model is\nlearned by fitting the derivative of the log density, the score, thus avoiding\nthe need to compute a normalization constant. Our approach improves the\ncomputational efficiency of an earlier solution by using a low-rank,\nNyström-like solution. The new solution retains the consistency and\nconvergence rates of the full-rank solution (exactly in Fisher distance, and\nnearly in other distances), with guarantees on the degree of cost and storage\nreduction. We evaluate the method in experiments on density estimation and in\nthe construction of an adaptive Hamiltonian Monte Carlo sampler. Compared to an\nexisting score learning approach using a denoising autoencoder, our estimator\nis empirically more data-efficient when estimating the score, runs faster, and\nhas fewer parameters (which can be tuned in a principled and interpretable\nway), in addition to providing statistical guarantees.\n",
"title": "Efficient and principled score estimation with Nyström kernel exponential families"
}
| null | null | null | null | true | null |
3863
| null |
Default
| null | null |
null |
{
"abstract": " We prove existence results for small presentations of model categories\ngeneralizing a theorem of D. Dugger from combinatorial model categories to more\ngeneral model categories. Some of these results are shown under the assumption\nof Vopěnka's principle. Our main theorem applies in particular to\ncofibrantly generated model categories where the domains of the generating\ncofibrations satisfy a slightly stronger smallness condition. As a consequence,\nassuming Vopěnka's principle, such a cofibrantly generated model category\nis Quillen equivalent to a combinatorial model category. Moreover, if there are\ngenerating sets which consist of presentable objects, then the same conclusion\nholds without the assumption of Vopěnka's principle. We also correct a\nmistake from previous work that made similar claims.\n",
"title": "Small presentations of model categories and Vopěnka's principle"
}
| null | null | null | null | true | null |
3864
| null |
Default
| null | null |
null |
{
"abstract": " Image matting is a longstanding problem in computational photography.\nAlthough, it has been studied for more than two decades, yet there is a\nchallenge of developing an automatic matting algorithm which does not require\nany human efforts. Most of the state-of-the-art matting algorithms require\nhuman intervention in the form of trimap or scribbles to generate the alpha\nmatte form the input image. In this paper, we present a simple and efficient\napproach to automatically generate the trimap from the input image and make the\nwhole matting process free from human-in-the-loop. We use learning based\nmatting method to generate the matte from the automatically generated trimap.\nExperimental results demonstrate that our method produces good quality trimap\nwhich results into accurate matte estimation. We validate our results by\nreplacing the automatically generated trimap by manually created trimap while\nusing the same image matting algorithm.\n",
"title": "Automatic Trimap Generation for Image Matting"
}
| null | null | null | null | true | null |
3865
| null |
Default
| null | null |
null |
{
"abstract": " We present the combined Chandra and Swift-BAT spectral analysis of seven\nSeyfert 2 galaxies selected from the Swift-BAT 100-month catalog. We selected\nnearby (z<=0.03) sources lacking of a ROSAT counterpart and never previously\nobserved with Chandra in the 0.3-10 keV energy range, and targeted these\nobjects with 10 ks Chandra ACIS-S observations. The X-ray spectral fitting over\nthe 0.3-150 keV energy range allows us to determine that all the objects are\nsignificantly obscured, having NH>=1E23 cm^(-2) at a >99% confidence level.\nMoreover, one to three sources are candidate Compton thick Active Galactic\nNuclei (CT-AGN), i.e., have NH>=1E24 cm^(-2). We also test the recent \"spectral\ncurvature\" method developed by Koss et al. (2016) to find candidate CT-AGN,\nfinding a good agreement between our results and their predictions. Since the\nselection criteria we adopted have been effective in detecting highly obscured\nAGN, further observations of these and other Seyfert 2 galaxies selected from\nthe Swift-BAT 100-month catalog will allow us to create a statistically\nsignificant sample of highly obscured AGN, therefore better understanding the\nphysics of the obscuration processes.\n",
"title": "X-ray spectral properties of seven heavily obscured Seyfert 2 galaxies"
}
| null | null | null | null | true | null |
3866
| null |
Default
| null | null |
null |
{
"abstract": " LISA is a proposed space-based laser interferometer detecting gravitational\nwaves by measuring distances between free-floating test masses housed in three\nsatellites in a triangular constellation with laser links in-between. Each\nsatellite contains two optical benches that are articulated by moving optical\nsubassemblies for compensating the breathing angle in the constellation. The\nphase reference distribution system, also known as backlink, forms an optical\nbi-directional path between the intra-satellite benches.\nIn this work we discuss phase reference implementations with a target\nnon-reciprocity of at most $2\\pi\\,\\mathrm{\\mu rad/\\sqrt{Hz}}$, equivalent to\n$1\\,\\mathrm{pm/\\sqrt{Hz}}$ for a wavelength of $1064\\,\\mathrm{nm}$ in the\nfrequency band from $0.1\\,\\mathrm{mHz}$ to $1\\,\\mathrm{Hz}$. One phase\nreference uses a steered free beam connection, the other one a fiber together\nwith additional laser frequencies. The noise characteristics of these\nimplementations will be compared in a single interferometric set-up with a\npreviously successfully tested direct fiber connection. We show the design of\nthis interferometer created by optical simulations including ghost beam\nanalysis, component alignment and noise estimation. First experimental results\nof a free beam laser link between two optical set-ups that are co-rotating by\n$\\pm 1^\\circ$ are presented. This experiment demonstrates sufficient thermal\nstability during rotation of less than $10^{-4}\\,\\mathrm{K/\\sqrt{Hz}}$ at\n$1\\,\\mathrm{mHz}$ and operation of the free beam steering mirror control over\nmore than 1 week.\n",
"title": "Towards the LISA Backlink: Experiment design for comparing optical phase reference distribution systems"
}
| null | null | null | null | true | null |
3867
| null |
Default
| null | null |
null |
{
"abstract": " Generative Adversarial Networks (GANs) have shown remarkable success as a\nframework for training models to produce realistic-looking data. In this work,\nwe propose a Recurrent GAN (RGAN) and Recurrent Conditional GAN (RCGAN) to\nproduce realistic real-valued multi-dimensional time series, with an emphasis\non their application to medical data. RGANs make use of recurrent neural\nnetworks in the generator and the discriminator. In the case of RCGANs, both of\nthese RNNs are conditioned on auxiliary information. We demonstrate our models\nin a set of toy datasets, where we show visually and quantitatively (using\nsample likelihood and maximum mean discrepancy) that they can successfully\ngenerate realistic time-series. We also describe novel evaluation methods for\nGANs, where we generate a synthetic labelled training dataset, and evaluate on\na real test set the performance of a model trained on the synthetic data, and\nvice-versa. We illustrate with these metrics that RCGANs can generate\ntime-series data useful for supervised training, with only minor degradation in\nperformance on real test data. This is demonstrated on digit classification\nfrom 'serialised' MNIST and by training an early warning system on a medical\ndataset of 17,000 patients from an intensive care unit. We further discuss and\nanalyse the privacy concerns that may arise when using RCGANs to generate\nrealistic synthetic medical time series data.\n",
"title": "Real-valued (Medical) Time Series Generation with Recurrent Conditional GANs"
}
| null | null | null | null | true | null |
3868
| null |
Default
| null | null |
null |
{
"abstract": " We provide examples of operators $T(D)+V$ with decaying potentials that have\nembedded eigenvalues. The decay of the potential depends on the curvature of\nthe Fermi surfaces of constant kinetic energy $T$. We make the connection to\ncounterexamples in Fourier restriction theory.\n",
"title": "Embedded eigenvalues of generalized Schrödinger operators"
}
| null | null | null | null | true | null |
3869
| null |
Default
| null | null |
null |
{
"abstract": " In real human robot interaction (HRI) scenarios, speech recognition\nrepresents a major challenge due to robot noise, background noise and\ntime-varying acoustic channel. This document describes the procedure used to\nobtain the Multichannel Robot Speech Recognition Database (MChRSR). It is\ncomposed of 12 hours of multichannel evaluation data recorded in a real mobile\nHRI scenario. This database was recorded with a PR2 robot performing different\ntranslational and azimuthal movements. Accordingly, 16 evaluation sets were\nobtained re-recording the clean set of the Aurora 4 database in different\nmovement conditions.\n",
"title": "Multichannel Robot Speech Recognition Database: MChRSR"
}
| null | null | null | null | true | null |
3870
| null |
Default
| null | null |
null |
{
"abstract": " We present a deterministic algorithm for Russian inflection. This algorithm\nis implemented in a publicly available web-service www.passare.ru which\nprovides functions for inflection of single words, word matching and synthesis\nof grammatically correct Russian text. The inflectional functions have been\ntested against the annotated corpus of Russian language OpenCorpora.\n",
"title": "The Algorithmic Inflection of Russian and Generation of Grammatically Correct Text"
}
| null | null | null | null | true | null |
3871
| null |
Default
| null | null |
null |
{
"abstract": " Objective: A clinical decision support tool that automatically interprets\nEEGs can reduce time to diagnosis and enhance real-time applications such as\nICU monitoring. Clinicians have indicated that a sensitivity of 95% with a\nspecificity below 5% was the minimum requirement for clinical acceptance. We\npropose a highperformance classification system based on principles of big data\nand machine learning. Methods: A hybrid machine learning system that uses\nhidden Markov models (HMM) for sequential decoding and deep learning networks\nfor postprocessing is proposed. These algorithms were trained and evaluated\nusing the TUH EEG Corpus, which is the world's largest publicly available\ndatabase of clinical EEG data. Results: Our approach delivers a sensitivity\nabove 90% while maintaining a specificity below 5%. This system detects three\nevents of clinical interest: (1) spike and/or sharp waves, (2) periodic\nlateralized epileptiform discharges, (3) generalized periodic epileptiform\ndischarges. It also detects three events used to model background noise: (1)\nartifacts, (2) eye movement (3) background. Conclusions: A hybrid HMM/deep\nlearning system can deliver a low false alarm rate on EEG event detection,\nmaking automated analysis a viable option for clinicians. Significance: The TUH\nEEG Corpus enables application of highly data consumptive machine learning\nalgorithms to EEG analysis. Performance is approaching clinical acceptance for\nreal-time applications.\n",
"title": "Automatic Analysis of EEGs Using Big Data and Hybrid Deep Learning Architectures"
}
| null | null | null | null | true | null |
3872
| null |
Default
| null | null |
null |
{
"abstract": " The same concept can mean different things or be instantiated in different\nforms depending on context, suggesting a degree of flexibility within the\nconceptual system. We propose that a compositional network model can be used to\ncapture and predict this flexibility. We modeled individual concepts (e.g.,\nBANANA, BOTTLE) as graph-theoretical networks, in which properties (e.g.,\nYELLOW, SWEET) were represented as nodes and their associations as edges. In\nthis framework, networks capture the within-concept statistics that reflect how\nproperties correlate with each other across instances of a concept. We ran a\nclassification analysis using graph eigendecomposition to validate these\nmodels, and find that these models can successfully discriminate between object\nconcepts. We then computed formal measures from these concept networks and\nexplored their relationship to conceptual structure. We find that diversity\ncoefficients and core-periphery structure can be interpreted as network-based\nmeasures of conceptual flexibility and stability, respectively. These results\nsupport the feasibility of a concept network framework and highlight its\nability to formally capture important characteristics of the conceptual system.\n",
"title": "Implementing a Concept Network Model"
}
| null | null | null | null | true | null |
3873
| null |
Default
| null | null |
null |
{
"abstract": " M. Hanzer and I. Matic have proved that the genuine unitary principal series\nrepresentations of the metaplectic groups are irreducible. A simple consequence\nof that paper is a criterion for the irreducibility of the non-unitary\nprincipal series representations of the metaplectic groups that we give in this\npaper.\n",
"title": "Two simple observations on representations of metaplectic groups"
}
| null | null | null | null | true | null |
3874
| null |
Default
| null | null |
null |
{
"abstract": " This article explores the geometric algebra of Minkowski spacetime, and its\nrelationship to the geometric algebra of Euclidean 4-space. Both of these\ngeometric algebras are algebraically isomorphic to the 2x2 matrix algebra over\nHamilton's famous quaternions, and provide a rich geometric framework for\nvarious important topics in mathematics and physics, including stereographic\nprojection and spinors, and both spherical and hyperbolic geometry. In\naddition, by identifying the time-like Minkowski unit vector with the extra\ndimension of Euclidean 4-space, David Hestenes' Space-Time Algebra of Minkowski\nspacetime is unified with William Baylis' Algebra of Physical Space.\n",
"title": "Spinors in Spacetime Algebra and Euclidean 4-Space"
}
| null | null | null | null | true | null |
3875
| null |
Default
| null | null |
null |
{
"abstract": " We are concerned with unbounded sets of $\\mathbb{R}^N$ whose boundary has\nconstant nonlocal (or fractional) mean curvature, which we call CNMC sets. This\nis the equation associated to critical points of the fractional perimeter\nfunctional under a volume constraint. We construct CNMC sets which are the\ncountable union of a certain bounded domain and all its translations through a\nperiodic integer lattice of dimension $M\\leq N$. Our CNMC sets form a $C^2$\nbranch emanating from the unit ball alone and where the parameter in the branch\nis essentially the distance to the closest lattice point. Thus, the new\ntranslated near-balls (or near-spheres) appear from infinity. We find their\nexact asymptotic shape as the parameter tends to infinity.\n",
"title": "Near-sphere lattices with constant nonlocal mean curvature"
}
| null | null | null | null | true | null |
3876
| null |
Default
| null | null |
null |
{
"abstract": " Mathematical models for physiological processes aid qualitative understanding\nof the impact of various parameters on the underlying process. We analyse two\nsuch models for human physiological processes: the Mackey-Glass and the Lasota\nequations, which model the change in the concentration of blood cells in the\nhuman body. We first study the local stability of these models, and derive\nbounds on various model parameters and the feedback delay for the concentration\nto equilibrate. We then deduce conditions for non-oscillatory convergence of\nthe solutions, which could ensure that the blood cell concentration does not\noscillate. Further, we define the convergence characteristics of the solutions\nwhich govern the rate at which the concentration equilibrates when the system\nis stable. Owing to the possibility that physiological parameters can seldom be\nestimated precisely, we also derive bounds for robust stability\\textemdash\nwhich enable one to ensure that the blood cell concentration equilibrates\ndespite parametric uncertainty. We also highlight that when the necessary and\nsufficient condition for local stability is violated, the system transits into\ninstability via a Hopf bifurcation, leading to limit cycles in the blood cell\nconcentration. We then outline a framework to characterise the type of the Hopf\nbifurcation and determine the asymptotic orbital stability of limit cycles. The\nanalysis is complemented with numerical examples, stability charts and\nbifurcation diagrams. The insights into the dynamical properties of the\nmathematical models may serve to guide the study of dynamical diseases.\n",
"title": "Stability, convergence, and limit cycles in some human physiological processes"
}
| null | null | null | null | true | null |
3877
| null |
Default
| null | null |
null |
{
"abstract": " We provide a derivation of the Poisson multi-Bernoulli mixture (PMBM) filter\nfor multi-target tracking with the standard point target measurements without\nusing probability generating functionals or functional derivatives. We also\nestablish the connection with the \\delta-generalised labelled multi-Bernoulli\n(\\delta-GLMB) filter, showing that a \\delta-GLMB density represents a\nmulti-Bernoulli mixture with labelled targets so it can be seen as a special\ncase of PMBM. In addition, we propose an implementation for linear/Gaussian\ndynamic and measurement models and how to efficiently obtain typical estimators\nin the literature from the PMBM. The PMBM filter is shown to outperform other\nfilters in the literature in a challenging scenario.\n",
"title": "Poisson multi-Bernoulli mixture filter: direct derivation and implementation"
}
| null | null | null | null | true | null |
3878
| null |
Default
| null | null |
null |
{
"abstract": " Surface properties are examined in a chiral d-wave superconductor with\nhexagonal symmetry, whose one-body Hamiltonian possesses the intrinsic\nspin-orbit coupling identical to the one characterizing the topological nature\nof the Kane-Mele honeycomb insulator. In the normal state spin-orbit coupling\ngives rise to spontaneous surface spin currents, whereas in the superconducting\nstate there exist besides the spin currents also charge surface currents, due\nto the chiral pairing symmetry. Interestingly, the combination of these two\ncurrents results in a surface spin polarization, whose spatial dependence is\nmarkedly different on the zigzag and armchair surfaces. We discuss various\npotential candidate materials, such as SrPtAs, which may exhibit these surface\nproperties.\n",
"title": "Surface magnetism in a chiral d-wave superconductor with hexagonal symmetry"
}
| null | null | null | null | true | null |
3879
| null |
Default
| null | null |
null |
{
"abstract": " Modern threats have emerged from the prevalence of social networks. Hostile\nactors, such as extremist groups or foreign governments, utilize these networks\nto run propaganda campaigns with different aims. For extremists, these\ncampaigns are designed for recruiting new members or inciting violence. For\nforeign governments, the aim may be to create instability in rival nations.\nProper social network counter-measures are needed to combat these threats. Here\nwe present one important counter-measure: penetrating social networks. This\nmeans making target users connect with or follow agents deployed in the social\nnetwork. Once such connections are established with the targets, the agents can\ninfluence them by sharing content which counters the influence campaign. In\nthis work we study how to penetrate a social network, which we call the\nfollow-back problem. The goal here is to find a policy that maximizes the\nnumber of targets that follow the agent.\nWe conduct an empirical study to understand what behavioral and network\nfeatures affect the probability of a target following an agent. We find that\nthe degree of the target and the size of the mutual neighborhood of the agent\nand target in the network affect this probability. Based on our empirical\nfindings, we then propose a model for targets following an agent. Using this\nmodel, we solve the follow-back problem exactly on directed acyclic graphs and\nderive a closed form expression for the expected number of follows an agent\nreceives under the optimal policy. We then formulate the follow-back problem on\nan arbitrary graph as an integer program. To evaluate our integer program based\npolicies, we conduct simulations on real social network topologies in Twitter.\nWe find that our polices result in more effective network penetration, with\nsignificant increases in the expected number of targets that follow the agent.\n",
"title": "Penetrating a Social Network: The Follow-back Problem"
}
| null | null | null | null | true | null |
3880
| null |
Default
| null | null |
null |
{
"abstract": " The contribution of $O^{2-}$ ions to antiferromagnetism in\n$La_{2-x}Ae_xCuO_4$ ($Ae = Sr, Ba)$ is highly sensitive to doped holes. In\ncontrast, the contribution of $Cu^{2+}$ ions to antiferromagnetism in\n$Nd_{2-x}Ce_xCuO_{4+y}$ is much less sensitive to doped electrons. The\ndifference causes the precarious and, respectively, robust antiferromagnetic\nphase of these cuprates. The same sensitivities affect the doping dependence of\nthe incommensurability of density waves, $\\delta (x)$. In the hole-doped\ncompounds this gives rise to a doping offset for magnetic and charge density\nwaves, $\\delta_{m,c}^p(x) \\propto \\sqrt{x-x_{0p}^N}$. Here $x_{0p}^N$ is the\ndoping concentration where the Néel temperature vanishes, $T_N(x_{0p}^N) =\n0$. No such doping offset occurs for density waves in the electron-doped\ncompound. Instead, excess oxygen (necessary for stability in crystal growth) of\nconcentration $y$ causes a different doping offset in the latter case,\n$\\delta_{m,c}^n(x) \\propto \\sqrt{x- 2y}$. The square-root formulas result from\nthe assumption of superlattice formation through partitioning of the $CuO_2$\nplane by pairs of itinerant charge carriers. Agreement of observed\nincommensurability $\\delta(x)$ with the formulas is very good for the\nhole-doped compounds and reasonable for the electron-doped compound. The\ndeviation in the latter case may be caused by residual excess oxygen.\n",
"title": "Universality of density waves in p-doped La2CuO4 and n-doped Nd2CuO4+y"
}
| null | null | null | null | true | null |
3881
| null |
Default
| null | null |
null |
{
"abstract": " Networks describe a range of social, biological and technical phenomena. An\nimportant property of a network is its degree correlation or assortativity,\ndescribing how nodes in the network associate based on their number of\nconnections. Social networks are typically thought to be distinct from other\nnetworks in being assortative (possessing positive degree correlations);\nwell-connected individuals associate with other well-connected individuals, and\npoorly-connected individuals associate with each other. We review the evidence\nfor this in the literature and find that, while social networks are more\nassortative than non-social networks, only when they are built using\ngroup-based methods do they tend to be positively assortative. Non-social\nnetworks tend to be disassortative. We go on to show that connecting\nindividuals due to shared membership of a group, a commonly used method, biases\ntowards assortativity unless a large enough number of censuses of the network\nare taken. We present a number of solutions to overcoming this bias by drawing\non advances in sociological and biological fields. Adoption of these methods\nacross all fields can greatly enhance our understanding of social networks and\nnetworks in general.\n",
"title": "The perceived assortativity of social networks: Methodological problems and solutions"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
3882
| null |
Validated
| null | null |
null |
{
"abstract": " We used a multiple-scale homogenization method to derive generalized sheet\ntransition conditions (GSTCs) for electromagnetic fields at the surface of a\nmetascreen---a metasurface with a \"fishnet\" structure. These surfaces are\ncharacterized by periodically-spaced arbitrary-shaped apertures in an otherwise\nrelatively impenetrable surface. The parameters in these GSTCs are interpreted\nas effective surface susceptibilities and surface porosities, which are related\nto the geometry of the apertures that constitute the metascreen. Finally, we\nemphasize the subtle but important difference between the GSTCs required for\nmetascreens and those required for metafilms (a metasurface with a \"cermet\"\nstructure, i.e., an array of isolated (non-touching) scatterers).\n",
"title": "Generalized Sheet Transition Conditions (GSTCs) for a Metascreen -- A Fishnet Metasurface"
}
| null | null | null | null | true | null |
3883
| null |
Default
| null | null |
null |
{
"abstract": " Mobile network operators can track subscribers via passive or active\nmonitoring of device locations. The recorded trajectories offer an\nunprecedented outlook on the activities of large user populations, which\nenables developing new networking solutions and services, and scaling up\nstudies across research disciplines. Yet, the disclosure of individual\ntrajectories raises significant privacy concerns: thus, these data are often\nprotected by restrictive non-disclosure agreements that limit their\navailability and impede potential usages. In this paper, we contribute to the\ndevelopment of technical solutions to the problem of privacy-preserving\npublishing of spatiotemporal trajectories of mobile subscribers. We propose an\nalgorithm that generalizes the data so that they satisfy\n$k^{\\tau,\\epsilon}$-anonymity, an original privacy criterion that thwarts\nattacks on trajectories. Evaluations with real-world datasets demonstrate that\nour algorithm attains its objective while retaining a substantial level of\naccuracy in the data. Our work is a step forward in the direction of open,\nprivacy-preserving datasets of spatiotemporal trajectories.\n",
"title": "$k^{τ,ε}$-anonymity: Towards Privacy-Preserving Publishing of Spatiotemporal Trajectory Data"
}
| null | null | null | null | true | null |
3884
| null |
Default
| null | null |
null |
{
"abstract": " Taipan is a multi-object spectroscopic galaxy survey starting in 2017 that\nwill cover 2pi steradians over the southern sky, and obtain optical spectra for\nabout two million galaxies out to z<0.4. Taipan will use the newly-refurbished\n1.2m UK Schmidt Telescope at Siding Spring Observatory with the new TAIPAN\ninstrument, which includes an innovative 'Starbugs' positioning system capable\nof rapidly and simultaneously deploying up to 150 spectroscopic fibres (and up\nto 300 with a proposed upgrade) over the 6-deg diameter focal plane, and a\npurpose-built spectrograph operating from 370 to 870nm with resolving power\nR>2000. The main scientific goals of Taipan are: (i) to measure the distance\nscale of the Universe (primarily governed by the local expansion rate, H_0) to\n1% precision, and the structure growth rate of structure to 5%; (ii) to make\nthe most extensive map yet constructed of the mass distribution and motions in\nthe local Universe, using peculiar velocities based on improved Fundamental\nPlane distances, which will enable sensitive tests of gravitational physics;\nand (iii) to deliver a legacy sample of low-redshift galaxies as a unique\nlaboratory for studying galaxy evolution as a function of mass and environment.\nThe final survey, which will be completed within 5 years, will consist of a\ncomplete magnitude-limited sample (i<17) of about 1.2x10^6 galaxies,\nsupplemented by an extension to higher redshifts and fainter magnitudes\n(i<18.1) of a luminous red galaxy sample of about 0.8x10^6 galaxies.\nObservations and data processing will be carried out remotely and in a\nfully-automated way, using a purpose-built automated 'virtual observer'\nsoftware and an automated data reduction pipeline. The Taipan survey is\ndeliberately designed to maximise its legacy value, by complementing and\nenhancing current and planned surveys of the southern sky at wavelengths from\nthe optical to the radio.\n",
"title": "The Taipan Galaxy Survey: Scientific Goals and Observing Strategy"
}
| null | null | null | null | true | null |
3885
| null |
Default
| null | null |
null |
{
"abstract": " Online learning algorithms, widely used to power search and content\noptimization on the web, must balance exploration and exploitation, potentially\nsacrificing the experience of current users for information that will lead to\nbetter decisions in the future. Recently, concerns have been raised about\nwhether the process of exploration could be viewed as unfair, placing too much\nburden on certain individuals or groups. Motivated by these concerns, we\ninitiate the study of the externalities of exploration - the undesirable side\neffects that the presence of one party may impose on another - under the linear\ncontextual bandits model. We introduce the notion of a group externality,\nmeasuring the extent to which the presence of one population of users impacts\nthe rewards of another. We show that this impact can in some cases be negative,\nand that, in a certain sense, no algorithm can avoid it. We then study\nexternalities at the individual level, interpreting the act of exploration as\nan externality imposed on the current user of a system by future users. This\ndrives us to ask under what conditions inherent diversity in the data makes\nexplicit exploration unnecessary. We build on a recent line of work on the\nsmoothed analysis of the greedy algorithm that always chooses the action that\ncurrently looks optimal, improving on prior results to show that a greedy\napproach almost matches the best possible Bayesian regret rate of any other\nalgorithm on the same problem instance whenever the diversity conditions hold,\nand that this regret is at most $\\tilde{O}(T^{1/3})$. Returning to group-level\neffects, we show that under the same conditions, negative group externalities\nessentially vanish under the greedy algorithm. Together, our results uncover a\nsharp contrast between the high externalities that exist in the worst case, and\nthe ability to remove all externalities if the data is sufficiently diverse.\n",
"title": "The Externalities of Exploration and How Data Diversity Helps Exploitation"
}
| null | null |
[
"Statistics"
] | null | true | null |
3886
| null |
Validated
| null | null |
null |
{
"abstract": " This paper is concerned with the blowup phenomena for initial value problem\nof semilinear wave equation with critical space-dependent damping term\n(DW:$V$). The main result of the present paper is to give a solution of the\nproblem and to provide a sharp estimate for lifespan for such a solution when\n$\\frac{N}{N-1}<p\\leq p_S(N+V_0)$, where $p_S(N)$ is the Strauss exponent for\n(DW:$0$). The main idea of the proof is due to the technique of test functions\nfor (DW:$0$) originated by Zhou--Han (2014, MR3169791). Moreover, we find a new\nthreshold value $V_0=\\frac{(N-1)^2}{N+1}$ for the coefficient of critical and\nsingular damping $|x|^{-1}$.\n",
"title": "Life-span of blowup solutions to semilinear wave equation with space-dependent critical damping"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3887
| null |
Validated
| null | null |
null |
{
"abstract": " The object of the present paper is to study invariant submanifolds of\n(LCS)n-manifolds with respect to quarter symmetric metric connection. It is\nshown that the mean curvature of an invariant submanifold of (LCS)n-manifold\nwith respect to quarter symmetric metric connection and Levi-Civita connection\nare equal. An example is constructed to illustrate the results of the paper. We\nalso obtain some equivalent conditions of such notion.\n",
"title": "Invariant submanifolds of (LCS)n-Manifolds with respect to quarter symmetric metric connection"
}
| null | null | null | null | true | null |
3888
| null |
Default
| null | null |
null |
{
"abstract": " Dynamic epidemic models have proven valuable for public health decision\nmakers as they provide useful insights into the understanding and prevention of\ninfectious diseases. However, inference for these types of models can be\ndifficult because the disease spread is typically only partially observed e.g.\nin form of reported incidences in given time periods. This chapter discusses\nhow to perform likelihood-based inference for partially observed Markov\nepidemic models when it is relatively easy to generate samples from the Markov\ntransmission model while the likelihood function is intractable. The first part\nof the chapter reviews the theoretical background of inference for partially\nobserved Markov processes (POMP) via iterated filtering. In the second part of\nthe chapter the performance of the method and associated practical difficulties\nare illustrated on two examples. In the first example a simulated outbreak data\nset consisting of the number of newly reported cases aggregated by week is\nfitted to a POMP where the underlying disease transmission model is assumed to\nbe a simple Markovian SIR model. The second example illustrates possible model\nextensions such as seasonal forcing and over-dispersion in both, the\ntransmission and observation model, which can be used, e.g., when analysing\nroutinely collected rotavirus surveillance data. Both examples are implemented\nusing the R-package pomp (King et al., 2016) and the code is made available\nonline.\n",
"title": "Iterated filtering methods for Markov process epidemic models"
}
| null | null | null | null | true | null |
3889
| null |
Default
| null | null |
null |
{
"abstract": " This paper considers the problem of designing maximum distance separable\n(MDS) codes over small fields with constraints on the support of their\ngenerator matrices. For any given $m\\times n$ binary matrix $M$, the GM-MDS\nconjecture, due to Dau et al., states that if $M$ satisfies the so-called MDS\ncondition, then for any field $\\mathbb{F}$ of size $q\\geq n+m-1$, there exists\nan $[n,m]_q$ MDS code whose generator matrix $G$, with entries in $\\mathbb{F}$,\nfits $M$ (i.e., $M$ is the support matrix of $G$). Despite all the attempts by\nthe coding theory community, this conjecture remains still open in general. It\nwas shown, independently by Yan et al. and Dau et al., that the GM-MDS\nconjecture holds if the following conjecture, referred to as the TM-MDS\nconjecture, holds: if $M$ satisfies the MDS condition, then the determinant of\na transformation matrix $T$, such that $TV$ fits $M$, is not identically zero,\nwhere $V$ is a Vandermonde matrix with distinct parameters. In this work, we\ngeneralize the TM-MDS conjecture, and present an algebraic-combinatorial\napproach based on polynomial-degree reduction for proving this conjecture. Our\nproof technique's strength is based primarily on reducing inherent\ncombinatorics in the proof. We demonstrate the strength of our technique by\nproving the TM-MDS conjecture for the cases where the number of rows ($m$) of\n$M$ is upper bounded by $5$. For this class of special cases of $M$ where the\nonly additional constraint is on $m$, only cases with $m\\leq 4$ were previously\nproven theoretically, and the previously used proof techniques are not\napplicable to cases with $m > 4$.\n",
"title": "An Algebraic-Combinatorial Proof Technique for the GM-MDS Conjecture"
}
| null | null | null | null | true | null |
3890
| null |
Default
| null | null |
null |
{
"abstract": " The \\emph{word problem} of a group $G = \\langle \\Sigma \\rangle$ can be\ndefined as the set of formal words in $\\Sigma^*$ that represent the identity in\n$G$. When viewed as formal languages, this gives a strong connection between\nclasses of groups and classes of formal languages. For example, Anisimov showed\nthat a group is finite if and only if its word problem is a regular language,\nand Muller and Schupp showed that a group is virtually-free if and only if its\nword problem is a context-free language. Above this, not much was known, until\nSalvati showed recently that the word problem of $\\mathbb{Z}^2$ is a multiple\ncontext-free language, giving first such example. We generalize Salvati's\nresult to show that the word problem of $\\mathbb{Z}^n$ is a multiple\ncontext-free language for any $n$.\n",
"title": "The Word Problem of $\\mathbb{Z}^n$ Is a Multiple Context-Free Language"
}
| null | null | null | null | true | null |
3891
| null |
Default
| null | null |
null |
{
"abstract": " Models in econophysics, i.e., the emerging field of statistical physics that\napplies the main concepts of traditional physics to economics, typically\nconsist of large systems of economic agents who are characterized by the amount\nof money they have. In the simplest model, at each time step, one agent gives\none dollar to another agent, with both agents being chosen independently and\nuniformly at random from the system. Numerical simulations of this model\nsuggest that, at least when the number of agents and the average amount of\nmoney per agent are large, the distribution of money converges to an\nexponential distribution reminiscent of the Boltzmann-Gibbs distribution of\nenergy in physics. The main objective of this paper is to give a rigorous proof\nof this result and show that the convergence to the exponential distribution is\nuniversal in the sense that it holds more generally when the economic agents\nare located on the vertices of a connected graph and interact locally with\ntheir neighbors rather than globally with all the other agents. We also study a\nclosely related model where, at each time step, agents buy with a probability\nproportional to the amount of money they have, and prove that in this case the\nlimiting distribution of money is Poissonian.\n",
"title": "Rigorous proof of the Boltzmann-Gibbs distribution of money on connected graphs"
}
| null | null | null | null | true | null |
3892
| null |
Default
| null | null |
null |
{
"abstract": " Spatially extended population dynamics models that incorporate intrinsic\nnoise serve as case studies for the role of fluctuations and correlations in\nbiological systems. Including spatial structure and stochastic noise in\npredator-prey competition invalidates the deterministic Lotka-Volterra picture\nof neutral population cycles. Stochastic models yield long-lived erratic\npopulation oscillations stemming from a resonant amplification mechanism. In\nspatially extended predator-prey systems, one observes noise-stabilized\nactivity and persistent correlations. Fluctuation-induced renormalizations of\nthe oscillation parameters can be analyzed perturbatively. The critical\ndynamics and the non-equilibrium relaxation kinetics at the predator extinction\nthreshold are characterized by the directed percolation universality class.\nSpatial or environmental variability results in more localized patches which\nenhances both species densities. Affixing variable rates to individual\nparticles and allowing for trait inheritance subject to mutations induces fast\nevolutionary dynamics for the rate distributions. Stochastic spatial variants\nof cyclic competition with rock-paper-scissors interactions illustrate\nconnections between population dynamics and evolutionary game theory, and\ndemonstrate how space can help maintain diversity. In two dimensions,\nthree-species cyclic competition models of the May-Leonard type are\ncharacterized by the emergence of spiral patterns whose properties are\nelucidated by a mapping onto a complex Ginzburg-Landau equation. Extensions to\ngeneral food networks can be classified on the mean-field level, which provides\nboth a fundamental understanding of ensuing cooperativity and emergence of\nalliances. Novel space-time patterns emerge as a result of the formation of\ncompeting alliances, such as coarsening domains that each incorporate\nrock-paper-scissors competition games.\n",
"title": "Stochastic population dynamics in spatially extended predator-prey systems"
}
| null | null | null | null | true | null |
3893
| null |
Default
| null | null |
null |
{
"abstract": " We consider the problem of learning a policy for a Markov decision process\nconsistent with data captured on the state-actions pairs followed by the\npolicy. We assume that the policy belongs to a class of parameterized policies\nwhich are defined using features associated with the state-action pairs. The\nfeatures are known a priori, however, only an unknown subset of them could be\nrelevant. The policy parameters that correspond to an observed target policy\nare recovered using $\\ell_1$-regularized logistic regression that best fits the\nobserved state-action samples. We establish bounds on the difference between\nthe average reward of the estimated and the original policy (regret) in terms\nof the generalization error and the ergodic coefficient of the underlying\nMarkov chain. To that end, we combine sample complexity theory and sensitivity\nanalysis of the stationary distribution of Markov chains. Our analysis suggests\nthat to achieve regret within order $O(\\sqrt{\\epsilon})$, it suffices to use\ntraining sample size on the order of $\\Omega(\\log n \\cdot poly(1/\\epsilon))$,\nwhere $n$ is the number of the features. We demonstrate the effectiveness of\nour method on a synthetic robot navigation example.\n",
"title": "Learning Policies for Markov Decision Processes from Data"
}
| null | null |
[
"Computer Science",
"Mathematics",
"Statistics"
] | null | true | null |
3894
| null |
Validated
| null | null |
null |
{
"abstract": " As autonomous vehicles become an every-day reality, high-accuracy pedestrian\ndetection is of paramount practical importance. Pedestrian detection is a\nhighly researched topic with mature methods, but most datasets focus on common\nscenes of people engaged in typical walking poses on sidewalks. But performance\nis most crucial for dangerous scenarios, such as children playing in the street\nor people using bicycles/skateboards in unexpected ways. Such \"in-the-tail\"\ndata is notoriously hard to observe, making both training and testing\ndifficult. To analyze this problem, we have collected a novel annotated dataset\nof dangerous scenarios called the Precarious Pedestrian dataset. Even given a\ndedicated collection effort, it is relatively small by contemporary standards\n(around 1000 images). To allow for large-scale data-driven learning, we explore\nthe use of synthetic data generated by a game engine. A significant challenge\nis selected the right \"priors\" or parameters for synthesis: we would like\nrealistic data with poses and object configurations that mimic true Precarious\nPedestrians. Inspired by Generative Adversarial Networks (GANs), we generate a\nmassive amount of synthetic data and train a discriminative classifier to\nselect a realistic subset, which we deem the Adversarial Imposters. We\ndemonstrate that this simple pipeline allows one to synthesize realistic\ntraining data by making use of rendering/animation engines within a GAN\nframework. Interestingly, we also demonstrate that such data can be used to\nrank algorithms, suggesting that Adversarial Imposters can also be used for\n\"in-the-tail\" validation at test-time, a notoriously difficult challenge for\nreal-world deployment.\n",
"title": "Expecting the Unexpected: Training Detectors for Unusual Pedestrians with Adversarial Imposters"
}
| null | null | null | null | true | null |
3895
| null |
Default
| null | null |
null |
{
"abstract": " A finite word is closed if it contains a factor that occurs both as a prefix\nand as a suffix but does not have internal occurrences, otherwise it is open.\nWe are interested in the {\\it oc-sequence} of a word, which is the binary\nsequence whose $n$-th element is $0$ if the prefix of length $n$ of the word is\nopen, or $1$ if it is closed. We exhibit results showing that this sequence is\ndeeply related to the combinatorial and periodic structure of a word. In the\ncase of Sturmian words, we show that these are uniquely determined (up to\nrenaming letters) by their oc-sequence. Moreover, we prove that the class of\nfinite Sturmian words is a maximal element with this property in the class of\nbinary factorial languages. We then discuss several aspects of Sturmian words\nthat can be expressed through this sequence. Finally, we provide a linear-time\nalgorithm that computes the oc-sequence of a finite word, and a linear-time\nalgorithm that reconstructs a finite Sturmian word from its oc-sequence.\n",
"title": "The sequence of open and closed prefixes of a Sturmian word"
}
| null | null | null | null | true | null |
3896
| null |
Default
| null | null |
null |
{
"abstract": " We construct a continuous time model for price-mediated contagion\nprecipitated by a common exogenous stress to the trading book of all firms in\nthe financial system. In this setting, firms are constrained so as to satisfy a\nrisk-weight based capital ratio requirement. We use this model to find\nanalytical bounds on the risk-weights for an asset as a function of the market\nliquidity. Under these appropriate risk-weights, we find existence and\nuniqueness for the joint system of firm behavior and the asset price. We\nfurther consider an analytical bound on the firm liquidations, which allows us\nto construct exact formulas for stress testing the financial system with\ndeterministic or random stresses. Numerical case studies are provided to\ndemonstrate various implications of this model and analytical bounds.\n",
"title": "Capital Regulation under Price Impacts and Dynamic Financial Contagion"
}
| null | null | null | null | true | null |
3897
| null |
Default
| null | null |
null |
{
"abstract": " In 2000, Dergachev and Kirillov introduced subalgebras of \"seaweed type\" in\n$\\mathfrak{gl}_n$ and computed their index using certain graphs, which we call\ntype-${\\sf A}$ meander graphs. Then the subalgebras of seaweed type, or just\n\"seaweeds\", have been defined by Panyushev (2001) for arbitrary reductive Lie\nalgebras. Recently, a meander graph approach to computing the index in types\n${\\sf B}$ and ${\\sf C}$ has been developed by the authors. In this article, we\nconsider the most difficult and interesting case of type ${\\sf D}$. Some new\nphenomena occurring here are related to the fact that the Dynkin diagram has a\nbranching node.\n",
"title": "On seaweed subalgebras and meander graphs in type D"
}
| null | null | null | null | true | null |
3898
| null |
Default
| null | null |
null |
{
"abstract": " Let $M$ be an even-dimensional, oriented closed manifold. We show that the\nrestriction of a singular Riemannian flow on $M$ to a small tubular\nneighborhood of each connected component of its singular stratum is\nfoliated-diffeomorphic to an isometric flow on the same neighborhood. We then\nprove a formula that computes characteristic numbers of $M$ as the sum of\nresidues associated to the infinitesimal foliation at the components of the\nsingular stratum of the flow.\n",
"title": "Singular Riemannian flows and characteristic numbers"
}
| null | null | null | null | true | null |
3899
| null |
Default
| null | null |
null |
{
"abstract": " Transformation models are a very important tool for applied statisticians and\neconometricians. In many applications, the dependent variable is transformed so\nthat homogeneity or normal distribution of the error holds. In this paper, we\nanalyze transformation models in a high-dimensional setting, where the set of\npotential covariates is large. We propose an estimator for the transformation\nparameter and we show that it is asymptotically normally distributed using an\northogonalized moment condition where the nuisance functions depend on the\ntarget parameter. In a simulation study, we show that the proposed estimator\nworks well in small samples. A common practice in labor economics is to\ntransform wage with the log-function. In this study, we test if this\ntransformation holds in CPS data from the United States.\n",
"title": "Transformation Models in High-Dimensions"
}
| null | null | null | null | true | null |
3900
| null |
Default
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.