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": " On the probability simplex, we can consider the standard information\ngeometric structure with the e- and m-affine connections mutually dual with\nrespect to the Fisher metric. The geometry naturally defines submanifolds\nsimultaneously autoparallel for the both affine connections, which we call {\\em\ndoubly autoparallel submanifolds}.\nIn this note we discuss their several interesting common properties. Further,\nwe algebraically characterize doubly autoparallel submanifolds on the\nprobability simplex and give their classification.\n",
"title": "Doubly autoparallel structure on the probability simplex"
}
| null | null | null | null | true | null |
4001
| null |
Default
| null | null |
null |
{
"abstract": " The search of unconventional magnetic and non-magnetic states is a major\ntopic in the study of frustrated magnetism. Canonical examples of those states\ninclude various spin liquids and spin nematics. However, discerning their\nexistence and the correct characterization is usually challenging. Here we\nintroduce a machine-learning protocol that can identify general nematic order\nand their order parameter from seemingly featureless spin configurations, thus\nproviding comprehensive insight on the presence or absence of hidden orders. We\ndemonstrate the capabilities of our method by extracting the analytical form of\nnematic order parameter tensors up to rank 6. This may prove useful in the\nsearch for novel spin states and for ruling out spurious spin liquid\ncandidates.\n",
"title": "Probing Hidden Spin Order with Interpretable Machine Learning"
}
| null | null | null | null | true | null |
4002
| null |
Default
| null | null |
null |
{
"abstract": " A wide variety of phenomena of engineering and scientific interest are of a\ncontinuous-time nature and can be modeled by stochastic differential equations\n(SDEs), which represent the evolution of the uncertainty in the states of a\nsystem. For systems of this class, some parameters of the SDE might be unknown\nand the measured data often includes noise, so state and parameter estimators\nare needed to perform inference and further analysis using the system state\npath. The distributions of SDEs which are nonlinear or subject to non-Gaussian\nmeasurement noise do not admit tractable analytic expressions, so state and\nparameter estimators for these systems are often approximations based on\nheuristics, such as the extended and unscented Kalman smoothers, or the\nprediction error method using nonlinear Kalman filters. However, the Onsager\nMachlup functional can be used to obtain fictitious densities for the\nparameters and state-paths of SDEs with analytic expressions. In this thesis,\nwe provide a unified theoretical framework for maximum a posteriori (MAP)\nestimation of general random variables, possibly infinite-dimensional, and show\nhow the Onsager--Machlup functional can be used to construct the joint MAP\nstate-path and parameter estimator for SDEs. We also prove that the minimum\nenergy estimator, which is often thought to be the MAP state-path estimator,\nactually gives the state paths associated to the MAP noise paths. Furthermore,\nwe prove that the discretized MAP state-path and parameter estimators, which\nhave emerged recently as powerful alternatives to nonlinear Kalman smoothers,\nconverge hypographically as the discretization step vanishes. Their\nhypographical limit, however, is the MAP estimator for SDEs when the\ntrapezoidal discretization is used and the minimum energy estimator when the\nEuler discretization is used, associating different interpretations to each\ndiscretized estimate.\n",
"title": "Maximum a Posteriori Joint State Path and Parameter Estimation in Stochastic Differential Equations"
}
| null | null |
[
"Mathematics",
"Statistics"
] | null | true | null |
4003
| null |
Validated
| null | null |
null |
{
"abstract": " The endogenous adaptation of agents, that may adjust their local contact\nnetwork in response to the risk of being infected, can have the perverse effect\nof increasing the overall systemic infectiveness of a disease. We study a\ndynamical model over two geographically distinct but interacting locations, to\nbetter understand theoretically the mechanism at play. Moreover, we provide\nempirical motivation from the Italian National Bovine Database, for the period\n2006-2013.\n",
"title": "Spreading of an infectious disease between different locations"
}
| null | null | null | null | true | null |
4004
| null |
Default
| null | null |
null |
{
"abstract": " Although the existence of quasi-bound rotational levels of the $X^+ \\\n^2\\Sigma_g^+$ ground state of H$_2^+$ has been predicted a long time ago, these\nstates have never been observed. Calculated positions and widths of quasi-bound\nrotational levels located close to the top of the centrifugal barriers have not\nbeen reported either. Given the role that such states play in the recombination\nof H(1s) and H$^+$ to form H$_2^+$, this lack of data may be regarded as one of\nthe largest unknown aspects of this otherwise accurately known fundamental\nmolecular cation. We present measurements of the positions and widths of the\nlowest-lying quasi-bound rotational levels of H$_2^+$ and compare the\nexperimental results with the positions and widths we calculate using a\npotential model for the $X^+$ state of H$_2^+$ which includes adiabatic,\nnonadiabatic, relativistic and radiative corrections to the Born-Oppenheimer\napproximation.\n",
"title": "Observation and calculation of the quasi-bound rovibrational levels of the electronic ground state of H$_2^+$"
}
| null | null | null | null | true | null |
4005
| null |
Default
| null | null |
null |
{
"abstract": " The dark ages of the Universe end with the formation of the first generation\nof stars residing in primeval galaxies. These objects were the first to produce\nultraviolet ionizing photons in a period when the cosmic gas changed from a\nneutral state to an ionized one, known as Epoch of Reionization (EoR). A\npivotal aspect to comprehend the EoR is to probe the intertwined relationship\nbetween the fraction of ionizing photons capable to escape dark haloes, also\nknown as the escape fraction ($f_{esc}$), and the physical properties of the\ngalaxy. This work develops a sound statistical model suitable to account for\nsuch non-linear relationships and the non-Gaussian nature of $f_{esc}$. This\nmodel simultaneously estimates the probability that a given primordial galaxy\nstarts the ionizing photon production and estimates the mean level of the\n$f_{esc}$ once it is triggered. The model was employed in the First Billion\nYears simulation suite, from which we show that the baryonic fraction and the\nrate of ionizing photons appear to have a larger impact on $f_{esc}$ than\npreviously thought. A naive univariate analysis of the same problem would\nsuggest smaller effects for these properties and a much larger impact for the\nspecific star formation rate, which is lessened after accounting for other\ngalaxy properties and non-linearities in the statistical model.\n",
"title": "A case study of hurdle and generalized additive models in astronomy: the escape of ionizing radiation"
}
| null | null | null | null | true | null |
4006
| null |
Default
| null | null |
null |
{
"abstract": " In this project, we propose a novel approach for estimating depth from RGB\nimages. Traditionally, most work uses a single RGB image to estimate depth,\nwhich is inherently difficult and generally results in poor performance, even\nwith thousands of data examples. In this work, we alternatively use multiple\nRGB images that were captured while changing the focus of the camera's lens.\nThis method leverages the natural depth information correlated to the different\npatterns of clarity/blur in the sequence of focal images, which helps\ndistinguish objects at different depths. Since no such data set exists for\nlearning this mapping, we collect our own data set using customized hardware.\nWe then use a convolutional neural network for learning the depth from the\nstacked focal images. Comparative studies were conducted on both a standard\nRGBD data set and our own data set (learning from both single and multiple\nimages), and results verified that stacked focal images yield better depth\nestimation than using just single RGB image.\n",
"title": "Out-of-focus: Learning Depth from Image Bokeh for Robotic Perception"
}
| null | null |
[
"Computer Science"
] | null | true | null |
4007
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper, we consider the problem of machine teaching, the inverse\nproblem of machine learning. Different from traditional machine teaching which\nviews the learners as batch algorithms, we study a new paradigm where the\nlearner uses an iterative algorithm and a teacher can feed examples\nsequentially and intelligently based on the current performance of the learner.\nWe show that the teaching complexity in the iterative case is very different\nfrom that in the batch case. Instead of constructing a minimal training set for\nlearners, our iterative machine teaching focuses on achieving fast convergence\nin the learner model. Depending on the level of information the teacher has\nfrom the learner model, we design teaching algorithms which can provably reduce\nthe number of teaching examples and achieve faster convergence than learning\nwithout teachers. We also validate our theoretical findings with extensive\nexperiments on different data distribution and real image datasets.\n",
"title": "Iterative Machine Teaching"
}
| null | null | null | null | true | null |
4008
| null |
Default
| null | null |
null |
{
"abstract": " Directed latent variable models that formulate the joint distribution as\n$p(x,z) = p(z) p(x \\mid z)$ have the advantage of fast and exact sampling.\nHowever, these models have the weakness of needing to specify $p(z)$, often\nwith a simple fixed prior that limits the expressiveness of the model.\nUndirected latent variable models discard the requirement that $p(z)$ be\nspecified with a prior, yet sampling from them generally requires an iterative\nprocedure such as blocked Gibbs-sampling that may require many steps to draw\nsamples from the joint distribution $p(x, z)$. We propose a novel approach to\nlearning the joint distribution between the data and a latent code which uses\nan adversarially learned iterative procedure to gradually refine the joint\ndistribution, $p(x, z)$, to better match with the data distribution on each\nstep. GibbsNet is the best of both worlds both in theory and in practice.\nAchieving the speed and simplicity of a directed latent variable model, it is\nguaranteed (assuming the adversarial game reaches the virtual training criteria\nglobal minimum) to produce samples from $p(x, z)$ with only a few sampling\niterations. Achieving the expressiveness and flexibility of an undirected\nlatent variable model, GibbsNet does away with the need for an explicit $p(z)$\nand has the ability to do attribute prediction, class-conditional generation,\nand joint image-attribute modeling in a single model which is not trained for\nany of these specific tasks. We show empirically that GibbsNet is able to learn\na more complex $p(z)$ and show that this leads to improved inpainting and\niterative refinement of $p(x, z)$ for dozens of steps and stable generation\nwithout collapse for thousands of steps, despite being trained on only a few\nsteps.\n",
"title": "GibbsNet: Iterative Adversarial Inference for Deep Graphical Models"
}
| null | null | null | null | true | null |
4009
| null |
Default
| null | null |
null |
{
"abstract": " For a graph $G$, let $odd(G)$ and $\\omega(G)$ denote the number of odd\ncomponents and the number of components of $G$, respectively. Then it is\nwell-known that $G$ has a 1-factor if and only if $odd(G-S)\\le |S|$ for all\n$S\\subset V(G)$. Also it is clear that $odd(G-S) \\le \\omega(G-S)$. In this\npaper we characterize a 1-tough graph $G$, which satisfies $\\omega(G-S) \\le\n|S|$ for all $\\emptyset \\ne S \\subset V(G)$, using an $H$-factor of a\nset-valued function $H:V(G) \\to \\{ \\{1\\}, \\{0,2\\} \\}$. Moreover, we generalize\nthis characterization to a graph that satisfies $\\omega(G-S) \\le f(S)$ for all\n$\\emptyset \\ne S \\subset V(G)$, where $f:V(G) \\to \\{1,3,5, \\ldots\\}$.\n",
"title": "Characterization of 1-Tough Graphs using Factors"
}
| null | null |
[
"Mathematics"
] | null | true | null |
4010
| null |
Validated
| null | null |
null |
{
"abstract": " Gradient boosting is a state-of-the-art prediction technique that\nsequentially produces a model in the form of linear combinations of simple\npredictors---typically decision trees---by solving an infinite-dimensional\nconvex optimization problem. We provide in the present paper a thorough\nanalysis of two widespread versions of gradient boosting, and introduce a\ngeneral framework for studying these algorithms from the point of view of\nfunctional optimization. We prove their convergence as the number of iterations\ntends to infinity and highlight the importance of having a strongly convex risk\nfunctional to minimize. We also present a reasonable statistical context\nensuring consistency properties of the boosting predictors as the sample size\ngrows. In our approach, the optimization procedures are run forever (that is,\nwithout resorting to an early stopping strategy), and statistical\nregularization is basically achieved via an appropriate $L^2$ penalization of\nthe loss and strong convexity arguments.\n",
"title": "Optimization by gradient boosting"
}
| null | null | null | null | true | null |
4011
| null |
Default
| null | null |
null |
{
"abstract": " A decentralized payment system is not secure if transactions are transferred\ndirectly between clients. In such a situation it is not possible to prevent a\nclient from redeeming some coins twice in separate transactions that means a\ndouble-spending attack. Bitcoin uses a simple method to preventing this attack\ni.e. all transactions are published in a unique log (blockchain). This approach\nrequires a global consensus on the blockchain that because of significant\nlatency for transaction confirmation is vulnerable against double-spending. The\nsolution is to accelerate confirmations. In this paper, we try to introduce an\nalternative for PoW because of all its major and significant problems that lead\nto collapsing decentralization of the Bitcoin, while a full decentralized\npayment system is the main goal of Bitcoin idea. As the network is growing and\nbecoming larger day-today , Bitcoin is approaching this risk. The method we\nintroduce is based on a distributed voting process: RDV: Register, Deposit,\nVote.\n",
"title": "RDV: Register, Deposit, Vote: Secure and Decentralized Consensus Mechanism for Blockchain Networks"
}
| null | null | null | null | true | null |
4012
| null |
Default
| null | null |
null |
{
"abstract": " Let $M$ and $N$ be two monomials of the same degree, and let $I$ be the\nsmallest Borel ideal containing $M$ and $N$. We show that the toric ring of $I$\nis Koszul by constructing a quadratic Gröbner basis for the associated toric\nideal. Our proofs use the construction of graphs corresponding to fibers of the\ntoric map. As a consequence, we conclude that the Rees algebra is also Koszul.\n",
"title": "The Rees algebra of a two-Borel ideal is Koszul"
}
| null | null |
[
"Mathematics"
] | null | true | null |
4013
| null |
Validated
| null | null |
null |
{
"abstract": " We give a new expression for the law of the eigenvalues of the discrete\nAnderson model on the finite interval $[0,N]$, in terms of two random processes\nstarting at both ends of the interval. Using this formula, we deduce that the\ntail of the eigenvectors behaves approximatelylike $\\exp(\\sigma\nB\\_{|n-k|}-\\gamma\\frac{|n-k|}{4})$ where $B\\_{s}$ is the Brownian motion and\n$k$ is uniformly chosen in $[0,N]$ independentlyof $B\\_{s}$. A similar result\nhas recently been shown by B. Rifkind and B. Virag in the critical case, that\nis, when the random potential is multiplied by a factor $\\frac{1}{\\sqrt{N}}$\n",
"title": "A forward--backward random process for the spectrum of 1D Anderson operators"
}
| null | null | null | null | true | null |
4014
| null |
Default
| null | null |
null |
{
"abstract": " Given a curve defined over an algebraically closed field which is complete\nwith respect to a nontrivial valuation, we study its tropical Jacobian. This is\ndone by first tropicalizing the curve, and then computing the Jacobian of the\nresulting weighted metric graph. In general, it is not known how to find the\nabstract tropicalization of a curve defined by polynomial equations, since an\nembedded tropicalization may not be faithful, and there is no known algorithm\nfor carrying out semistable reduction in practice. We solve this problem in the\ncase of hyperelliptic curves by studying admissible covers. We also describe\nhow to take a weighted metric graph and compute its period matrix, which gives\nits tropical Jacobian and tropical theta divisor. Lastly, we describe the\npresent status of reversing this process, namely how to compute a curve which\nhas a given matrix as its period matrix.\n",
"title": "From Curves to Tropical Jacobians and Back"
}
| null | null |
[
"Mathematics"
] | null | true | null |
4015
| null |
Validated
| null | null |
null |
{
"abstract": " We consider importance sampling to estimate the probability $\\mu$ of a union\nof $J$ rare events $H_j$ defined by a random variable $\\boldsymbol{x}$. The\nsampler we study has been used in spatial statistics, genomics and\ncombinatorics going back at least to Karp and Luby (1983). It works by sampling\none event at random, then sampling $\\boldsymbol{x}$ conditionally on that event\nhappening and it constructs an unbiased estimate of $\\mu$ by multiplying an\ninverse moment of the number of occuring events by the union bound. We prove\nsome variance bounds for this sampler. For a sample size of $n$, it has a\nvariance no larger than $\\mu(\\bar\\mu-\\mu)/n$ where $\\bar\\mu$ is the union\nbound. It also has a coefficient of variation no larger than\n$\\sqrt{(J+J^{-1}-2)/(4n)}$ regardless of the overlap pattern among the $J$\nevents. Our motivating problem comes from power system reliability, where the\nphase differences between connected nodes have a joint Gaussian distribution\nand the $J$ rare events arise from unacceptably large phase differences. In the\ngrid reliability problems even some events defined by $5772$ constraints in\n$326$ dimensions, with probability below $10^{-22}$, are estimated with a\ncoefficient of variation of about $0.0024$ with only $n=10{,}000$ sample\nvalues.\n",
"title": "Importance sampling the union of rare events with an application to power systems analysis"
}
| null | null | null | null | true | null |
4016
| null |
Default
| null | null |
null |
{
"abstract": " Most network studies rely on an observed network that differs from the\nunderlying network which is obfuscated by measurement errors. It is well known\nthat such errors can have a severe impact on the reliability of network\nmetrics, especially on centrality measures: a more central node in the observed\nnetwork might be less central in the underlying network.\nWe introduce a metric for the reliability of centrality measures -- called\nsensitivity. Given two randomly chosen nodes, the sensitivity means the\nprobability that the more central node in the observed network is also more\ncentral in the underlying network. The sensitivity concept relies on the\nunderlying network which is usually not accessible. Therefore, we propose two\nmethods to approximate the sensitivity. The iterative method, which simulates\npossible underlying networks for the estimation and the imputation method,\nwhich uses the sensitivity of the observed network for the estimation. Both\nmethods rely on the observed network and assumptions about the underlying type\nof measurement error (e.g., the percentage of missing edges or nodes).\nOur experiments on real-world networks and random graphs show that the\niterative method performs well in many cases. In contrast, the imputation\nmethod does not yield useful estimations for networks other than\nErdős-Rényi graphs.\n",
"title": "Estimating the sensitivity of centrality measures w.r.t. measurement errors"
}
| null | null | null | null | true | null |
4017
| null |
Default
| null | null |
null |
{
"abstract": " Let ${\\cal X }=XX^{\\prime}$ be a random matrix associated with a centered\n$r$-column centered Gaussian vector $X$ with a covariance matrix $P$. In this\narticle we compute expectations of matrix-products of the form $\\prod_{1\\leq\ni\\leq n}({\\cal X } P^{v_i})$ for any $n\\geq 1$ and any multi-index parameters\n$v_i\\in\\mathbb{N}$. We derive closed form formulae and a simple sequential\nalgorithm to compute these matrices w.r.t. the parameter $n$. The second part\nof the article is dedicated to a non commutative binomial formula for the\ncentral matrix-moments $\\mathbb{E}\\left(\\left[{\\cal X }-P\\right]^n\\right)$. The\nmatrix product moments discussed in this study are expressed in terms of\npolynomial formulae w.r.t. the powers of the covariance matrix, with\ncoefficients depending on the trace of these matrices. We also derive a series\nof estimates w.r.t. the Loewner order on quadratic forms. For instance we shall\nprove the rather crude estimate $\\mathbb{E}\\left(\\left[{\\cal X\n}-P\\right]^n\\right)\\leq \\mathbb{E}\\left({\\cal X }^n-P^n\\right)$, for any $n\\geq\n1$\n",
"title": "Matrix product moments in normal variables"
}
| null | null | null | null | true | null |
4018
| null |
Default
| null | null |
null |
{
"abstract": " Bandwidth selection is crucial in the kernel estimation of density level\nsets. Risk based on the symmetric difference between the estimated and true\nlevel sets is usually used to measure their proximity. In this paper we provide\nan asymptotic $L^p$ approximation to this risk, where $p$ is characterized by\nthe weight function in the risk. In particular the excess risk corresponds to\nan $L^2$ type of risk, and is adopted in an optimal bandwidth selection rule\nfor nonparametric level set estimation of $d$-dimensional density functions\n($d\\geq 1$).\n",
"title": "Asymptotics and Optimal Bandwidth Selection for Nonparametric Estimation of Density Level Sets"
}
| null | null | null | null | true | null |
4019
| null |
Default
| null | null |
null |
{
"abstract": " Immunogenicity is a major problem during the development of biotherapeutics\nsince it can lead to rapid clearance of the drug and adverse reactions. The\nchallenge for biotherapeutic design is therefore to identify mutants of the\nprotein sequence that minimize immunogenicity in a target population whilst\nretaining pharmaceutical activity and protein function. Current approaches are\nmoderately successful in designing sequences with reduced immunogenicity, but\ndo not account for the varying frequencies of different human leucocyte antigen\nalleles in a specific population and in addition, since many designs are\nnon-functional, require costly experimental post-screening. Here we report a\nnew method for de-immunization design using multi-objective combinatorial\noptimization that simultaneously optimizes the likelihood of a functional\nprotein sequence at the same time as minimizing its immunogenicity tailored to\na target population. We bypass the need for three-dimensional protein structure\nor molecular simulations to identify functional designs by automatically\ngenerating sequences using probabilistic models that have been used previously\nfor mutation effect prediction and structure prediction. As proof-of-principle\nwe designed sequences of the C2 domain of Factor VIII and tested them\nexperimentally, resulting in a good correlation with the predicted\nimmunogenicity of our model.\n",
"title": "Population-specific design of de-immunized protein biotherapeutics"
}
| null | null | null | null | true | null |
4020
| null |
Default
| null | null |
null |
{
"abstract": " Probit regression was first proposed by Bliss in 1934 to study mortality\nrates of insects. Since then, an extensive body of work has analyzed and used\nprobit or related binary regression methods (such as logistic regression) in\nnumerous applications and fields. This paper provides a fresh angle to such\nwell-established binary regression methods. Concretely, we demonstrate that\nlinearizing the probit model in combination with linear estimators performs on\npar with state-of-the-art nonlinear regression methods, such as posterior mean\nor maximum aposteriori estimation, for a broad range of real-world regression\nproblems. We derive exact, closed-form, and nonasymptotic expressions for the\nmean-squared error of our linearized estimators, which clearly separates them\nfrom nonlinear regression methods that are typically difficult to analyze. We\nshowcase the efficacy of our methods and results for a number of synthetic and\nreal-world datasets, which demonstrates that linearized binary regression finds\npotential use in a variety of inference, estimation, signal processing, and\nmachine learning applications that deal with binary-valued observations or\nmeasurements.\n",
"title": "Linearized Binary Regression"
}
| null | null | null | null | true | null |
4021
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, first, we prove that the Diophantine system\n\\[f(z)=f(x)+f(y)=f(u)-f(v)=f(p)f(q)\\] has infinitely many integer solutions for\n$f(X)=X(X+a)$ with nonzero integers $a\\equiv 0,1,4\\pmod{5}$. Second, we show\nthat the above Diophantine system has an integer parametric solution for\n$f(X)=X(X+a)$ with nonzero integers $a$, if there are integers $m,n,k$ such\nthat \\[\\begin{cases} \\begin{split} (n^2-m^2) (4mnk(k+a+1) + a(m^2+2mn-n^2))\n&\\equiv0\\pmod{(m^2+n^2)^2},\\\\ (m^2+2mn-n^2) ((m^2-2mn-n^2)k(k+a+1) - 2amn)\n&\\equiv0 \\pmod{(m^2+n^2)^2}, \\end{split} \\end{cases}\\] where $k\\equiv0\\pmod{4}$\nwhen $a$ is even, and $k\\equiv2\\pmod{4}$ when $a$ is odd. Third, we get that\nthe Diophantine system \\[f(z)=f(x)+f(y)=f(u)-f(v)=f(p)f(q)=\\frac{f(r)}{f(s)}\\]\nhas a five-parameter rational solution for $f(X)=X(X+a)$ with nonzero rational\nnumber $a$ and infinitely many nontrivial rational parametric solutions for\n$f(X)=X(X+a)(X+b)$ with nonzero integers $a,b$ and $a\\neq b$. At last, we raise\nsome related questions.\n",
"title": "Arithmetic properties of polynomials"
}
| null | null | null | null | true | null |
4022
| null |
Default
| null | null |
null |
{
"abstract": " A lot of scientific works are published in different areas of science,\ntechnology, engineering and mathematics. It is not easy, even for experts, to\njudge the quality of authors, papers and venues (conferences and journals). An\nobjective measure to assign scores to these entities and to rank them is very\nuseful. Although, several metrics and indexes have been proposed earlier, they\nsuffer from various problems. In this paper, we propose a graph-based analytics\nframework to assign scores and to rank authors, papers and venues. Our\nalgorithm considers only the link structures of the underlying graphs. It does\nnot take into account other aspects, such as the associated texts and the\nreputation of these entities. In the limit of large number of iterations, the\nsolution of the iterative equations gives the unique entity scores. This\nframework can be easily extended to other interdependent networks.\n",
"title": "A Graph Analytics Framework for Ranking Authors, Papers and Venues"
}
| null | null | null | null | true | null |
4023
| null |
Default
| null | null |
null |
{
"abstract": " The main theorem is incorrectly stated.\n",
"title": "Inner Cohomology of the General Linear Group"
}
| null | null |
[
"Mathematics"
] | null | true | null |
4024
| null |
Validated
| null | null |
null |
{
"abstract": " We study fractional quantum Hall states at filling fractions in the Jain\nsequences using the framework of composite Dirac fermions. Synthesizing\nprevious work, we write down an effective field theory consistent with all\nsymmetry requirements, including Galilean invariance and particle-hole\nsymmetry. Employing a Fermi liquid description, we demonstrate the appearance\nof the Girvin--Macdonlald--Platzman algebra and compute the dispersion relation\nof neutral excitations and various response functions. Our results satisfy\nrequirements of particle-hole symmetry. We show that while the dispersion\nrelation obtained from the HLR theory is particle-hole symmetric, correlation\nfunctions obtained from HLR are not. The results of the Dirac theory are shown\nto be consistent with the Haldane bound on the projected structure factor,\nwhile those of the HLR theory violate it.\n",
"title": "Particle-hole symmetry and composite fermions in fractional quantum Hall states"
}
| null | null | null | null | true | null |
4025
| null |
Default
| null | null |
null |
{
"abstract": " We prove that Artin groups from a class containing all large-type Artin\ngroups are systolic. This provides a concise yet precise description of their\ngeometry. Immediate consequences are new results concerning large-type Artin\ngroups: biautomaticity; existence of $EZ$-boundaries; the Novikov conjecture;\ndescriptions of finitely presented subgroups, of virtually solvable subgroups,\nand of centralizers for infinite order elements; the Burghelea conjecture and\nthe Bass conjecture; existence of low-dimensional models for classifying spaces\nfor some families of subgroups.\n",
"title": "Large-type Artin groups are systolic"
}
| null | null | null | null | true | null |
4026
| null |
Default
| null | null |
null |
{
"abstract": " The chemotactic dynamics of cells and organisms that have no specialized\ngradient sensing organelles is not well understood. In fact, chemotaxis of this\nsort of organism is especially challenging to explain when the external\nchemical gradient is so small as to make variations of concentrations minute\nover the length of each of the organisms. Experimental evidence lends support\nto the conjecture that chemotactic behavior of chains of cells can be achieved\nvia cell-to-cell communication. This is the chemotactic basis for the Local\nExcitation, Global Inhibition (LEGI) model.\nA generalization of the model for the communication component of the LEGI\nmodel is proposed. Doing so permits us to study in detail how gradient sensing\nchanges as a function of the structure of the communication term. The key\nfindings of this study are, an accounting of how gradient sensing is affected\nby the competition of communication and diffusive processes; the determination\nof the scale dependence of the model outcomes; the sensitivity of communication\nto parameters in the model. Together with an essential analysis of the dynamics\nof the model, these findings can prove useful in suggesting experiments aimed\nat determining the viability of a communication mechanism in chemotactic\ndynamics of chains and networks of cells exposed to a chemical concentration\ngradient.\n",
"title": "Gradient Sensing via Cell Communication"
}
| null | null | null | null | true | null |
4027
| null |
Default
| null | null |
null |
{
"abstract": " A general procedure for constructing Yetter-Drinfeld modules from quantum\nprincipal bundles is introduced. As an application a Yetter-Drinfeld structure\nis put on the cotangent space of the Heckenberger-Kolb calculi of the quantum\nGrassmannians. For the special case of quantum projective space the associated\nbraiding is shown to be non-diagonal and of Hecke type. Moreover, its Nichols\nalgebra is shown to be finite-dimensional and equal to the anti-holomorphic\npart of the total differential calculus.\n",
"title": "Nichols Algebras and Quantum Principal Bundles"
}
| null | null |
[
"Mathematics"
] | null | true | null |
4028
| null |
Validated
| null | null |
null |
{
"abstract": " We present a method to reconstruct autocorrelated signals together with their\nautocorrelation structure from nonlinear, noisy measurements for arbitrary\nmonotonous nonlinear instrument response. In the presented formulation the\nalgorithm provides a significant speedup compared to prior implementations,\nallowing for a wider range of application. The nonlinearity can be used to\nmodel instrument characteristics or to enforce properties on the underlying\nsignal, such as positivity. Uncertainties on any posterior quantities can be\nprovided due to independent samples from an approximate posterior distribution.\nWe demonstrate the methods applicability via simulated and real measurements,\nusing different measurement instruments, nonlinearities and dimensionality.\n",
"title": "Inference of signals with unknown correlation structure from nonlinear measurements"
}
| null | null | null | null | true | null |
4029
| null |
Default
| null | null |
null |
{
"abstract": " An optimization-based approach for the Tucker tensor approximation of\nparameter-dependent data tensors and solutions of tensor differential equations\nwith low Tucker rank is presented. The problem of updating the tensor\ndecomposition is reformulated as fitting problem subject to the tangent space\nwithout relying on an orthogonality gauge condition. A discrete Euler scheme is\nestablished in an alternating least squares framework, where the quadratic\nsubproblems reduce to trace optimization problems, that are shown to be\nexplicitly solvable and accessible using SVD of small size. In the presence of\nsmall singular values, instability for larger ranks is reduced, since the\nmethod does not need the (pseudo) inverse of matricizations of the core tensor.\nRegularization of Tikhonov type can be used to compensate for the lack of\nuniqueness in the tangent space. The method is validated numerically and shown\nto be stable also for larger ranks in the case of small singular values of the\ncore unfoldings. Higher order explicit integrators of Runge-Kutta type can be\ncomposed.\n",
"title": "An optimization approach for dynamical Tucker tensor approximation"
}
| null | null | null | null | true | null |
4030
| null |
Default
| null | null |
null |
{
"abstract": " The optical absorption of CdWO$_4$ is reported at high pressures up to 23\nGPa. The onset of a phase transition was detected at 19.5 GPa, in good\nagreement with a previous Raman spectroscopy study. The crystal structure of\nthe high-pressure phase of CdWO$_4$ was solved at 22 GPa employing\nsingle-crystal synchrotron x-ray diffraction. The symmetry changes from space\ngroup $P$2/$c$ in the low-pressure wolframite phase to $P2_1/c$ in the\nhigh-pressure post-wolframite phase accompanied by a doubling of the unit-cell\nvolume. The octahedral oxygen coordination of the tungsten and cadmium ions is\nincreased to [7]-fold and [6+1]-fold, respectively, at the phase transition.\nThe compressibility of the low-pressure phase of CdWO$_4$ has been reevaluated\nwith powder x-ray diffraction up to 15 GPa finding a bulk modulus of $B_0$ =\n123 GPa. The direct band gap of the low-pressure phase increases with\ncompression up to 16.9 GPa at 12 meV/GPa. At this point an indirect band gap\ncrosses the direct band gap and decreases at -2 meV/GPa up to 19.5 GPa where\nthe phase transition starts. At the phase transition the band gap collapses by\n0.7 eV and another direct band gap decreases at -50 meV/GPa up to the maximum\nmeasured pressure. The structural stability of the post-wolframite structure is\nconfirmed by \\textit{ab initio} calculations finding the post-wolframite-type\nphase to be more stable than the wolframite at 18 GPa. Lattice dynamic\ncalculations based on space group $P2_1/c$ explain well the Raman-active modes\npreviously measured in the high-pressure post-wolframite phase. The\npressure-induced band gap crossing in the wolframite phase as well as the\npressure dependence of the direct band gap in the high-pressure phase are\nfurther discussed with respect to the calculations.\n",
"title": "Optical and structural study of the pressure-induced phase transition of CdWO$_4$"
}
| null | null |
[
"Physics"
] | null | true | null |
4031
| null |
Validated
| null | null |
null |
{
"abstract": " Convolutional neural networks (CNNs) have recently emerged as a popular\nbuilding block for natural language processing (NLP). Despite their success,\nmost existing CNN models employed in NLP share the same learned (and static)\nset of filters for all input sentences. In this paper, we consider an approach\nof using a small meta network to learn context-sensitive convolutional filters\nfor text processing. The role of meta network is to abstract the contextual\ninformation of a sentence or document into a set of input-aware filters. We\nfurther generalize this framework to model sentence pairs, where a\nbidirectional filter generation mechanism is introduced to encapsulate\nco-dependent sentence representations. In our benchmarks on four different\ntasks, including ontology classification, sentiment analysis, answer sentence\nselection, and paraphrase identification, our proposed model, a modified CNN\nwith context-sensitive filters, consistently outperforms the standard CNN and\nattention-based CNN baselines. By visualizing the learned context-sensitive\nfilters, we further validate and rationalize the effectiveness of proposed\nframework.\n",
"title": "Learning Context-Sensitive Convolutional Filters for Text Processing"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
4032
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper we study right $S$-Noetherian rings and modules, extending of\nnotions introduced by Anderson and Dumitrescu in commutative algebra to\nnoncommutative rings. Two characterizations of right $S$-Noetherian rings are\ngiven in terms of completely prime right ideals and point annihilator sets. We\nalso prove an existence result for completely prime point annihilators of\ncertain $S$-Noetherian modules with the following consequence in commutative\nalgebra: If a module $M$ over a commutative ring is $S$-Noetherian with respect\nto a multiplicative set $S$ that contains no zero-divisors for $M$, then $M$\nhas an associated prime.\n",
"title": "On right $S$-Noetherian rings and $S$-Noetherian modules"
}
| null | null | null | null | true | null |
4033
| null |
Default
| null | null |
null |
{
"abstract": " The oddball paradigm is widely applied to the investigation of multiple\ncognitive functions. Prior studies have explored the cortical oscillation and\npower spectral differing from the resting-state conduction to oddball paradigm,\nbut whether brain networks existing the significant difference is still\nunclear. Our study addressed how the brain reconfigures its architecture from a\nresting-state condition (i.e., baseline) to P300 stimulus task in the visual\noddball paradigm. In this study, electroencephalogram (EEG) datasets were\ncollected from 24 postgraduate students, who were required to only mentally\ncount the number of target stimulus; afterwards the functional EEG networks\nconstructed in different frequency bands were compared between baseline and\noddball task conditions to evaluate the reconfiguration of functional network\nin the brain. Compared to the baseline, our results showed the significantly (p\n< 0.05) enhanced delta/theta EEG connectivity and decreased alpha default mode\nnetwork in the progress of brain reconfiguration to the P300 task. Furthermore,\nthe reconfigured coupling strengths were demonstrated to relate to P300\namplitudes, which were then regarded as input features to train a classifier to\ndifferentiate the high and low P300 amplitudes groups with an accuracy of\n77.78%. The findings of our study help us to understand the changes of\nfunctional brain connectivity from resting-state to oddball stimulus task, and\nthe reconfigured network pattern has the potential for the selection of good\nsubjects for P300-based brain- computer interface.\n",
"title": "Reconfiguration of Brain Network between Resting-state and Oddball Paradigm"
}
| null | null |
[
"Quantitative Biology"
] | null | true | null |
4034
| null |
Validated
| null | null |
null |
{
"abstract": " A common problem in machine learning is to rank a set of n items based on\npairwise comparisons. Here ranking refers to partitioning the items into sets\nof pre-specified sizes according to their scores, which includes identification\nof the top-k items as the most prominent special case. The score of a given\nitem is defined as the probability that it beats a randomly chosen other item.\nFinding an exact ranking typically requires a prohibitively large number of\ncomparisons, but in practice, approximate rankings are often adequate.\nAccordingly, we study the problem of finding approximate rankings from pairwise\ncomparisons. We analyze an active ranking algorithm that counts the number of\ncomparisons won, and decides whether to stop or which pair of items to compare\nnext, based on confidence intervals computed from the data collected in\nprevious steps. We show that this algorithm succeeds in recovering approximate\nrankings using a number of comparisons that is close to optimal up to\nlogarithmic factors. We also present numerical results, showing that in\npractice, approximation can drastically reduce the number of comparisons\nrequired to estimate a ranking.\n",
"title": "Approximate Ranking from Pairwise Comparisons"
}
| null | null | null | null | true | null |
4035
| null |
Default
| null | null |
null |
{
"abstract": " Human activities from hunting to emailing are performed in a fractal-like\nscale invariant pattern. These patterns are considered efficient for hunting or\nforaging, but are they efficient for gathering information? Here we link the\nscale invariant pattern of inter-touch intervals on the smartphone to optimal\nstrategies for information gathering. We recorded touchscreen touches in 65\nindividuals for a month and categorized the activity into checking for\ninformation vs. sharing content. For both categories, the inter-touch intervals\nwere well described by power-law fits spanning 5 orders of magnitude, from 1 s\nto several hours. The power-law exponent typically found for checking was 1.5\nand for generating it was 1.3. Next, by using computer simulations we addressed\nwhether the checking pattern was efficient - in terms of minimizing futile\nattempts yielding no new information. We find that the best performing power\nlaw exponent depends on the duration of the assessment and the exponent of 1.5\nwas the most efficient in the short-term i.e. in the few minutes range.\nFinally, we addressed whether how people generated and shared content was in\ntune with the checking pattern. We assumed that the unchecked posts must be\nminimized for maximal efficiency and according to our analysis the most\nefficient temporal pattern to share content was the exponent of 1.3 - which was\nalso the pattern displayed by the smartphone users. The behavioral organization\nfor content generation is different from content consumption across time\nscales. We propose that this difference is a signature of optimal behavior and\nthe short-term assessments used in modern human actions.\n",
"title": "Optimised information gathering in smartphone users"
}
| null | null | null | null | true | null |
4036
| null |
Default
| null | null |
null |
{
"abstract": " In this paper the problem of selecting $p$ out of $n$ available items is\ndiscussed, such that their total cost is minimized. We assume that costs are\nnot known exactly, but stem from a set of possible outcomes.\nRobust recoverable and two-stage models of this selection problem are\nanalyzed. In the two-stage problem, up to $p$ items is chosen in the first\nstage, and the solution is completed once the scenario becomes revealed in the\nsecond stage. In the recoverable problem, a set of $p$ items is selected in the\nfirst stage, and can be modified by exchanging up to $k$ items in the second\nstage, after a scenario reveals.\nWe assume that uncertain costs are modeled through bounded uncertainty sets,\ni.e., the interval uncertainty sets with an additional linear (budget)\nconstraint, in their discrete and continuous variants. Polynomial algorithms\nfor recoverable and two-stage selection problems with continuous bounded\nuncertainty, and compact mixed integer formulations in the case of discrete\nbounded uncertainty are constructed.\n",
"title": "On Recoverable and Two-Stage Robust Selection Problems with Budgeted Uncertainty"
}
| null | null |
[
"Computer Science",
"Mathematics"
] | null | true | null |
4037
| null |
Validated
| null | null |
null |
{
"abstract": " We study anisotropic undersampling schemes like those used in\nmulti-dimensional NMR spectroscopy and MR imaging, which sample exhaustively in\ncertain time dimensions and randomly in others.\nOur analysis shows that anisotropic undersampling schemes are equivalent to\ncertain block-diagonal measurement systems. We develop novel exact formulas for\nthe sparsity/undersampling tradeoffs in such measurement systems. Our formulas\npredict finite-N phase transition behavior differing substantially from the\nwell known asymptotic phase transitions for classical Gaussian undersampling.\nExtensive empirical work shows that our formulas accurately describe observed\nfinite-N behavior, while the usual formulas based on universality are\nsubstantially inaccurate.\nWe also vary the anisotropy, keeping the total number of samples fixed, and\nfor each variation we determine the precise sparsity/undersampling tradeoff\n(phase transition). We show that, other things being equal, the ability to\nrecover a sparse object decreases with an increasing number of\nexhaustively-sampled dimensions.\n",
"title": "Sparsity/Undersampling Tradeoffs in Anisotropic Undersampling, with Applications in MR Imaging/Spectroscopy"
}
| null | null | null | null | true | null |
4038
| null |
Default
| null | null |
null |
{
"abstract": " Given a graph, the sparsest cut problem asks for a subset of vertices whose\nedge expansion (the normalized cut given by the subset) is minimized. In this\npaper, we study a generalization of this problem seeking for $ k $ disjoint\nsubsets of vertices (clusters) whose all edge expansions are small and\nfurthermore, the number of vertices remained in the exterior of the subsets\n(outliers) is also small. We prove that although this problem is $ NP-$hard for\ntrees, it can be solved in polynomial time for all weighted trees, provided\nthat we restrict the search space to subsets which induce connected subgraphs.\nThe proposed algorithm is based on dynamic programming and runs in the worst\ncase in $ O(k^2 n^3) $, when $ n $ is the number of vertices and $ k $ is the\nnumber of clusters. It also runs in linear time when the number of clusters and\nthe number of outliers is bounded by a constant.\n",
"title": "Multi-way sparsest cut problem on trees with a control on the number of parts and outliers"
}
| null | null | null | null | true | null |
4039
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we propose a well-justified synthetic approach of the\nprojective space. We define the concepts of plane and space of incidence and\nalso the Gallucci's axiom as an axiom to our classical projective space. To\nthis purpose we prove from our space axioms, the theorems of Desargues, Pappus,\nthe fundamental theorem of projectivities, and the fundamental theorem of\ncentral-axial collinearities, respectively. Our building up do not use any\ninformation on analytical projective geometry, as the concept of cross-ratio\nand the homogeneous coordinates of points.\n",
"title": "Gallucci's axiom revisited"
}
| null | null | null | null | true | null |
4040
| null |
Default
| null | null |
null |
{
"abstract": " A recent stacking analysis of Planck HFI data of galaxy clusters (Hurier\n2016) allowed to derive the cluster temperatures by using the relativistic\ncorrections to the Sunyaev-Zel'dovich effect (SZE). However, the temperatures\nof high-temperature clusters, as derived from this analysis, resulted to be\nbasically higher than the temperatures derived from X-ray measurements, at a\nmoderate statistical significance of $1.5\\sigma$. This discrepancy has been\nattributed by Hurier (2016) to calibration issues. In this paper we discuss an\nalternative explanation for this discrepancy in terms of a non-thermal SZE\nastrophysical component. We find that this explanation can work if non-thermal\nelectrons in galaxy clusters have a low value of their minimum momentum\n($p_1\\sim0.5-1$), and if their pressure is of the order of $20-30\\%$ of the\nthermal gas pressure. Both these conditions are hard to obtain if the\nnon-thermal electrons are mixed with the hot gas in the intra cluster medium,\nbut can be possibly obtained if the non-thermal electrons are mainly confined\nin bubbles with high content of non-thermal plasma and low content of thermal\nplasma, or in giant radio lobes/relics located in the outskirts of clusters. In\norder to derive more precise results on the properties of non-thermal electrons\nin clusters, and in view of more solid detections of a discrepancy between\nX-rays and SZE derived clusters temperatures that cannot be explained in other\nways, it would be necessary to reproduce the full analysis done by Hurier\n(2016) by adding systematically the non-thermal component of the SZE.\n",
"title": "Effect of the non-thermal Sunyaev-Zel'dovich Effect on the temperature determination of galaxy clusters"
}
| null | null | null | null | true | null |
4041
| null |
Default
| null | null |
null |
{
"abstract": " Background: Model-based analysis of movements can help better understand\nhuman motor control. Here, the models represent the human body as an\narticulated multi-body system that reflects the characteristics of the human\nbeing studied.\nResults: We present an open-source toolbox that allows for the creation of\nhuman models with easy-to-setup, customizable configurations. The toolbox\nscripts are written in Matlab/Octave and provide a command-based interface as\nwell as a graphical interface to construct, visualize and export models.\nBuilt-in software modules provide functionalities such as automatic scaling of\nmodels based on subject height and weight, custom scaling of segment lengths,\nmass and inertia, addition of body landmarks, and addition of motion capture\nmarkers. Users can set up custom definitions of joints, segments and other body\nproperties using the many included examples as templates. In addition to the\nhuman, any number of objects (e.g. exoskeletons, orthoses, prostheses, boxes)\ncan be added to the modeling environment.\nConclusions: The ModelFactory toolbox is published as open-source software\nunder the permissive zLib license. The toolbox fulfills an important function\nby making it easier to create human models, and should be of interest to human\nmovement researchers.\nThis document is the author's version of this article.\n",
"title": "ModelFactory: A Matlab/Octave based toolbox to create human body models"
}
| null | null | null | null | true | null |
4042
| null |
Default
| null | null |
null |
{
"abstract": " In this study, we propose a new statical approach for high-dimensionality\nreduction of heterogenous data that limits the curse of dimensionality and\ndeals with missing values. To handle these latter, we propose to use the Random\nForest imputation's method. The main purpose here is to extract useful\ninformation and so reducing the search space to facilitate the data exploration\nprocess. Several illustrative numeric examples, using data coming from publicly\navailable machine learning repositories are also included. The experimental\ncomponent of the study shows the efficiency of the proposed analytical\napproach.\n",
"title": "Dimensionality reduction with missing values imputation"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
4043
| null |
Validated
| null | null |
null |
{
"abstract": " The recent direct observation of gravitational waves (GW) from merging black\nholes opens up the possibility of exploring the theory of gravity in the strong\nregime at an unprecedented level. It is therefore interesting to explore which\nextensions to General Relativity (GR) could be detected. We construct an\nEffective Field Theory (EFT) satisfying the following requirements. It is\ntestable with GW observations; it is consistent with other experiments,\nincluding short distance tests of GR; it agrees with widely accepted principles\nof physics, such as locality, causality and unitarity; and it does not involve\nnew light degrees of freedom. The most general theory satisfying these\nrequirements corresponds to adding to the GR Lagrangian operators constructed\nout of powers of the Riemann tensor, suppressed by a scale comparable to the\ncurvature of the observed merging binaries. The presence of these operators\nmodifies the gravitational potential between the compact objects, as well as\ntheir effective mass and current quadrupoles, ultimately correcting the\nwaveform of the emitted GW.\n",
"title": "An effective formalism for testing extensions to General Relativity with gravitational waves"
}
| null | null |
[
"Physics"
] | null | true | null |
4044
| null |
Validated
| null | null |
null |
{
"abstract": " By formally invoking the Wiener-Hopf method, we explicitly solve a\none-dimensional, singular integral equation for the excitation of a slowly\ndecaying electromagnetic wave, called surface plasmon-polariton (SPP), of small\nwavelength on a semi-infinite, flat conducting sheet irradiated by a plane wave\nin two spatial dimensions. This setting is germane to wave diffraction by edges\nof large sheets of single-layer graphene. Our analytical approach includes: (i)\nformulation of a functional equation in the Fourier domain; (ii) evaluation of\na split function, which is expressed by a contour integral and is a key\ningredient of the Wiener-Hopf factorization; and (iii) extraction of the SPP as\na simple-pole residue of a Fourier integral. Our analytical solution is in good\nagreement with a finite-element numerical computation.\n",
"title": "On the Wiener-Hopf method for surface plasmons: Diffraction from semi-infinite metamaterial sheet"
}
| null | null |
[
"Mathematics"
] | null | true | null |
4045
| null |
Validated
| null | null |
null |
{
"abstract": " Humans are remarkably proficient at controlling their limbs and tools from a\nwide range of viewpoints and angles, even in the presence of optical\ndistortions. In robotics, this ability is referred to as visual servoing:\nmoving a tool or end-point to a desired location using primarily visual\nfeedback. In this paper, we study how viewpoint-invariant visual servoing\nskills can be learned automatically in a robotic manipulation scenario. To this\nend, we train a deep recurrent controller that can automatically determine\nwhich actions move the end-point of a robotic arm to a desired object. The\nproblem that must be solved by this controller is fundamentally ambiguous:\nunder severe variation in viewpoint, it may be impossible to determine the\nactions in a single feedforward operation. Instead, our visual servoing system\nmust use its memory of past movements to understand how the actions affect the\nrobot motion from the current viewpoint, correcting mistakes and gradually\nmoving closer to the target. This ability is in stark contrast to most visual\nservoing methods, which either assume known dynamics or require a calibration\nphase. We show how we can learn this recurrent controller using simulated data\nand a reinforcement learning objective. We then describe how the resulting\nmodel can be transferred to a real-world robot by disentangling perception from\ncontrol and only adapting the visual layers. The adapted model can servo to\npreviously unseen objects from novel viewpoints on a real-world Kuka IIWA\nrobotic arm. For supplementary videos, see:\nthis https URL\n",
"title": "Sim2Real View Invariant Visual Servoing by Recurrent Control"
}
| null | null |
[
"Computer Science"
] | null | true | null |
4046
| null |
Validated
| null | null |
null |
{
"abstract": " We consider the 3-point blow-up of the manifold $ (S^2 \\times S^2, \\sigma\n\\oplus \\sigma)$ where $\\sigma$ is the standard symplectic form which gives area\n1 to the sphere $S^2$, and study its group of symplectomorphisms $\\rm{Symp} (\nS^2 \\times S^2 \\#\\, 3\\overline{\\mathbb C\\mathbb P}\\,\\!^2, \\omega)$. So far, the\nmonotone case was studied by J. Evans and he proved that this group is\ncontractible. Moreover, J. Li, T. J. Li and W. Wu showed that the group\nSymp$_{h}(S^2 \\times S^2 \\#\\, 3\\overline{ \\mathbb C\\mathbb P}\\,\\!^2,\\omega) $\nof symplectomorphisms that act trivially on homology is always connected and\nrecently they also computed its fundamental group. We describe, in full detail,\nthe rational homotopy Lie algebra of this group.\nWe show that some particular circle actions contained in the\nsymplectomorphism group generate its full topology. More precisely, they give\nthe generators of the homotopy graded Lie algebra of $\\rm{Symp} (S^2 \\times S^2\n\\#\\, 3\\overline{ \\mathbb C\\mathbb P}\\,\\!^2, \\omega)$. Our study depends on\nKarshon's classification of Hamiltonian circle actions and the inflation\ntechnique introduced by Lalonde-McDuff. As an application, we deduce the rank\nof the homotopy groups of $\\rm{Symp}({\\mathbb C\\mathbb P}^2 \\#\\,\n5\\overline{\\mathbb C\\mathbb P}\\,\\!^2, \\tilde \\omega)$, in the case of small\nblow-ups.\n",
"title": "The homotopy Lie algebra of symplectomorphism groups of 3-fold blow-ups of $(S^2 \\times S^2, σ_{std} \\oplus σ_{std}) $"
}
| null | null | null | null | true | null |
4047
| null |
Default
| null | null |
null |
{
"abstract": " In the present investigation, the development of axial velocity profile, the\nrequirement for development length ($L^*_{fd}=L/D_{h}$) and the pressure drop\nin the entrance region of circular and parallel plate micro-channels have been\ncritically analysed for a large range of operating conditions ($10^{-2}\\le\nRe\\le 10^{4}$, $10^{-4}\\le Kn\\le 0.2$ and $0\\le C_2\\le 0.5$). For this purpose,\nthe conventional Navier-Stokes equations have been numerically solved using the\nfinite volume method on non-staggered grid, while employing the second-order\nvelocity slip condition at the wall with $C_1=1$. The results indicate that\nalthough the magnitude of local velocity slip at the wall is always greater\nthan that for the fully-developed section, the local wall shear stress,\nparticularly for higher $Kn$ and $C_2$, could be considerably lower than its\nfully-developed value. This effect, which is more prominent for lower $Re$,\nsignificantly affects the local and the fully-developed incremental pressure\ndrop number $K(x)$ and $K_{fd}$, respectively. As a result, depending upon the\noperating condition, $K_{fd}$, as well as $K(x)$, could assume negative values.\nThis never reported observation implies that in the presence of enhanced\nvelocity slip at the wall, the pressure gradient in the developing region could\neven be less than that in the fully-developed section. From simulated data, it\nhas been observed that both $L^*_{fd}$ and $K_{fd}$ are characterised by the\nlow and the high $Re$ asymptotes, using which, extremely accurate correlations\nfor them have been proposed for both geometries. Although owing to the complex\nnature, no correlation could be derived for $K(x)$ and an exact knowledge of\n$K(x)$ is necessary for evaluating the actual pressure drop for a duct length\n$L^*<L^*_{fd}$, a method has been proposed that provides a conservative\nestimate of the pressure drop for both $K_{fd}>0$ and $K_{fd}\\le0$.\n",
"title": "Pressure Drop and Flow development in the Entrance Region of Micro-Channels with Second Order Slip Boundary Conditions and the Requirement for Development Length"
}
| null | null |
[
"Physics"
] | null | true | null |
4048
| null |
Validated
| null | null |
null |
{
"abstract": " We describe a method to identify poor households in data-scarce countries by\nleveraging information contained in nationally representative household\nsurveys. It employs standard statistical learning techniques---cross-validation\nand parameter regularization---which together reduce the extent to which the\nmodel is over-fitted to match the idiosyncracies of observed survey data. The\nautomated framework satisfies three important constraints of this development\nsetting: i) The prediction model uses at most ten questions, which limits the\ncosts of data collection; ii) No computation beyond simple arithmetic is needed\nto calculate the probability that a given household is poor, immediately after\ndata on the ten indicators is collected; and iii) One specification of the\nmodel (i.e. one scorecard) is used to predict poverty throughout a country that\nmay be characterized by significant sub-national differences. Using survey data\nfrom Zambia, the model's out-of-sample predictions distinguish poor households\nfrom non-poor households using information contained in ten questions.\n",
"title": "Household poverty classification in data-scarce environments: a machine learning approach"
}
| null | null |
[
"Statistics"
] | null | true | null |
4049
| null |
Validated
| null | null |
null |
{
"abstract": " Recent research in computational linguistics has developed algorithms which\nassociate matrices with adjectives and verbs, based on the distribution of\nwords in a corpus of text. These matrices are linear operators on a vector\nspace of context words. They are used to construct the meaning of composite\nexpressions from that of the elementary constituents, forming part of a\ncompositional distributional approach to semantics. We propose a Matrix Theory\napproach to this data, based on permutation symmetry along with Gaussian\nweights and their perturbations. A simple Gaussian model is tested against word\nmatrices created from a large corpus of text. We characterize the cubic and\nquartic departures from the model, which we propose, alongside the Gaussian\nparameters, as signatures for comparison of linguistic corpora. We propose that\nperturbed Gaussian models with permutation symmetry provide a promising\nframework for characterizing the nature of universality in the statistical\nproperties of word matrices. The matrix theory framework developed here\nexploits the view of statistics as zero dimensional perturbative quantum field\ntheory. It perceives language as a physical system realizing a universality\nclass of matrix statistics characterized by permutation symmetry.\n",
"title": "Linguistic Matrix Theory"
}
| null | null | null | null | true | null |
4050
| null |
Default
| null | null |
null |
{
"abstract": " Ponzi schemes are financial frauds where, under the promise of high profits,\nusers put their money, recovering their investment and interests only if enough\nusers after them continue to invest money. Originated in the offline world 150\nyears ago, Ponzi schemes have since then migrated to the digital world,\napproaching first on the Web, and more recently hanging over cryptocurrencies\nlike Bitcoin. Smart contract platforms like Ethereum have provided a new\nopportunity for scammers, who have now the possibility of creating\n\"trustworthy\" frauds that still make users lose money, but at least are\nguaranteed to execute \"correctly\". We present a comprehensive survey of Ponzi\nschemes on Ethereum, analysing their behaviour and their impact from various\nviewpoints. Perhaps surprisingly, we identify a remarkably high number of Ponzi\nschemes, despite the hosting platform has been operating for less than two\nyears.\n",
"title": "Dissecting Ponzi schemes on Ethereum: identification, analysis, and impact"
}
| null | null | null | null | true | null |
4051
| null |
Default
| null | null |
null |
{
"abstract": " It is well known that it is challenging to train deep neural networks and\nrecurrent neural networks for tasks that exhibit long term dependencies. The\nvanishing or exploding gradient problem is a well known issue associated with\nthese challenges. One approach to addressing vanishing and exploding gradients\nis to use either soft or hard constraints on weight matrices so as to encourage\nor enforce orthogonality. Orthogonal matrices preserve gradient norm during\nbackpropagation and may therefore be a desirable property. This paper explores\nissues with optimization convergence, speed and gradient stability when\nencouraging or enforcing orthogonality. To perform this analysis, we propose a\nweight matrix factorization and parameterization strategy through which we can\nbound matrix norms and therein control the degree of expansivity induced during\nbackpropagation. We find that hard constraints on orthogonality can negatively\naffect the speed of convergence and model performance.\n",
"title": "On orthogonality and learning recurrent networks with long term dependencies"
}
| null | null | null | null | true | null |
4052
| null |
Default
| null | null |
null |
{
"abstract": " In typical neural machine translation~(NMT), the decoder generates a sentence\nword by word, packing all linguistic granularities in the same time-scale of\nRNN. In this paper, we propose a new type of decoder for NMT, which splits the\ndecode state into two parts and updates them in two different time-scales.\nSpecifically, we first predict a chunk time-scale state for phrasal modeling,\non top of which multiple word time-scale states are generated. In this way, the\ntarget sentence is translated hierarchically from chunks to words, with\ninformation in different granularities being leveraged. Experiments show that\nour proposed model significantly improves the translation performance over the\nstate-of-the-art NMT model.\n",
"title": "Chunk-Based Bi-Scale Decoder for Neural Machine Translation"
}
| null | null | null | null | true | null |
4053
| null |
Default
| null | null |
null |
{
"abstract": " Deep neural networks (DNN) excel at extracting patterns. Through\nrepresentation learning and automated feature engineering on large datasets,\nsuch models have been highly successful in computer vision and natural language\napplications. Designing optimal network architectures from a principled or\nrational approach however has been less than successful, with the best\nsuccessful approaches utilizing an additional machine learning algorithm to\ntune the network hyperparameters. However, in many technical fields, there\nexist established domain knowledge and understanding about the subject matter.\nIn this work, we develop a novel furcated neural network architecture that\nutilizes domain knowledge as high-level design principles of the network. We\ndemonstrate proof-of-concept by developing IL-Net, a furcated network for\npredicting the properties of ionic liquids, which is a class of complex\nmulti-chemicals entities. Compared to existing state-of-the-art approaches, we\nshow that furcated networks can improve model accuracy by approximately 20-35%,\nwithout using additional labeled data. Lastly, we distill two key design\nprinciples for furcated networks that can be adapted to other domains.\n",
"title": "IL-Net: Using Expert Knowledge to Guide the Design of Furcated Neural Networks"
}
| null | null | null | null | true | null |
4054
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we introduce an algorithm for performing spectral clustering\nefficiently. Spectral clustering is a powerful clustering algorithm that\nsuffers from high computational complexity, due to eigen decomposition. In this\nwork, we first build the adjacency matrix of the corresponding graph of the\ndataset. To build this matrix, we only consider a limited number of points,\ncalled landmarks, and compute the similarity of all data points with the\nlandmarks. Then, we present a definition of the Laplacian matrix of the graph\nthat enable us to perform eigen decomposition efficiently, using a deep\nautoencoder. The overall complexity of the algorithm for eigen decomposition is\n$O(np)$, where $n$ is the number of data points and $p$ is the number of\nlandmarks. At last, we evaluate the performance of the algorithm in different\nexperiments.\n",
"title": "Fast Spectral Clustering Using Autoencoders and Landmarks"
}
| null | null | null | null | true | null |
4055
| null |
Default
| null | null |
null |
{
"abstract": " Advances in mobile computing technologies have made it possible to monitor\nand apply data-driven interventions across complex systems in real time. Markov\ndecision processes (MDPs) are the primary model for sequential decision\nproblems with a large or indefinite time horizon. Choosing a representation of\nthe underlying decision process that is both Markov and low-dimensional is\nnon-trivial. We propose a method for constructing a low-dimensional\nrepresentation of the original decision process for which: 1. the MDP model\nholds; 2. a decision strategy that maximizes mean utility when applied to the\nlow-dimensional representation also maximizes mean utility when applied to the\noriginal process. We use a deep neural network to define a class of potential\nprocess representations and estimate the process of lowest dimension within\nthis class. The method is illustrated using data from a mobile study on heavy\ndrinking and smoking among college students.\n",
"title": "Sufficient Markov Decision Processes with Alternating Deep Neural Networks"
}
| null | null | null | null | true | null |
4056
| null |
Default
| null | null |
null |
{
"abstract": " Interfacing a ferromagnet with a polarized ferroelectric gate generates a\nnon-uniform, interfacial spin density coupled to the ferroelectric polarization\nallowing so for an electric field control of effective transversal field to\nmagnetization. Here we study the dynamic magnetization switching behavior of\nsuch a multilayer system based on the Landau-Lifshitz-Baryakhtar equation,\ndemonstrating that interfacial magnetoelectric coupling is utilizable as a\nhighly localized and efficient tool for manipulating magnetism.\n",
"title": "Gate-controlled magnonic-assisted switching of magnetization in ferroelectric/ferromagnetic junctions"
}
| null | null |
[
"Physics"
] | null | true | null |
4057
| null |
Validated
| null | null |
null |
{
"abstract": " By combining bulk sensitive soft-X-ray angular-resolved photoemission\nspectroscopy and accurate first-principles calculations we explored the bulk\nelectronic properties of WTe$_2$, a candidate type-II Weyl semimetal featuring\na large non-saturating magnetoresistance. Despite the layered geometry\nsuggesting a two-dimensional electronic structure, we find a three-dimensional\nelectronic dispersion. We report an evident band dispersion in the reciprocal\ndirection perpendicular to the layers, implying that electrons can also travel\ncoherently when crossing from one layer to the other. The measured Fermi\nsurface is characterized by two well-separated electron and hole pockets at\neither side of the $\\Gamma$ point, differently from previous more surface\nsensitive ARPES experiments that additionally found a significant quasiparticle\nweight at the zone center. Moreover, we observe a significant sensitivity of\nthe bulk electronic structure of WTe$_2$ around the Fermi level to electronic\ncorrelations and renormalizations due to self-energy effects, previously\nneglected in first-principles descriptions.\n",
"title": "Three-Dimensional Electronic Structure of type-II Weyl Semimetal WTe$_2$"
}
| null | null |
[
"Physics"
] | null | true | null |
4058
| null |
Validated
| null | null |
null |
{
"abstract": " We consider multi-agent stochastic optimization problems over reproducing\nkernel Hilbert spaces (RKHS). In this setting, a network of interconnected\nagents aims to learn decision functions, i.e., nonlinear statistical models,\nthat are optimal in terms of a global convex functional that aggregates data\nacross the network, with only access to locally and sequentially observed\nsamples. We propose solving this problem by allowing each agent to learn a\nlocal regression function while enforcing consensus constraints. We use a\npenalized variant of functional stochastic gradient descent operating\nsimultaneously with low-dimensional subspace projections. These subspaces are\nconstructed greedily by applying orthogonal matching pursuit to the sequence of\nkernel dictionaries and weights. By tuning the projection-induced bias, we\npropose an algorithm that allows for each individual agent to learn, based upon\nits locally observed data stream and message passing with its neighbors only, a\nregression function that is close to the globally optimal regression function.\nThat is, we establish that with constant step-size selections agents' functions\nconverge to a neighborhood of the globally optimal one while satisfying the\nconsensus constraints as the penalty parameter is increased. Moreover, the\ncomplexity of the learned regression functions is guaranteed to remain finite.\nOn both multi-class kernel logistic regression and multi-class kernel support\nvector classification with data generated from class-dependent Gaussian mixture\nmodels, we observe stable function estimation and state of the art performance\nfor distributed online multi-class classification. Experiments on the Brodatz\ntextures further substantiate the empirical validity of this approach.\n",
"title": "Decentralized Online Learning with Kernels"
}
| null | null |
[
"Computer Science",
"Mathematics",
"Statistics"
] | null | true | null |
4059
| null |
Validated
| null | null |
null |
{
"abstract": " Let $\\mathbb{F}_q$ denote the finite field of order $q,$ $n$ be a positive\ninteger coprime to $q$ and $t \\geq 2$ be an integer. In this paper, we\nenumerate all the complementary-dual cyclic $\\mathbb{F}_q$-linear\n$\\mathbb{F}_{q^t}$-codes of length $n$ by placing $\\ast$, ordinary and\nHermitian trace bilinear forms on $\\mathbb{F}_{q^t}^n.$\n",
"title": "Enumeration of complementary-dual cyclic $\\mathbb{F}_{q}$-linear $\\mathbb{F}_{q^t}$-codes"
}
| null | null | null | null | true | null |
4060
| null |
Default
| null | null |
null |
{
"abstract": " Bilinear models provide an appealing framework for mixing and merging\ninformation in Visual Question Answering (VQA) tasks. They help to learn high\nlevel associations between question meaning and visual concepts in the image,\nbut they suffer from huge dimensionality issues. We introduce MUTAN, a\nmultimodal tensor-based Tucker decomposition to efficiently parametrize\nbilinear interactions between visual and textual representations. Additionally\nto the Tucker framework, we design a low-rank matrix-based decomposition to\nexplicitly constrain the interaction rank. With MUTAN, we control the\ncomplexity of the merging scheme while keeping nice interpretable fusion\nrelations. We show how our MUTAN model generalizes some of the latest VQA\narchitectures, providing state-of-the-art results.\n",
"title": "MUTAN: Multimodal Tucker Fusion for Visual Question Answering"
}
| null | null | null | null | true | null |
4061
| null |
Default
| null | null |
null |
{
"abstract": " Two new high-precision measurements of the deuterium abundance from absorbers\nalong the line of sight to the quasar PKS1937--1009 were presented. The\nabsorbers have lower neutral hydrogen column densities (N(HI) $\\approx$\n18\\,cm$^{-2}$) than for previous high-precision measurements, boding well for\nfurther extensions of the sample due to the plenitude of low column density\nabsorbers. The total high-precision sample now consists of 12 measurements with\na weighted average deuterium abundance of D/H = $2.55\\pm0.02\\times10^{-5}$. The\nsample does not favour a dipole similar to the one detected for the fine\nstructure constant. The increased precision also calls for improved\nnucleosynthesis predictions. For that purpose we have updated the public\nAlterBBN code including new reactions, updated nuclear reaction rates, and the\npossibility of adding new physics such as dark matter. The standard Big Bang\nNucleosynthesis prediction of D/H = $2.456\\pm0.057\\times10^{-5}$ is consistent\nwith the observed value within 1.7 standard deviations.\n",
"title": "Nucleosynthesis Predictions and High-Precision Deuterium Measurements"
}
| null | null | null | null | true | null |
4062
| null |
Default
| null | null |
null |
{
"abstract": " This paper proposes a scalable algorithmic framework for spectral reduction\nof large undirected graphs. The proposed method allows computing much smaller\ngraphs while preserving the key spectral (structural) properties of the\noriginal graph. Our framework is built upon the following two key components: a\nspectrum-preserving node aggregation (reduction) scheme, as well as a spectral\ngraph sparsification framework with iterative edge weight scaling. We show that\nthe resulting spectrally-reduced graphs can robustly preserve the first few\nnontrivial eigenvalues and eigenvectors of the original graph Laplacian. In\naddition, the spectral graph reduction method has been leveraged to develop\nmuch faster algorithms for multilevel spectral graph partitioning as well as\nt-distributed Stochastic Neighbor Embedding (t-SNE) of large data sets. We\nconducted extensive experiments using a variety of large graphs and data sets,\nand obtained very promising results. For instance, we are able to reduce the\n\"coPapersCiteseer\" graph with 0.43 million nodes and 16 million edges to a much\nsmaller graph with only 13K (32X fewer) nodes and 17K (950X fewer) edges in\nabout 16 seconds; the spectrally-reduced graphs also allow us to achieve up to\n1100X speedup for spectral graph partitioning and up to 60X speedup for t-SNE\nvisualization of large data sets.\n",
"title": "Nearly-Linear Time Spectral Graph Reduction for Scalable Graph Partitioning and Data Visualization"
}
| null | null | null | null | true | null |
4063
| null |
Default
| null | null |
null |
{
"abstract": " We introduce the first index that can be built in $o(n)$ time for a text of\nlength $n$, and also queried in $o(m)$ time for a pattern of length $m$. On a\nconstant-size alphabet, for example, our index uses\n$O(n\\log^{1/2+\\varepsilon}n)$ bits, is built in $O(n/\\log^{1/2-\\varepsilon} n)$\ndeterministic time, and finds the $\\mathrm{occ}$ pattern occurrences in time\n$O(m/\\log n + \\sqrt{\\log n}\\log\\log n + \\mathrm{occ})$, where $\\varepsilon>0$\nis an arbitrarily small constant. As a comparison, the most recent classical\ntext index uses $O(n\\log n)$ bits, is built in $O(n)$ time, and searches in\ntime $O(m/\\log n + \\log\\log n + \\mathrm{occ})$. We build on a novel text\nsampling based on difference covers, which enjoys properties that allow us\nefficiently computing longest common prefixes in constant time. We extend our\nresults to the secondary memory model as well, where we give the first\nconstruction in $o(Sort(n))$ time of a data structure with suffix array\nfunctionality, which can search for patterns in the almost optimal time, with\nan additive penalty of $O(\\sqrt{\\log_{M/B} n}\\log\\log n)$, where $M$ is the\nsize of main memory available and $B$ is the disk block size.\n",
"title": "Text Indexing and Searching in Sublinear Time"
}
| null | null | null | null | true | null |
4064
| null |
Default
| null | null |
null |
{
"abstract": " We study the temperature dependence of the Rashba-split bands in the bismuth\ntellurohalides BiTe$X$ $(X=$ I, Br, Cl) from first principles. We find that\nincreasing temperature reduces the Rashba splitting, with the largest effect\nobserved in BiTeI with a reduction of the Rashba parameter of $40$% when\ntemperature increases from $0$ K to $300$ K. These results highlight the\ninadequacy of previous interpretations of the observed Rashba splitting in\nterms of static-lattice calculations alone. Notably, we find the opposite\ntrend, a strengthening of the Rashba splitting with rising temperature, in the\npressure-stabilized topological-insulator phase of BiTeI. We propose that the\nopposite trends with temperature on either side of the topological phase\ntransition could be an experimental signature for identifying it. The predicted\ntemperature dependence is consistent with optical conductivity measurements,\nand should also be observable using photoemission spectroscopy, which could\nprovide further insights into the nature of spin splitting and topology in the\nbismuth tellurohalides.\n",
"title": "Temperature dependence of the bulk Rashba splitting in the bismuth tellurohalides"
}
| null | null | null | null | true | null |
4065
| null |
Default
| null | null |
null |
{
"abstract": " We give a definition of viscosity solution for the minimal surface system and\nprove a version of Allard regularity theorem in this setting.\n",
"title": "Viscosity solutions and the minimal surface system"
}
| null | null | null | null | true | null |
4066
| null |
Default
| null | null |
null |
{
"abstract": " The recently introduced acoustic ray-tracing semiclassical (RTS) method is\nvalidated for a set of practically relevant boundary conditions. RTS is a\nfrequency domain geometrical method which directly reproduces the acoustic\nGreen's function. As previously demonstrated for a rectangular room and weakly\nabsorbing boundaries with a real and frequency-independent impedance, RTS is\ncapable of modeling also the lowest modes of such a room, which makes it a\nuseful method for low frequency sound field modeling in enclosures. In\npractice, rooms are furnished with diverse types of materials and acoustic\nelements, resulting in a frequency-dependent, phase-modifying\nabsorption/reflection. In a realistic setting, we test the RTS method with two\nadditional boundary conditions: a local reaction boundary simulating a\nresonating membrane absorber and an extended reaction boundary representing a\nporous layer backed by a rigid boundary described within the Delany-Bazley-Miki\nmodel, as well as a combination thereof. The RTS-modeled spatially dependent\npressure response and octave band decay curves with the corresponding\nreverberation times are compared to those obtained by the finite element\nmethod.\n",
"title": "Ray-tracing semiclassical low frequency acoustic modeling with local and extended reaction boundaries"
}
| null | null | null | null | true | null |
4067
| null |
Default
| null | null |
null |
{
"abstract": " The World Wide Web conference is a well-established and mature venue with an\nalready long history. Over the years it has been attracting papers reporting\nmany important research achievements centered around the Web. In this work we\naim at understanding the evolution of WWW conference series by detecting\ncrucial years and important topics. We propose a simple yet novel approach\nbased on tracking the classification errors of the conference papers according\nto their predicted publication years.\n",
"title": "Towards Understanding the Evolution of the WWW Conference"
}
| null | null |
[
"Computer Science"
] | null | true | null |
4068
| null |
Validated
| null | null |
null |
{
"abstract": " In this semi-tutorial paper, we first review the information-theoretic\napproach to account for the computational costs incurred during the search for\noptimal actions in a sequential decision-making problem. The traditional (MDP)\nframework ignores computational limitations while searching for optimal\npolicies, essentially assuming that the acting agent is perfectly rational and\naims for exact optimality. Using the free-energy, a variational principle is\nintroduced that accounts not only for the value of a policy alone, but also\nconsiders the cost of finding this optimal policy. The solution of the\nvariational equations arising from this formulation can be obtained using\nfamiliar Bellman-like value iterations from dynamic programming (DP) and the\nBlahut-Arimoto (BA) algorithm from rate distortion theory. Finally, we\ndemonstrate the utility of the approach for generating hierarchies of state\nabstractions that can be used to best exploit the available computational\nresources. A numerical example showcases these concepts for a path-planning\nproblem in a grid world environment.\n",
"title": "Hierarchical State Abstractions for Decision-Making Problems with Computational Constraints"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
4069
| null |
Validated
| null | null |
null |
{
"abstract": " We consider the generalized Milne problem in non-conservative plane-parallel\noptically thick atmosphere consisting of two components - the free electrons\nand small dust particles. Recall, that the traditional Milne problem describes\nthe propagation of radiation through the conservative (without absorption)\noptically thick atmosphere when the source of thermal radiation located far\nbelow the surface. In such case, the flux of propagating light is the same at\nevery distance in an atmosphere. In the generalized Milne problem, the flux\nchanges inside the atmosphere. The solutions of the both Milne problems give\nthe angular distribution and polarization degree of emerging radiation. The\nconsidered problem depends on two dimensionless parameters W and (a+b), which\ndepend on three parameters: $\\eta$ - the ratio of optical depth due to free\nelectrons to optical depth due to small dust grains; the absorption factor\n$\\varepsilon$ of dust grains and two coefficients - $\\bar b_1$ and $\\bar b_2$,\ndescribing the averaged anisotropic dust grains. These coefficients obey the\nrelation $\\bar b_1+3\\bar b_2=1$. The goal of the paper is to study the\ndependence of the radiation angular distribution and degree of polarization of\nemerging light on these parameters. Here we consider only continuum radiation.\n",
"title": "The generalized Milne problem in gas-dusty atmosphere"
}
| null | null | null | null | true | null |
4070
| null |
Default
| null | null |
null |
{
"abstract": " In this work we exploit agglomeration based $h$-multigrid preconditioners to\nspeed-up the iterative solution of discontinuous Galerkin discretizations of\nthe Stokes and Navier-Stokes equations. As a distinctive feature $h$-coarsened\nmesh sequences are generated by recursive agglomeration of a fine grid,\nadmitting arbitrarily unstructured grids of complex domains, and agglomeration\nbased discontinuous Galerkin discretizations are employed to deal with\nagglomerated elements of coarse levels. Both the expense of building coarse\ngrid operators and the performance of the resulting multigrid iteration are\ninvestigated. For the sake of efficiency coarse grid operators are inherited\nthrough element-by-element $L^2$ projections, avoiding the cost of numerical\nintegration over agglomerated elements. Specific care is devoted to the\nprojection of viscous terms discretized by means of the BR2 dG method. We\ndemonstrate that enforcing the correct amount of stabilization on coarse grids\nlevels is mandatory for achieving uniform convergence with respect to the\nnumber of levels. The numerical solution of steady and unsteady, linear and\nnon-linear problems is considered tackling challenging 2D test cases and 3D\nreal life computations on parallel architectures. Significant execution time\ngains are documented.\n",
"title": "h-multigrid agglomeration based solution strategies for discontinuous Galerkin discretizations of incompressible flow problems"
}
| null | null | null | null | true | null |
4071
| null |
Default
| null | null |
null |
{
"abstract": " We study how to detect clusters in a graph defined by a stream of edges,\nwithout storing the entire graph. We extend the approach to dynamic graphs\ndefined by the most recent edges of the stream and to several streams. The {\\em\ncontent correlation }of two streams $\\rho(t)$ is the Jaccard similarity of\ntheir clusters in the windows before time $t$. We propose a simple and\nefficient method to approximate this correlation online and show that for\ndynamic random graphs which follow a power law degree distribution, we can\nguarantee a good approximation. As an application, we follow Twitter streams\nand compute their content correlations online. We then propose a {\\em search by\ncorrelation} where answers to sets of keywords are entirely based on the small\ncorrelations of the streams. Answers are ordered by the correlations, and\nexplanations can be traced with the stored clusters.\n",
"title": "The content correlation of multiple streaming edges"
}
| null | null |
[
"Computer Science"
] | null | true | null |
4072
| null |
Validated
| null | null |
null |
{
"abstract": " We construct fundamental solutions of second-order parabolic systems of\ndivergence form with bounded and measurable leading coefficients and divergence\nfree first-order coefficients in the class of $BMO^{-1}_x$, under the\nassumption that weak solutions of the system satisfy a certain local\nboundedness estimate. We also establish Gaussian upper bound for such\nfundamental solutions under the same conditions.\n",
"title": "Fundamental solutions for second order parabolic systems with drift terms"
}
| null | null |
[
"Mathematics"
] | null | true | null |
4073
| null |
Validated
| null | null |
null |
{
"abstract": " GAUGE INVARIANCE: The Sachs-Wolfe formula describing the Cosmic Microwave\nBackground (CMB) temperature anisotropies is one of the most important\nrelations in cosmology. Despite its importance, the gauge invariance of this\nformula has only been discussed at first order. Here we discuss the subtle\nissue of second-order gauge transformations on the CMB. By introducing two\nrules (needed to handle the subtle issues), we prove the gauge invariance of\nthe second-order Sachs-Wolfe formula and provide several compact expressions\nwhich can be useful for the study of gauge transformations on cosmology. Our\nresults go beyond a simple technicality: we discuss from a physical point of\nview several aspects that improve our understanding of the CMB. We also\nelucidate how crucial it is to understand gauge transformations on the CMB in\norder to avoid errors and/or misconceptions as occurred in the past. THE RIVER\nFRAME: we introduce a cosmological frame which we call the river frame. In this\nframe, photons and any object can be thought as fishes swimming in the river\nand relations are easily expressed in either the metric or the covariant\nformalism then ensuring a transparent geometric meaning. Finally, our results\nshow that the river frame is useful to make perturbative and non-perturbative\nanalysis. In particular, it was already used to obtain the fully nonlinear\ngeneralization of the Sachs-Wolfe formula and is used here to describe\nsecond-order perturbations.\n",
"title": "CMB in the river frame and gauge invariance at second order"
}
| null | null |
[
"Physics"
] | null | true | null |
4074
| null |
Validated
| null | null |
null |
{
"abstract": " The noisy matrix completion problem, which aims to recover a low-rank matrix\n$\\mathbf{X}$ from a partial, noisy observation of its entries, arises in many\nstatistical, machine learning, and engineering applications. In this paper, we\npresent a new, information-theoretic approach for active sampling (or\ndesigning) of matrix entries for noisy matrix completion, based on the maximum\nentropy design principle. One novelty of our method is that it implicitly makes\nuse of uncertainty quantification (UQ) -- a measure of uncertainty for\nunobserved matrix entries -- to guide the active sampling procedure. The\nproposed framework reveals several novel insights on the role of compressive\nsensing (e.g., coherence) and coding design (e.g., Latin squares) on the\nsampling performance and UQ for noisy matrix completion. Using such insights,\nwe develop an efficient posterior sampler for UQ, which is then used to guide a\nclosed-form sampling scheme for matrix entries. Finally, we illustrate the\neffectiveness of this integrated sampling / UQ methodology in simulation\nstudies and two applications to collaborative filtering.\n",
"title": "Active matrix completion with uncertainty quantification"
}
| null | null | null | null | true | null |
4075
| null |
Default
| null | null |
null |
{
"abstract": " For $Q$ a polynomial with integer coefficients and $x, y \\geq 2$, we prove\nupper bounds for the quantity $\\Psi_Q(x, y) = |\\{n\\leq x: p\\mid Q(n)\\Rightarrow\np\\leq y\\}|$.\nWe apply our results to a problem of De Koninck, Doyon and Luca on integers\ndivisible by the square of their largest prime factor. As a corollary to our\narguments, we improve the known level of distribution of the set $\\{n^2-D\\}$\nfor well-factorable moduli, previously due to Iwaniec. We also consider the\nChebyshev problem of estimating $\\max\\{P^+(n^2-D), n\\leq x\\}$ and make\nexplicit, in Deshouillers-Iwaniec's state-of-the-art result, the dependence on\nthe Selberg eigenvalue conjecture.\n",
"title": "Majoration du nombre de valeurs friables d'un polynôme"
}
| null | null | null | null | true | null |
4076
| null |
Default
| null | null |
null |
{
"abstract": " Implementing the modal method in the electromagnetic grating diffraction\nproblem delivered by the curvilinear coordinate transformation yields a general\nanalytical solution to the 1D grating diffraction problem in a form of a\nT-matrix. Simultaneously it is shown that the validity of the Rayleigh\nexpansion is defined by the validity of the modal expansion in a transformed\nmedium delivered by the coordinate transformation.\n",
"title": "General analytical solution for the electromagnetic grating diffraction problem"
}
| null | null | null | null | true | null |
4077
| null |
Default
| null | null |
null |
{
"abstract": " We give an abstract formulation of the formal theory partial differential\nequations (PDEs) in synthetic differential geometry, one that would seamlessly\ngeneralize the traditional theory to a range of enhanced contexts, such as\nsuper-geometry, higher (stacky) differential geometry, or even a combination of\nboth. A motivation for such a level of generality is the eventual goal of\nsolving the open problem of covariant geometric pre-quantization of locally\nvariational field theories, which may include fermions and (higher) gauge\nfields. (abridged)\n",
"title": "Synthetic geometry of differential equations: I. Jets and comonad structure"
}
| null | null | null | null | true | null |
4078
| null |
Default
| null | null |
null |
{
"abstract": " The distribution of matter in the universe is, to first order, lognormal.\nImproving this approximation requires characterization of the third moment\n(skewness) of the log density field. Thus, using Millennium Simulation\nphenomenology and building on previous work, we present analytic fits for the\nmean, variance, and skewness of the log density field $A$. We further show that\na Generalized Extreme Value (GEV) distribution accurately models $A$; we submit\nthat this GEV behavior is the result of strong intrapixel correlations, without\nwhich the smoothed distribution would tend (by the Central Limit Theorem)\ntoward a Gaussian. Our GEV model yields cumulative distribution functions\naccurate to within 1.7 per cent for near-concordance cosmologies, over a range\nof redshifts and smoothing scales.\n",
"title": "Precision Prediction for the Cosmological Density Distribution"
}
| null | null | null | null | true | null |
4079
| null |
Default
| null | null |
null |
{
"abstract": " Workhorse theories throughout all of physics derive effective Hamiltonians to\ndescribe slow time evolution, even though low-frequency modes are actually\ncoupled to high-frequency modes. Such effective Hamiltonians are accurate\nbecause of \\textit{adiabatic decoupling}: the high-frequency modes `dress' the\nlow-frequency modes, and renormalize their Hamiltonian, but they do not\nsteadily inject energy into the low-frequency sector. Here, however, we\nidentify a broad class of dynamical systems in which adiabatic decoupling fails\nto hold, and steady energy transfer across a large gap in natural frequency\n(`steady downconversion') instead becomes possible, through nonlinear\nresonances of a certain form. Instead of adiabatic decoupling, the special\nfeatures of multiple time scale dynamics lead in these cases to efficiency\nconstraints that somewhat resemble thermodynamics.\n",
"title": "Hamiltonian analogs of combustion engines: a systematic exception to adiabatic decoupling"
}
| null | null | null | null | true | null |
4080
| null |
Default
| null | null |
null |
{
"abstract": " Accurate noise modelling is important for training of deep learning\nreconstruction algorithms. While noise models are well known for traditional\nimaging techniques, the noise distribution of a novel sensor may be difficult\nto determine a priori. Therefore, we propose learning arbitrary noise\ndistributions. To do so, this paper proposes a fully connected neural network\nmodel to map samples from a uniform distribution to samples of any explicitly\nknown probability density function. During the training, the Jensen-Shannon\ndivergence between the distribution of the model's output and the target\ndistribution is minimized. We experimentally demonstrate that our model\nconverges towards the desired state. It provides an alternative to existing\nsampling methods such as inversion sampling, rejection sampling, Gaussian\nmixture models and Markov-Chain-Monte-Carlo. Our model has high sampling\nefficiency and is easily applied to any probability distribution, without the\nneed of further analytical or numerical calculations.\n",
"title": "Towards Arbitrary Noise Augmentation - Deep Learning for Sampling from Arbitrary Probability Distributions"
}
| null | null | null | null | true | null |
4081
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we introduce an unbiased gradient simulation algorithms for\nsolving convex optimization problem with stochastic function compositions. We\nshow that the unbiased gradient generated from the algorithm has finite\nvariance and finite expected computation cost. We then combined the unbiased\ngradient simulation with two variance reduced algorithms (namely SVRG and SCSG)\nand showed that the proposed optimization algorithms based on unbiased gradient\nsimulations exhibit satisfactory convergence properties. Specifically, in the\nSVRG case, the algorithm with simulated gradient can be shown to converge\nlinearly to optima in expectation and almost surely under strong convexity.\nFinally, for the numerical experiment,we applied the algorithms to two\nimportant cases of stochastic function compositions optimization: maximizing\nthe Cox's partial likelihood model and training conditional random fields.\n",
"title": "Unbiased Simulation for Optimizing Stochastic Function Compositions"
}
| null | null | null | null | true | null |
4082
| null |
Default
| null | null |
null |
{
"abstract": " A robot's ability to understand or ground natural language instructions is\nfundamentally tied to its knowledge about the surrounding world. We present an\napproach to grounding natural language utterances in the context of factual\ninformation gathered through natural-language interactions and past visual\nobservations. A probabilistic model estimates, from a natural language\nutterance, the objects,relations, and actions that the utterance refers to, the\nobjectives for future robotic actions it implies, and generates a plan to\nexecute those actions while updating a state representation to include newly\nacquired knowledge from the visual-linguistic context. Grounding a command\nnecessitates a representation for past observations and interactions; however,\nmaintaining the full context consisting of all possible observed objects,\nattributes, spatial relations, actions, etc., over time is intractable.\nInstead, our model, Temporal Grounding Graphs, maintains a learned state\nrepresentation for a belief over factual groundings, those derived from\nnatural-language interactions, and lazily infers new groundings from visual\nobservations using the context implied by the utterance. This work\nsignificantly expands the range of language that a robot can understand by\nincorporating factual knowledge and observations of its workspace in its\ninference about the meaning and grounding of natural-language utterances.\n",
"title": "Temporal Grounding Graphs for Language Understanding with Accrued Visual-Linguistic Context"
}
| null | null | null | null | true | null |
4083
| null |
Default
| null | null |
null |
{
"abstract": " The growing demand on efficient and distributed optimization algorithms for\nlarge-scale data stimulates the popularity of Alternative Direction Methods of\nMultipliers (ADMM) in numerous areas, such as compressive sensing, matrix\ncompletion, and sparse feature learning. While linear equality constrained\nproblems have been extensively explored to be solved by ADMM, there lacks a\ngeneric framework for ADMM to solve problems with nonlinear equality\nconstraints, which are common in practical application (e.g., orthogonality\nconstraints). To address this problem, in this paper, we proposed a new generic\nADMM framework for handling nonlinear equality constraints, called neADMM.\nFirst, we propose the generalized problem formulation and systematically\nprovide the sufficient condition for the convergence of neADMM. Second, we\nprove a sublinear convergence rate based on variational inequality framework\nand also provide an novel accelerated strategy on the update of the penalty\nparameter. In addition, several practical applications under the generic\nframework of neADMM are provided. Experimental results on several applications\ndemonstrate the usefulness of our neADMM.\n",
"title": "Nonconvex generalizations of ADMM for nonlinear equality constrained problems"
}
| null | null |
[
"Computer Science",
"Mathematics"
] | null | true | null |
4084
| null |
Validated
| null | null |
null |
{
"abstract": " The analysis of neuroimaging data poses several strong challenges, in\nparticular, due to its high dimensionality, its strong spatio-temporal\ncorrelation and the comparably small sample sizes of the respective datasets.\nTo address these challenges, conventional decoding approaches such as the\nsearchlight reduce the complexity of the decoding problem by considering local\nclusters of voxels only. Thereby, neglecting the distributed spatial patterns\nof brain activity underlying many cognitive states. In this work, we introduce\nthe DLight framework, which overcomes these challenges by utilizing a long\nshort-term memory unit (LSTM) based deep neural network architecture to analyze\nthe spatial dependency structure of whole-brain fMRI data. In order to maintain\ninterpretability of the neuroimaging data, we adapt the layer-wise relevance\npropagation (LRP) method. Thereby, we enable the neuroscientist user to study\nthe learned association of the LSTM between the data and the cognitive state of\nthe individual. We demonstrate the versatility of DLight by applying it to a\nlarge fMRI dataset of the Human Connectome Project. We show that the decoding\nperformance of our method scales better with large datasets, and moreover\noutperforms conventional decoding approaches, while still detecting\nphysiologically appropriate brain areas for the cognitive states classified. We\nalso demonstrate that DLight is able to detect these areas on several levels of\ndata granularity (i.e., group, subject, trial, time point).\n",
"title": "Interpretable LSTMs For Whole-Brain Neuroimaging Analyses"
}
| null | null | null | null | true | null |
4085
| null |
Default
| null | null |
null |
{
"abstract": " Wet etching is an essential and complex step in semiconductor device\nprocessing. Metal-Assisted Chemical Etching (MacEtch) is fundamentally a wet\nbut anisotropic etching method. In the MacEtch technique, there are still a\nnumber of unresolved challenges preventing the optimal fabrication of\nhigh-aspect-ratio semiconductor micro- and nanostructures, such as undesired\netching, uncontrolled catalyst movement, non-uniformity and micro-porosity in\nthe metal-free areas. Here, an optimized MacEtch process using with a\nnanostructured Au catalyst is proposed for fabrication of Si high aspect ratio\nmicrostructures. The addition of isopropanol as surfactant in the HF-H2O2 water\nsolution improves the uniformity and the control of the H2 gas release. An\nadditional KOH etching removes eventually the unwanted nanowires left by the\nMacEtch through the nanoporous catalyst film. We demonstrate the benefits of\nthe isopropanol addition for reducing the etching rate and the nanoporosity of\netched structures with a monothonical decrease as a function of the isopropanol\nconcentration.\n",
"title": "Effect of Isopropanol on Gold Assisted Chemical Etching of Silicon Microstructures"
}
| null | null | null | null | true | null |
4086
| null |
Default
| null | null |
null |
{
"abstract": " An inherently abstract nature of source code makes programs difficult to\nunderstand. In our research, we designed three techniques utilizing concrete\nvalues of variables and other expressions during program execution.\nRuntimeSearch is a debugger extension searching for a given string in all\nexpressions at runtime. DynamiDoc generates documentation sentences containing\nexamples of arguments, return values and state changes. RuntimeSamp augments\nsource code lines in the IDE (integrated development environment) with sample\nvariable values. In this post-doctoral article, we briefly describe these three\napproaches and related motivational studies, surveys and evaluations. We also\nreflect on the PhD study, providing advice for current students. Finally,\nshort-term and long-term future work is described.\n",
"title": "Integrating Runtime Values with Source Code to Facilitate Program Comprehension"
}
| null | null | null | null | true | null |
4087
| null |
Default
| null | null |
null |
{
"abstract": " Self-nested trees present a systematic form of redundancy in their subtrees\nand thus achieve optimal compression rates by DAG compression. A method for\nquantifying the degree of self-similarity of plants through self-nested trees\nhas been introduced by Godin and Ferraro in 2010. The procedure consists in\ncomputing a self-nested approximation, called the nearest embedding self-nested\ntree, that both embeds the plant and is the closest to it. In this paper, we\npropose a new algorithm that computes the nearest embedding self-nested tree\nwith a smaller overall complexity, but also the nearest embedded self-nested\ntree. We show from simulations that the latter is mostly the closest to the\ninitial data, which suggests that this better approximation should be used as a\nprivileged measure of the degree of self-similarity of plants.\n",
"title": "Nearest Embedded and Embedding Self-Nested Trees"
}
| null | null | null | null | true | null |
4088
| null |
Default
| null | null |
null |
{
"abstract": " Let $G$ be a connected complex reductive algebraic group with Lie algebra\n$\\mathfrak{g}$. The Lusztig--Vogan bijection relates two bases for the bounded\nderived category of $G$-equivariant coherent sheaves on the nilpotent cone\n$\\mathcal{N}$ of $\\mathfrak{g}$. One basis is indexed by $\\Lambda^+$, the set\nof dominant weights of $G$, and the other by $\\Omega$, the set of pairs\n$(\\mathcal{O}, \\mathcal{E})$ consisting of a nilpotent orbit $\\mathcal{O}\n\\subset \\mathcal{N}$ and an irreducible $G$-equivariant vector bundle\n$\\mathcal{E} \\rightarrow \\mathcal{O}$. The existence of the Lusztig--Vogan\nbijection $\\gamma \\colon \\Omega \\rightarrow \\Lambda^+$ was proven by\nBezrukavnikov, and an algorithm computing $\\gamma$ in type $A$ was given by\nAchar. Herein we present a combinatorial description of $\\gamma$ in type $A$\nthat subsumes and dramatically simplifies Achar's algorithm.\n",
"title": "Computing the Lusztig--Vogan Bijection"
}
| null | null | null | null | true | null |
4089
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we propose an improvement for flowpipe-construction-based\nreachability analysis techniques for hybrid systems. Such methods apply\niterative successor computations to pave the reachable region of the state\nspace by state sets in an over-approximative manner. As the computational costs\nsteeply increase with the dimension, in this work we analyse the possibilities\nfor improving scalability by dividing the search space in sub-spaces and\nexecute reachability computations in the sub-spaces instead of the global\nspace. We formalise such an algorithm and provide experimental evaluations to\ncompare the efficiency as well as the precision of our sub-space search to the\noriginal search in the global space.\n",
"title": "Divide and Conquer: Variable Set Separation in Hybrid Systems Reachability Analysis"
}
| null | null | null | null | true | null |
4090
| null |
Default
| null | null |
null |
{
"abstract": " We give a necessary and sufficient condition for the maximum principle of\nSchrödinger operators in terms of the bottom of the spectrum of\ntime-changed processes. As a corollary, we obtain a sufficient condition for\nthe Liouville property of Schrödinger operators.\n",
"title": "The bottom of the spectrum of time-changed processes and the maximum principle of Schrödinger operators"
}
| null | null | null | null | true | null |
4091
| null |
Default
| null | null |
null |
{
"abstract": " In modern stream cipher, there are many algorithms, such as ZUC, LTE\nencryption algorithm and LTE integrity algorithm, using bit-component sequences\nof $p$-ary $m$-sequences as the input of the algorithm. Therefore, analyzing\ntheir statistical property (For example, autocorrelation, linear complexity and\n2-adic complexity) of bit-component sequences of $p$-ary $m$-sequences is\nbecoming an important research topic. In this paper, we first derive some\nautocorrelation properties of LSB (Least Significant Bit) sequences of $p$-ary\n$m$-sequences, i.e., we convert the problem of computing autocorrelations of\nLSB sequences of period $p^n-1$ for any positive $n\\geq2$ to the problem of\ndetermining autocorrelations of LSB sequence of period $p-1$. Then, based on\nthis property and computer calculation, we list some autocorrelation\ndistributions of LSB sequences of $p$-ary $m$-sequences with order $n$ for some\nsmall primes $p$'s, such as $p=3,5,7,11,17,31$. Additionally, using their\nautocorrelation distributions and the method inspired by Hu, we give the lower\nbounds on the 2-adic complexities of these LSB sequences. Our results show that\nthe main parts of all the lower bounds on the 2-adic complexity of these LSB\nsequencesare larger than $\\frac{N}{2}$, where $N$ is the period of these\nsequences. Therefor, these bounds are large enough to resist the analysis of\nRAA (Rational Approximation Algorithm) for FCSR (Feedback with Carry Shift\nRegister). Especially, for a Mersenne prime $p=2^k-1$, since all its\nbit-component sequences of a $p$-ary $m$-sequence are shift equivalent, our\nresults hold for all its bit-component sequences.\n",
"title": "Autocorrelation and Lower Bound on the 2-Adic Complexity of LSB Sequence of $p$-ary $m$-Sequence"
}
| null | null | null | null | true | null |
4092
| null |
Default
| null | null |
null |
{
"abstract": " The telecommunications industry is highly competitive, which means that the\nmobile providers need a business intelligence model that can be used to achieve\nan optimal level of churners, as well as a minimal level of cost in marketing\nactivities. Machine learning applications can be used to provide guidance on\nmarketing strategies. Furthermore, data mining techniques can be used in the\nprocess of customer segmentation. The purpose of this paper is to provide a\ndetailed analysis of the C.5 algorithm, within naive Bayesian modelling for the\ntask of segmenting telecommunication customers behavioural profiling according\nto their billing and socio-demographic aspects. Results have been\nexperimentally implemented.\n",
"title": "Integration of Machine Learning Techniques to Evaluate Dynamic Customer Segmentation Analysis for Mobile Customers"
}
| null | null | null | null | true | null |
4093
| null |
Default
| null | null |
null |
{
"abstract": " A vital aspect in energy storage planning and operation is to accurately\nmodel its operational cost, which mainly comes from the battery cell\ndegradation. Battery degradation can be viewed as a complex material fatigue\nprocess that based on stress cycles. Rainflow algorithm is a popular way for\ncycle identification in material fatigue process, and has been extensively used\nin battery degradation assessment. However, the rainflow algorithm does not\nhave a closed form, which makes the major difficulty to include it in\noptimization. In this paper, we prove the rainflow cycle-based cost is convex.\nConvexity enables the proposed degradation model to be incorporated in\ndifferent battery optimization problems and guarantees the solution quality. We\nprovide a subgradient algorithm to solve the problem. A case study on PJM\nregulation market demonstrates the effectiveness of the proposed degradation\nmodel in maximizing the battery operating profits as well as extending its\nlifetime.\n",
"title": "A Convex Cycle-based Degradation Model for Battery Energy Storage Planning and Operation"
}
| null | null | null | null | true | null |
4094
| null |
Default
| null | null |
null |
{
"abstract": " We present results of an experiment showing the first successful\ndemonstration of a cascaded micro-bunching scheme. Two modulator-chicane\npre-bunchers arranged in series and a high power mid-IR laser seed are used to\nmodulate a 52 MeV electron beam into a train of sharp microbunches phase-locked\nto the external drive laser. This configuration allows to increase the fraction\nof electrons trapped in a strongly tapered inverse free electron laser (IFEL)\nundulator to 96\\%, with up to 78\\% of the particles accelerated to the final\ndesign energy yielding a significant improvement compared to the classical\nsingle buncher scheme. These results represent a critical advance in\nlaser-based longitudinal phase space manipulations and find application both in\nhigh gradient advanced acceleration as well as in high peak and average power\ncoherent radiation sources.\n",
"title": "Demonstration of cascaded modulator-chicane micro-bunching of a relativistic electron beam"
}
| null | null | null | null | true | null |
4095
| null |
Default
| null | null |
null |
{
"abstract": " We show that fundamental groups of compact, orientable, irreducible\n3-manifolds with toroidal boundary are Grothendieck rigid.\n",
"title": "Grothendieck rigidity of 3-manifold groups"
}
| null | null | null | null | true | null |
4096
| null |
Default
| null | null |
null |
{
"abstract": " The pseudoscalars in Garret Sobczyk's paper \\emph{Simplicial Calculus with\nGeometric Algebra} are not well defined. Therefore his calculus does not have a\nproper foundation.\n",
"title": "Sobczyk's simplicial calculus does not have a proper foundation"
}
| null | null |
[
"Mathematics"
] | null | true | null |
4097
| null |
Validated
| null | null |
null |
{
"abstract": " We determine the information scrambling rate $\\lambda_{L}$ due to\nelectron-electron Coulomb interaction in graphene. $\\lambda_{L}$ characterizes\nthe growth of chaos and has been argued to give information about the\nthermalization and hydrodynamic transport coefficients of a many-body system.\nWe demonstrate that $\\lambda_{L}$ behaves for strong coupling similar to\ntransport and energy relaxation rates. A weak coupling analysis, however,\nreveals that scrambling is related to dephasing or single particle relaxation.\nFurthermore, $\\lambda_{L}$ is found to be parametrically larger than the\ncollision rate relevant for hydrodynamic processes, such as electrical\nconduction or viscous flow, and the rate of energy relaxation, relevant for\nthermalization. Thus, while scrambling is obviously necessary for\nthermalization and quantum transport, it does generically not set the time\nscale for these processes. In addition we derive a quantum kinetic theory for\ninformation scrambling that resembles the celebrated Boltzmann equation and\noffers a physically transparent insight into quantum chaos in many-body\nsystems.\n",
"title": "Hierarchy of Information Scrambling, Thermalization, and Hydrodynamic Flow in Graphene"
}
| null | null | null | null | true | null |
4098
| null |
Default
| null | null |
null |
{
"abstract": " In recent years, deep neural networks have yielded state-of-the-art\nperformance on several tasks. Although some recent works have focused on\ncombining deep learning with recommendation, we highlight three issues of\nexisting models. First, these models cannot work on both explicit and implicit\nfeedback, since the network structures are specially designed for one\nparticular case. Second, due to the difficulty on training deep neural\nnetworks, existing explicit models do not fully exploit the expressive\npotential of deep learning. Third, neural network models are easier to overfit\non the implicit setting than shallow models. To tackle these issues, we present\na generic recommender framework called Neural Collaborative Autoencoder (NCAE)\nto perform collaborative filtering, which works well for both explicit feedback\nand implicit feedback. NCAE can effectively capture the subtle hidden\nrelationships between interactions via a non-linear matrix factorization\nprocess. To optimize the deep architecture of NCAE, we develop a three-stage\npre-training mechanism that combines supervised and unsupervised feature\nlearning. Moreover, to prevent overfitting on the implicit setting, we propose\nan error reweighting module and a sparsity-aware data-augmentation strategy.\nExtensive experiments on three real-world datasets demonstrate that NCAE can\nsignificantly advance the state-of-the-art.\n",
"title": "Neural Collaborative Autoencoder"
}
| null | null | null | null | true | null |
4099
| null |
Default
| null | null |
null |
{
"abstract": " The influence of the surface curvature on the surface tension of small\ndroplets in equilibrium with a surrounding vapour, or small bubbles in\nequilibrium with a surrounding liquid, can be expanded as $\\gamma(R) = \\gamma_0\n+ c_1\\gamma_0/R + O(1/R^2)$, where $R = R_\\gamma$ is the radius of the surface\nof tension and $\\gamma_0$ is the surface tension of the planar interface,\ncorresponding to zero curvature. According to Tolman's law, the first-order\ncoefficient in this expansion is assumed to be related to the planar limit\n$\\delta_0$ of the Tolman length, i.e., the difference $\\delta = R_\\rho -\nR_\\gamma$ between the equimolar radius and the radius of the surface of\ntension, by $c_1 = -2\\delta_0$.\nWe show here that the deduction of Tolman's law from interfacial\nthermodynamics relies on an inaccurate application of the Gibbs adsorption\nequation to dispersed phases (droplets or bubbles). A revision of the\nunderlying theory reveals that the adsorption equation needs to be employed in\nan alternative manner to that suggested by Tolman. Accordingly, we develop a\ngeneralized Gibbs adsorption equation which consistently takes the size\ndependence of interfacial properties into account, and show that from this\nequation, a relation between the Tolman length and the influence of the size of\nthe dispersed phase on the surface tension cannot be deduced, invalidating the\nargument which was put forward by Tolman [J. Chem. Phys. 17 (1949) 333].\n",
"title": "Reexamination of Tolman's law and the Gibbs adsorption equation for curved interfaces"
}
| null | null | null | null | true | null |
4100
| null |
Default
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.