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": " Recently, Lawson has shown that the 2-primary Brown-Peterson spectrum does\nnot admit the structure of an $E_{12}$ ring spectrum, thus answering a question\nof May in the negative. We extend Lawson's result to odd primes by proving that\nthe p-primary Brown-Peterson spectrum does not admit the structure of an\n$E_{2(p^2+2)}$ ring spectrum. We also show that there can be no map $MU \\to BP$\nof $E_{2p+3}$ ring spectra at any prime.\n",
"title": "The Brown-Peterson spectrum is not $E_{2(p^2+2)}$ at odd primes"
}
| null | null | null | null | true | null |
4301
| null |
Default
| null | null |
null |
{
"abstract": " This paper is devoted to expressiveness of hypergraphs for which uncertainty\npropagation by local computations via Shenoy/Shafer method applies. It is\ndemonstrated that for this propagation method for a given joint belief\ndistribution no valuation of hyperedges of a hypergraph may provide with\nsimpler hypergraph structure than valuation of hyperedges by conditional\ndistributions. This has vital implication that methods recovering belief\nnetworks from data have no better alternative for finding the simplest\nhypergraph structure for belief propagation. A method for recovery\ntree-structured belief networks has been developed and specialized for\nDempster-Shafer belief functions\n",
"title": "Beliefs in Markov Trees - From Local Computations to Local Valuation"
}
| null | null | null | null | true | null |
4302
| null |
Default
| null | null |
null |
{
"abstract": " Tree-grass coexistence in savanna ecosystems depends strongly on\nenvironmental disturbances out of which crucial is fire. Most modeling attempts\nin the literature lack stochastic approach to fire occurrences which is\nessential to reflect their unpredictability. Existing models that actually\ninclude stochasticity of fire are usually analyzed only numerically. We\nintroduce new minimalistic model of tree-grass coexistence where fires occur\naccording to stochastic process. We use the tools of linear semigroup theory to\nprovide more careful mathematical analysis of the model. Essentially we show\nthat there exists a unique stationary distribution of tree and grass biomasses.\n",
"title": "A model for random fire induced tree-grass coexistence in savannas"
}
| null | null | null | null | true | null |
4303
| null |
Default
| null | null |
null |
{
"abstract": " An approach is presented for making predictions about functional time series.\nThe method is applied to data coming from periodically correlated processes and\nelectricity demand, obtaining accurate point forecasts and narrow prediction\nbands that cover high proportions of the forecasted functional datum, for a\ngiven confidence level. The method is computationally efficient and\nsubstantially different to other functional time series methods, offering a new\ninsight for the analysis of these data structures.\n",
"title": "A depth-based method for functional time series forecasting"
}
| null | null | null | null | true | null |
4304
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we investigate periodic vibrations of a group of particles\nwith a dihedral configuration in the plane governed by the Lennard-Jones and\nCoulomb forces. Using the gradient equivariant degree, we provide a full\ntopological classification of the periodic solutions with both temporal and\nspatial symmetries. In the process, we provide with general formulae for the\nspectrum of the linearized system which allows us to obtain the critical\nfrequencies of the particle motions which indicate the set of all critical\nperiods of small amplitude periodic solutions emerging from a given stationary\nsymmetric orbit of solutions.\n",
"title": "Dihedral Molecular Configurations Interacting by Lennard-Jones and Coulomb Forces"
}
| null | null | null | null | true | null |
4305
| null |
Default
| null | null |
null |
{
"abstract": " Content analysis of news stories (whether manual or automatic) is a\ncornerstone of the communication studies field. However, much research is\nconducted at the level of individual news articles, despite the fact that news\nevents (especially significant ones) are frequently presented as \"stories\" by\nnews outlets: chains of connected articles covering the same event from\ndifferent angles. These stories are theoretically highly important in terms of\nincreasing public recall of news items and enhancing the agenda-setting power\nof the press. Yet thus far, the field has lacked an efficient method for\ndetecting groups of articles which form stories in a way that enables their\nanalysis.\nIn this work, we present a novel, automated method for identifying linked\nnews stories from within a corpus of articles. This method makes use of\ntechniques drawn from the field of information retrieval to identify textual\ncloseness of pairs of articles, and then clustering techniques taken from the\nfield of network analysis to group these articles into stories. We demonstrate\nthe application of the method to a corpus of 61,864 articles, and show how it\ncan efficiently identify valid story clusters within the corpus. We use the\nresults to make observations about the prevalence and dynamics of stories\nwithin the UK news media, showing that more than 50% of news production takes\nplace within stories.\n",
"title": "Understanding news story chains using information retrieval and network clustering techniques"
}
| null | null | null | null | true | null |
4306
| null |
Default
| null | null |
null |
{
"abstract": " New social and economic activities massively exploit big data and machine\nlearning algorithms to do inference on people's lives. Applications include\nautomatic curricula evaluation, wage determination, and risk assessment for\ncredits and loans. Recently, many governments and institutions have raised\nconcerns about the lack of fairness, equity and ethics in machine learning to\ntreat these problems. It has been shown that not including sensitive features\nthat bias fairness, such as gender or race, is not enough to mitigate the\ndiscrimination when other related features are included. Instead, including\nfairness in the objective function has been shown to be more efficient.\nWe present novel fair regression and dimensionality reduction methods built\non a previously proposed fair classification framework. Both methods rely on\nusing the Hilbert Schmidt independence criterion as the fairness term. Unlike\nprevious approaches, this allows us to simplify the problem and to use multiple\nsensitive variables simultaneously. Replacing the linear formulation by kernel\nfunctions allows the methods to deal with nonlinear problems. For both linear\nand nonlinear formulations the solution reduces to solving simple matrix\ninversions or generalized eigenvalue problems. This simplifies the evaluation\nof the solutions for different trade-off values between the predictive error\nand fairness terms. We illustrate the usefulness of the proposed methods in toy\nexamples, and evaluate their performance on real world datasets to predict\nincome using gender and/or race discrimination as sensitive variables, and\ncontraceptive method prediction under demographic and socio-economic sensitive\ndescriptors.\n",
"title": "Fair Kernel Learning"
}
| null | null |
[
"Statistics"
] | null | true | null |
4307
| null |
Validated
| null | null |
null |
{
"abstract": " Bearing only cooperative localization has been used successfully on aerial\nand ground vehicles. In this paper we present an extension of the approach to\nthe underwater domain. The focus is on adapting the technique to handle the\nchallenging visibility conditions underwater. Furthermore, data from inertial,\nmagnetic, and depth sensors are utilized to improve the robustness of the\nestimation. In addition to robotic applications, the presented technique can be\nused for cave mapping and for marine archeology surveying, both by human\ndivers. Experimental results from different environments, including a fresh\nwater, low visibility, lake in South Carolina; a cavern in Florida; and coral\nreefs in Barbados during the day and during the night, validate the robustness\nand the accuracy of the proposed approach.\n",
"title": "Underwater Surveying via Bearing only Cooperative Localization"
}
| null | null | null | null | true | null |
4308
| null |
Default
| null | null |
null |
{
"abstract": " We revisit the relation between the neutrino masses and the spontaneous\nbreaking of the B-L gauge symmetry. We discuss the main scenarios for Dirac and\nMajorana neutrinos and point out two simple mechanisms for neutrino masses. In\nthis context the neutrino masses can be generated either at tree level or at\nquantum level and one predicts the existence of very light sterile neutrinos\nwith masses below the eV scale. The predictions for lepton number violating\nprocesses such as mu to e and mu to e gamma are discussed in detail. The impact\nfrom the cosmological constraints on the effective number of relativistic\ndegree of freedom is investigated.\n",
"title": "Sterile Neutrinos and B-L Symmetry"
}
| null | null |
[
"Physics"
] | null | true | null |
4309
| null |
Validated
| null | null |
null |
{
"abstract": " This paper is concerned with the generation of Gaussian invariant states in\ncascades of open quantum harmonic oscillators governed by linear quantum\nstochastic differential equations. We carry out infinitesimal perturbation\nanalysis of the covariance matrix for the invariant Gaussian state of such a\nsystem and the related purity functional subject to inaccuracies in the energy\nand coupling matrices of the subsystems. This leads to the problem of balancing\nthe state-space realizations of the component oscillators through symplectic\nsimilarity transformations in order to minimize the mean square sensitivity of\nthe purity functional to small random perturbations of the parameters. This\nresults in a quadratic optimization problem with an effective solution in the\ncase of cascaded one-mode oscillators, which is demonstrated by a numerical\nexample. We also discuss a connection of the sensitivity index with classical\nstatistical distances and outline infinitesimal perturbation analysis for\ntranslation invariant cascades of identical oscillators. The findings of the\npaper are applicable to robust state generation in quantum stochastic networks.\n",
"title": "Effects of parametric uncertainties in cascaded open quantum harmonic oscillators and robust generation of Gaussian invariant states"
}
| null | null | null | null | true | null |
4310
| null |
Default
| null | null |
null |
{
"abstract": " This paper proposes a computer-based recursion algorithm for automatic charge\nof power device of electric vehicles carrying electromagnet. The charging\nsystem includes charging cable with one end connecting gang socket,\nelectromagnetic gear driving the connecting socket and a charging pile breaking\nor closing, and detecting part for detecting electric vehicle static call or\nstart state. The gang socket mentioned above is linked to electromagnetic gear,\nand the detecting part is connected with charging management system containing\nthe intelligent charging power module which controls the electromagnetic drive\naction to close socket with a charging pile at static state and to break at\nstart state. Our work holds an electric automobile with convenience, safety low\nmaintenance cost.\n",
"title": "A computer-based recursion algorithm for automatic charge of power device of electric vehicles carrying electromagnet"
}
| null | null | null | null | true | null |
4311
| null |
Default
| null | null |
null |
{
"abstract": " We consider a chain of Abelian Klebanov-Tarnopolsky fermionic tensor models\ncoupled through quartic nearest-neighbor interactions. We characterize the\ngauge-singlet spectrum for small chains ($L=2,3,4,5$) and observe that the\nspectral statistics exhibits strong evidences in favor of quasi-many body\nlocalization.\n",
"title": "Abelian Tensor Models on the Lattice"
}
| null | null | null | null | true | null |
4312
| null |
Default
| null | null |
null |
{
"abstract": " The reliable measurement of confidence in classifiers' predictions is very\nimportant for many applications and is, therefore, an important part of\nclassifier design. Yet, although deep learning has received tremendous\nattention in recent years, not much progress has been made in quantifying the\nprediction confidence of neural network classifiers. Bayesian models offer a\nmathematically grounded framework to reason about model uncertainty, but\nusually come with prohibitive computational costs. In this paper we propose a\nsimple, scalable method to achieve a reliable confidence score, based on the\ndata embedding derived from the penultimate layer of the network. We\ninvestigate two ways to achieve desirable embeddings, by using either a\ndistance-based loss or Adversarial Training. We then test the benefits of our\nmethod when used for classification error prediction, weighting an ensemble of\nclassifiers, and novelty detection. In all tasks we show significant\nimprovement over traditional, commonly used confidence scores.\n",
"title": "Distance-based Confidence Score for Neural Network Classifiers"
}
| null | null | null | null | true | null |
4313
| null |
Default
| null | null |
null |
{
"abstract": " Benford's law is an empirical edict stating that the lower digits appear more\noften than higher ones as the first few significant digits in statistics of\nnatural phenomena and mathematical tables. A marked proportion of such analyses\nis restricted to the first significant digit. We employ violation of Benford's\nlaw, up to the first four significant digits, for investigating magnetization\nand correlation data of paradigmatic quantum many-body systems to detect\ncooperative phenomena, focusing on the finite-size scaling exponents thereof.\nWe find that for the transverse field quantum XY model, behavior of the very\nfirst significant digit of an observable, at an arbitrary point of the\nparameter space, is enough to capture the quantum phase transition in the model\nwith a relatively high scaling exponent. A higher number of significant digits\ndo not provide an appreciable further advantage, in particular, in terms of an\nincrease in scaling exponents. Since the first significant digit of a physical\nquantity is relatively simple to obtain in experiments, the results have\npotential implications for laboratory observations in noisy environments.\n",
"title": "Benford analysis of quantum critical phenomena: First digit provides high finite-size scaling exponent while first two and further are not much better"
}
| null | null | null | null | true | null |
4314
| null |
Default
| null | null |
null |
{
"abstract": " Two popular classes of methods for approximate inference are Markov chain\nMonte Carlo (MCMC) and variational inference. MCMC tends to be accurate if run\nfor a long enough time, while variational inference tends to give better\napproximations at shorter time horizons. However, the amount of time needed for\nMCMC to exceed the performance of variational methods can be quite high,\nmotivating more fine-grained tradeoffs. This paper derives a distribution over\nvariational parameters, designed to minimize a bound on the divergence between\nthe resulting marginal distribution and the target, and gives an example of how\nto sample from this distribution in a way that interpolates between the\nbehavior of existing methods based on Langevin dynamics and stochastic gradient\nvariational inference (SGVI).\n",
"title": "A Divergence Bound for Hybrids of MCMC and Variational Inference and an Application to Langevin Dynamics and SGVI"
}
| null | null | null | null | true | null |
4315
| null |
Default
| null | null |
null |
{
"abstract": " Bell's theorem has fascinated physicists and philosophers since his 1964\npaper, which was written in response to the 1935 paper of Einstein, Podolsky,\nand Rosen. Bell's theorem and its many extensions have led to the claim that\nquantum mechanics and by inference nature herself are nonlocal in the sense\nthat a measurement on a system by an observer at one location has an immediate\neffect on a distant \"entangled\" system (one with which the original system has\npreviously interacted). Einstein was repulsed by such \"spooky action at a\ndistance\" and was led to question whether quantum mechanics could provide a\ncomplete description of physical reality. In this paper I argue that quantum\nmechanics does not require spooky action at a distance of any kind and yet it\nis entirely reasonable to question the assumption that quantum mechanics can\nprovide a complete description of physical reality. The magic of entangled\nquantum states has little to do with entanglement and everything to do with\nsuperposition, a property of all quantum systems and a foundational tenet of\nquantum mechanics.\n",
"title": "Making Sense of Bell's Theorem and Quantum Nonlocality"
}
| null | null | null | null | true | null |
4316
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, the objects of our investigation are some dyadic operators,\nincluding dyadic shifts, multilinear paraproducts and multilinear Haar\nmultipliers. We mainly focus on the continuity and compactness of these\noperators. First, we consider the continuity properties of these operators.\nThen, by the Fréchet-Kolmogorov-Riesz-Tsuji theorem, the non-compactness\nproperties of these dyadic operators will be studied. Moreover, we show that\ntheir commutators are compact with \\textit{CMO} functions, which is quite\ndifferent from the non-compaceness properties of these dyadic operators. These\nresults are similar to those for Calderón-Zygmund singular integral\noperators.\n",
"title": "On Some properties of dyadic operators"
}
| null | null | null | null | true | null |
4317
| null |
Default
| null | null |
null |
{
"abstract": " At the beginning of a dynamic game, players may have exogenous theories about\nhow the opponents are going to play. Suppose that these theories are commonly\nknown. Then, players will refine their first-order beliefs, and challenge their\nown theories, through strategic reasoning. I develop and characterize\nepistemically a new solution concept, Selective Rationalizability, which\naccomplishes this task under the following assumption: when the observed\nbehavior is not compatible with the beliefs in players' rationality and\ntheories of all orders, players keep the orders of belief in rationality that\nare per se compatible with the observed behavior, and drop the incompatible\nbeliefs in the theories. Thus, Selective Rationalizability captures Common\nStrong Belief in Rationality (Battigalli and Siniscalchi, 2002) and refines\nExtensive-Form Rationalizability (Pearce, 1984; BS, 2002), whereas\nStrong-$\\Delta$-Rationalizability (Battigalli, 2003; Battigalli and\nSiniscalchi, 2003) captures the opposite epistemic priority choice. Selective\nRationalizability can be extended to encompass richer epistemic priority\norderings among different theories of opponents' behavior. This allows to\nestablish a surprising connection with strategic stability (Kohlberg and\nMertens, 1986).\n",
"title": "Rationalizability and Epistemic Priority Orderings"
}
| null | null | null | null | true | null |
4318
| null |
Default
| null | null |
null |
{
"abstract": " Wearable devices are transforming computing and the human-computer\ninteraction and they are a primary means for motion recognition of reflexive\nsystems. We review basic wearable deployments and their open wireless\ncommunications. An algorithm that uses accelerometer data to provide a control\nand communication signal is described. Challenges in the further deployment of\nwearable device in the field of body area network and biometric verification\nare discussed.\n",
"title": "Communications for Wearable Devices"
}
| null | null | null | null | true | null |
4319
| null |
Default
| null | null |
null |
{
"abstract": " We study standard and nonlocal nonlinear Schrödinger (NLS) equations\nobtained from the coupled NLS system of equations (Ablowitz-Kaup-Newell-Segur\n(AKNS) equations) by using standard and nonlocal reductions respectively. By\nusing the Hirota bilinear method we first find soliton solutions of the coupled\nNLS system of equations then using the reduction formulas we find the soliton\nsolutions of the standard and nonlocal NLS equations. We give examples for\nparticular values of the parameters and plot the function $|q(t,x)|^2$ for the\nstandard and nonlocal NLS equations.\n",
"title": "Nonlocal Nonlinear Schrödinger Equations and Their Soliton Solutions"
}
| null | null | null | null | true | null |
4320
| null |
Default
| null | null |
null |
{
"abstract": " Prediction is an appealing objective for self-supervised learning of\nbehavioral skills, particularly for autonomous robots. However, effectively\nutilizing predictive models for control, especially with raw image inputs,\nposes a number of major challenges. How should the predictions be used? What\nhappens when they are inaccurate? In this paper, we tackle these questions by\nproposing a method for learning robotic skills from raw image observations,\nusing only autonomously collected experience. We show that even an imperfect\nmodel can complete complex tasks if it can continuously retry, but this\nrequires the model to not lose track of the objective (e.g., the object of\ninterest). To enable a robot to continuously retry a task, we devise a\nself-supervised algorithm for learning image registration, which can keep track\nof objects of interest for the duration of the trial. We demonstrate that this\nidea can be combined with a video-prediction based controller to enable complex\nbehaviors to be learned from scratch using only raw visual inputs, including\ngrasping, repositioning objects, and non-prehensile manipulation. Our\nreal-world experiments demonstrate that a model trained with 160 robot hours of\nautonomously collected, unlabeled data is able to successfully perform complex\nmanipulation tasks with a wide range of objects not seen during training.\n",
"title": "Robustness via Retrying: Closed-Loop Robotic Manipulation with Self-Supervised Learning"
}
| null | null | null | null | true | null |
4321
| null |
Default
| null | null |
null |
{
"abstract": " Manifold learning based methods have been widely used for non-linear\ndimensionality reduction (NLDR). However, in many practical settings, the need\nto process streaming data is a challenge for such methods, owing to the high\ncomputational complexity involved. Moreover, most methods operate under the\nassumption that the input data is sampled from a single manifold, embedded in a\nhigh dimensional space. We propose a method for streaming NLDR when the\nobserved data is either sampled from multiple manifolds or irregularly sampled\nfrom a single manifold. We show that existing NLDR methods, such as Isomap,\nfail in such situations, primarily because they rely on smoothness and\ncontinuity of the underlying manifold, which is violated in the scenarios\nexplored in this paper. However, the proposed algorithm is able to learn\neffectively in presence of multiple, and potentially intersecting, manifolds,\nwhile allowing for the input data to arrive as a massive stream.\n",
"title": "S-Isomap++: Multi Manifold Learning from Streaming Data"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
4322
| null |
Validated
| null | null |
null |
{
"abstract": " We use data on extreme radio scintillation to demonstrate that this\nphenomenon is associated with hot stars in the solar neighbourhood. The ionized\ngas responsible for the scattering is found at distances up to 1.75pc from the\nhost star, and on average must comprise 1.E5 distinct structures per star. We\ndetect azimuthal velocities of the plasma, relative to the host star, up to 9.7\nkm/s, consistent with warm gas expanding at the sound speed. The circumstellar\nplasma structures that we infer are similar in several respects to the cometary\nknots seen in the Helix, and in other planetary nebulae. There the ionized gas\nappears as a skin around tiny molecular clumps. Our analysis suggests that\nmolecular clumps are ubiquitous circumstellar features, unrelated to the\nevolutionary state of the star. The total mass in such clumps is comparable to\nthe stellar mass.\n",
"title": "Extreme radio-wave scattering associated with hot stars"
}
| null | null | null | null | true | null |
4323
| null |
Default
| null | null |
null |
{
"abstract": " From longitudinal biomedical studies to social networks, graphs have emerged\nas a powerful framework for describing evolving interactions between agents in\ncomplex systems. In such studies, after pre-processing, the data can be\nrepresented by a set of graphs, each representing a system's state at different\npoints in time. The analysis of the system's dynamics depends on the selection\nof the appropriate analytical tools. After characterizing similarities between\nstates, a critical step lies in the choice of a distance between graphs capable\nof reflecting such similarities. While the literature offers a number of\ndistances that one could a priori choose from, their properties have been\nlittle investigated and no guidelines regarding the choice of such a distance\nhave yet been provided. In particular, most graph distances consider that the\nnodes are exchangeable and do not take into account node identities. Accounting\nfor the alignment of the graphs enables us to enhance these distances'\nsensitivity to perturbations in the network and detect important changes in\ngraph dynamics. Thus the selection of an adequate metric is a decisive --yet\ndelicate--practical matter.\nIn the spirit of Goldenberg, Zheng and Fienberg's seminal 2009 review, the\npurpose of this article is to provide an overview of commonly-used graph\ndistances and an explicit characterization of the structural changes that they\nare best able to capture. We use as a guiding thread to our discussion the\napplication of these distances to the analysis of both a longitudinal\nmicrobiome dataset and a brain fMRI study. We show examples of using\npermutation tests to detect the effect of covariates on the graphs'\nvariability. Synthetic examples provide intuition as to the qualities and\ndrawbacks of the different distances. Above all, we provide some guidance for\nchoosing one distance over another in certain types of applications.\n",
"title": "Tracking network dynamics: a survey of distances and similarity metrics"
}
| null | null | null | null | true | null |
4324
| null |
Default
| null | null |
null |
{
"abstract": " Both neural networks and decision trees are popular machine learning methods\nand are widely used to solve problems from diverse domains. These two\nclassifiers are commonly used base classifiers in an ensemble framework. In\nthis paper, we first present a new variant of oblique decision tree based on a\nlinear classifier, then construct an ensemble classifier based on the fusion of\na fast neural network, random vector functional link network and oblique\ndecision trees. Random Vector Functional Link Network has an elegant closed\nform solution with extremely short training time. The neural network partitions\neach training bag (obtained using bagging) at the root level into C subsets\nwhere C is the number of classes in the dataset and subsequently, C oblique\ndecision trees are trained on such partitions. The proposed method provides a\nrich insight into the data by grouping the confusing or hard to classify\nsamples for each class and thus, provides an opportunity to employ fine-grained\nclassification rule over the data. The performance of the ensemble classifier\nis evaluated on several multi-class datasets where it demonstrates a superior\nperformance compared to other state-of- the-art classifiers.\n",
"title": "Enhancing Multi-Class Classification of Random Forest using Random Vector Functional Neural Network and Oblique Decision Surfaces"
}
| null | null |
[
"Statistics"
] | null | true | null |
4325
| null |
Validated
| null | null |
null |
{
"abstract": " Let $A$ be a regular ring containing a field of characteristic zero and let\n$R = A[X_1,\\ldots, X_m]$. Consider $R$ as standard graded with $deg \\ A = 0$\nand $deg \\ X_i = 1$ for all $i$. In this paper we present a comprehensive study\nof graded components of local cohomology modules $H^i_I(R)$ where $I$ is an\n\\emph{arbitrary} homogeneous ideal in $R$. Our study seems to be the first in\nthis regard.\n",
"title": "Graded components of local cohomology modules"
}
| null | null | null | null | true | null |
4326
| null |
Default
| null | null |
null |
{
"abstract": " Selecting the right drugs for the right patients is a primary goal of\nprecision medicine. In this manuscript, we consider the problem of cancer drug\nselection in a learning-to-rank framework. We have formulated the cancer drug\nselection problem as to accurately predicting 1). the ranking positions of\nsensitive drugs and 2). the ranking orders among sensitive drugs in cancer cell\nlines based on their responses to cancer drugs. We have developed a new\nlearning-to-rank method, denoted as pLETORg , that predicts drug ranking\nstructures in each cell line via using drug latent vectors and cell line latent\nvectors. The pLETORg method learns such latent vectors through explicitly\nenforcing that, in the drug ranking list of each cell line, the sensitive drugs\nare pushed above insensitive drugs, and meanwhile the ranking orders among\nsensitive drugs are correct. Genomics information on cell lines is leveraged in\nlearning the latent vectors. Our experimental results on a benchmark cell\nline-drug response dataset demonstrate that the new pLETORg significantly\noutperforms the state-of-the-art method in prioritizing new sensitive drugs.\n",
"title": "Drug Selection via Joint Push and Learning to Rank"
}
| null | null | null | null | true | null |
4327
| null |
Default
| null | null |
null |
{
"abstract": " The temperature dependence of the electrical resistivity of the\nheterostructures consisting of single crystalline LaMnO$_3$ samples with\ndifferent crystallographic orientations covered by the epitaxial ferroelectric\nBa$_{0.8}$Sr$_{0.2}$TiO$_3$ film has been studied. Results obtained for the\nheterostructure have been compared with the electrical resistivity of the\nsingle crystalline LaMnO$_3$ without the film. It was found that for the\nsamples with the films where the polarization axis is perpendicular to the\ncrystal surface the electrical resistivity strongly decreases, and at the\ntemperature below ~160 K undergoes the insulator-metal transition. Ab-initio\ncalculations were also performed for the structural and electronic properties\nof the BaTiO$_3$/LaMnO$_3$ heterostructure. Transition to the 2D electron gas\nat the interface is shown.\n",
"title": "Two-dimensional electron gas at the interface of the ferroelectric-antiferromagnetic heterostructure Ba_0.8Sr_0.2TiO_3/LaMnO_3"
}
| null | null |
[
"Physics"
] | null | true | null |
4328
| null |
Validated
| null | null |
null |
{
"abstract": " We develop a novel method for training of GANs for unsupervised and class\nconditional generation of images, called Linear Discriminant GAN (LD-GAN). The\ndiscriminator of an LD-GAN is trained to maximize the linear separability\nbetween distributions of hidden representations of generated and targeted\nsamples, while the generator is updated based on the decision hyper-planes\ncomputed by performing LDA over the hidden representations. LD-GAN provides a\nconcrete metric of separation capacity for the discriminator, and we\nexperimentally show that it is possible to stabilize the training of LD-GAN\nsimply by calibrating the update frequencies between generators and\ndiscriminators in the unsupervised case, without employment of normalization\nmethods and constraints on weights. In the class conditional generation tasks,\nthe proposed method shows improved training stability together with better\ngeneralization performance compared to WGAN that employs an auxiliary\nclassifier.\n",
"title": "Linear Discriminant Generative Adversarial Networks"
}
| null | null | null | null | true | null |
4329
| null |
Default
| null | null |
null |
{
"abstract": " Monoclonal antibodies constitute one of the most important strategies to\ntreat patients suffering from cancers such as hematological malignancies and\nsolid tumors. In order to guarantee the quality of those preparations prepared\nat hospital, quality control has to be developed. The aim of this study was to\nexplore a noninvasive, nondestructive, and rapid analytical method to ensure\nthe quality of the final preparation without causing any delay in the process.\nWe analyzed four mAbs (Inlfiximab, Bevacizumab, Ramucirumab and Rituximab)\ndiluted at therapeutic concentration in chloride sodium 0.9% using Raman\nspectroscopy. To reduce the prediction errors obtained with traditional\nchemometric data analysis, we explored a data-driven approach using statistical\nmachine learning methods where preprocessing and predictive models are jointly\noptimized. We prepared a data analytics workflow and submitted the problem to a\ncollaborative data challenge platform called Rapid Analytics and Model\nPrototyping (RAMP). This allowed to use solutions from about 300 data\nscientists during five days of collaborative work. The prediction of the four\nmAbs samples was considerably improved with a misclassification rate and the\nmean error rate of 0.8% and 4%, respectively.\n",
"title": "Machine learning for classification and quantification of monoclonal antibody preparations for cancer therapy"
}
| null | null | null | null | true | null |
4330
| null |
Default
| null | null |
null |
{
"abstract": " It is well known that the addition of noise in a multistable system can\ninduce random transitions between stable states. The rate of transition can be\ncharacterised in terms of the noise-free system's dynamics and the added noise:\nfor potential systems in the presence of asymptotically low noise the\nwell-known Kramers' escape time gives an expression for the mean escape time.\nThis paper examines some general properties and examples of transitions between\nlocal steady and oscillatory attractors within networks: the transition rates\nat each node may be affected by the dynamics at other nodes. We use first\npassage time theory to explain some properties of scalings noted in the\nliterature for an idealised model of initiation of epileptic seizures in small\nsystems of coupled bistable systems with both steady and oscillatory\nattractors. We focus on the case of sequential escapes where a steady attractor\nis only marginally stable but all nodes start in this state. As the nodes\nescape to the oscillatory regime, we assume that the transitions back are very\ninfrequent in comparison. We quantify and characterise the resulting sequences\nof noise-induced escapes. For weak enough coupling we show that a master\nequation approach gives a good quantitative understanding of sequential\nescapes, but for strong coupling this description breaks down.\n",
"title": "Sequential noise-induced escapes for oscillatory network dynamics"
}
| null | null | null | null | true | null |
4331
| null |
Default
| null | null |
null |
{
"abstract": " Conditional term rewriting is an intuitive yet complex extension of term\nrewriting. In order to benefit from the simpler framework of unconditional\nrewriting, transformations have been defined to eliminate the conditions of\nconditional term rewrite systems.\nRecent results provide confluence criteria for conditional term rewrite\nsystems via transformations, yet they are restricted to CTRSs with certain\nsyntactic properties like weak left-linearity. These syntactic properties imply\nthat the transformations are sound for the given CTRS.\nThis paper shows how to use transformations to prove confluence of\noperationally terminating, right-stable deterministic conditional term rewrite\nsystems without the necessity of soundness restrictions. For this purpose, it\nis shown that certain rewrite strategies, in particular almost U-eagerness and\ninnermost rewriting, always imply soundness.\n",
"title": "Confluence of Conditional Term Rewrite Systems via Transformations"
}
| null | null | null | null | true | null |
4332
| null |
Default
| null | null |
null |
{
"abstract": " Magnetic skyrmions are topological spin structures having immense potential\nfor energy efficient spintronic devices. However, observations of skyrmions at\nroom temperature are limited to patterned nanostructures. Here, we report the\nobservation of stable skyrmions in unpatterned Ta/Co2FeAl(CFA)/MgO thin film\nheterostructures at room temperature and in zero external magnetic field\nemploying magnetic force microscopy. The skyrmions are observed in a trilayer\nstructure comprised of heavy metal (HM)/ferromagnet (FM)/Oxide interfaces which\nresult in strong interfacial Dzyaloshinskii-Moriya interaction (i-DMI) as\nevidenced by Brillouin light scattering measurements, in agreement with the\nresults of micromagnetic simulations. We also emphasize on room temperature\nobservation of multiple skyrmions which can be stabilized for suitable choices\nof CFA layer thickness, perpendicular magnetic anisotropy, and i-DMI. These\nresults open up a new paradigm for designing room temperature spintronic\ndevices based on skyrmions in FM continuous thin films.\n",
"title": "Observation of Skyrmions at Room Temperature in Co2FeAl Heusler Alloy Ultrathin Films"
}
| null | null |
[
"Physics"
] | null | true | null |
4333
| null |
Validated
| null | null |
null |
{
"abstract": " Spectral estimation (SE) aims to identify how the energy of a signal (e.g., a\ntime series) is distributed across different frequencies. This can become\nparticularly challenging when only partial and noisy observations of the signal\nare available, where current methods fail to handle uncertainty appropriately.\nIn this context, we propose a joint probabilistic model for signals,\nobservations and spectra, where SE is addressed as an exact inference problem.\nAssuming a Gaussian process prior over the signal, we apply Bayes' rule to find\nthe analytic posterior distribution of the spectrum given a set of\nobservations. Besides its expressiveness and natural account of spectral\nuncertainty, the proposed model also provides a functional-form representation\nof the power spectral density, which can be optimised efficiently. Comparison\nwith previous approaches, in particular against Lomb-Scargle, is addressed\ntheoretically and also experimentally in three different scenarios. Code and\ndemo available at this https URL.\n",
"title": "Bayesian Nonparametric Spectral Estimation"
}
| null | null | null | null | true | null |
4334
| null |
Default
| null | null |
null |
{
"abstract": " We present a new walking foot-placement controller based on 3LP, a 3D model\nof bipedal walking that is composed of three pendulums to simulate falling,\nswing and torso dynamics. Taking advantage of linear equations and closed-form\nsolutions of the 3LP model, our proposed controller projects intermediate\nstates of the biped back to the beginning of the phase for which a discrete LQR\ncontroller is designed. After the projection, a proper control policy is\ngenerated by this LQR controller and used at the intermediate time. This\ncontrol paradigm reacts to disturbances immediately and includes rules to\naccount for swing dynamics and leg-retraction. We apply it to a simulated Atlas\nrobot in position-control, always commanded to perform in-place walking. The\nstance hip joint in our robot keeps the torso upright to let the robot\nnaturally fall, and the swing hip joint tracks the desired footstep location.\nCombined with simple Center of Pressure (CoP) damping rules in the low-level\ncontroller, our foot-placement enables the robot to recover from strong pushes\nand produce periodic walking gaits when subject to persistent sources of\ndisturbance, externally or internally. These gaits are imprecise, i.e.,\nemergent from asymmetry sources rather than precisely imposing a desired\nvelocity to the robot. Also in extreme conditions, restricting linearity\nassumptions of the 3LP model are often violated, but the system remains robust\nin our simulations. An extensive analysis of closed-loop eigenvalues, viable\nregions and sensitivity to push timings further demonstrate the strengths of\nour simple controller.\n",
"title": "Imprecise dynamic walking with time-projection control"
}
| null | null | null | null | true | null |
4335
| null |
Default
| null | null |
null |
{
"abstract": " We study the magnetic field effects on the diluted spin-ice materials using\nthe replica-exchange Monte Carlo simulation. We observe five plateaus in the\nmagnetization curve of the diluted nearest-neighbor spin-ice model on the\npyrochlore lattice when a magnetic field is applied in the [111] direction.\nThis is in contrast to the case of the pure model with two plateaus. The origin\nof five plateaus is investigated from the spin configuration of two\ncorner-sharing tetrahedra in the case of the diluted model.\n",
"title": "Interplay of dilution and magnetic field in the nearest-neighbor spin-ice model on the pyrochlore lattice"
}
| null | null |
[
"Physics"
] | null | true | null |
4336
| null |
Validated
| null | null |
null |
{
"abstract": " An RNN-based forecasting approach is used to early detect anomalies in\nindustrial multivariate time series data from a simulated Tennessee Eastman\nProcess (TEP) with many cyber-attacks. This work continues a previously\nproposed LSTM-based approach to the fault detection in simpler data. It is\nconsidered necessary to adapt the RNN network to deal with data containing\nstochastic, stationary, transitive and a rich variety of anomalous behaviours.\nThere is particular focus on early detection with special NAB-metric. A\ncomparison with the DPCA approach is provided. The generated data set is made\npublicly available.\n",
"title": "RNN-based Early Cyber-Attack Detection for the Tennessee Eastman Process"
}
| null | null | null | null | true | null |
4337
| null |
Default
| null | null |
null |
{
"abstract": " Polymer solar cells are considered as very promising candidates for\ndevelopment of photovoltaics of the future. They are cheap and easy to\nfabricate, however, up to now, they possess fundamental drawback, low\neffectiveness. In the most popular BHJ (bulk heterojunction) architecture the\nactual record of efficiency is about 13 percent. One ask the question how\nfundamental this limitation is. In our paper we propose the simple model which\nexamines the limitations of efficiency by analysis of geometrical aspects of\nthe BHJ architecture. In this paper we considered two dimensional model. We\ncalculated the effective length of the donor-acceptor border in the random\nmixture of donor and acceptor nanocrystals and further compared it with an\nideal comb architecture. It turns out that in the BHJ architecture, this\neffective length is about 2 times smaller than in the comb architecture.\n",
"title": "Modelling of limitations of bulk heterojunction architecture in organic solar cells"
}
| null | null |
[
"Physics"
] | null | true | null |
4338
| null |
Validated
| null | null |
null |
{
"abstract": " A `flutter machine' is introduced for the investigation of a singular\ninterface between the classical and reversible Hopf bifurcations that is\ntheoretically predicted to be generic in nonconservative reversible systems\nwith vanishing dissipation. In particular, such a singular interface exists for\nthe Pflüger viscoelastic column moving in a resistive medium, which is proven\nby means of the perturbation theory of multiple eigenvalues with the Jordan\nblock. The laboratory setup, consisting of a cantilevered viscoelastic rod\nloaded by a positional force with non-zero curl produced by dry friction,\ndemonstrates high sensitivity of the classical Hopf bifurcation onset {to the\nratio between} the weak air drag and Kelvin-Voigt damping in the Pflüger\ncolumn. Thus, the Whitney umbrella singularity is experimentally confirmed,\nresponsible for discontinuities accompanying dissipation-induced instabilities\nin a broad range of physical contexts.\n",
"title": "Detecting singular weak-dissipation limit for flutter onset in reversible systems"
}
| null | null | null | null | true | null |
4339
| null |
Default
| null | null |
null |
{
"abstract": " This paper proposes a novel robotic hand design for assembly tasks. The idea\nis to combine two simple grippers -- an inner gripper which is used for precise\nalignment, and an outer gripper which is used for stable holding. Conventional\nrobotic hands require complicated compliant mechanisms or complicated control\nstrategy and force sensing to conduct assemble tasks, which makes them costly\nand difficult to pick and arrange small objects like screws or washers.\nCompared to the conventional hands, the proposed design provides a low-cost\nsolution for aligning, picking up, and arranging various objects by taking\nadvantages of the geometric constraints of the positioning fingers and gravity.\nIt is able to deal with small screws and washers, and eliminate the position\nerrors of cylindrical objects or objects with cylindrical holes. In the\nexperiments, both real-world tasks and quantitative analysis are performed to\nvalidate the aligning, picking, and arrangements abilities of the design.\n",
"title": "A Hand Combining Two Simple Grippers to Pick up and Arrange Objects for Assembly"
}
| null | null | null | null | true | null |
4340
| null |
Default
| null | null |
null |
{
"abstract": " The celebrated Time Hierarchy Theorem for Turing machines states, informally,\nthat more problems can be solved given more time. The extent to which a time\nhierarchy-type theorem holds in the distributed LOCAL model has been open for\nmany years. It is consistent with previous results that all natural problems in\nthe LOCAL model can be classified according to a small constant number of\ncomplexities, such as $O(1),O(\\log^* n), O(\\log n), 2^{O(\\sqrt{\\log n})}$, etc.\nIn this paper we establish the first time hierarchy theorem for the LOCAL\nmodel and prove that several gaps exist in the LOCAL time hierarchy.\n1. We define an infinite set of simple coloring problems called Hierarchical\n$2\\frac{1}{2}$-Coloring}. A correctly colored graph can be confirmed by simply\nchecking the neighborhood of each vertex, so this problem fits into the class\nof locally checkable labeling (LCL) problems. However, the complexity of the\n$k$-level Hierarchical $2\\frac{1}{2}$-Coloring problem is $\\Theta(n^{1/k})$,\nfor $k\\in\\mathbb{Z}^+$. The upper and lower bounds hold for both general graphs\nand trees, and for both randomized and deterministic algorithms.\n2. Consider any LCL problem on bounded degree trees. We prove an\nautomatic-speedup theorem that states that any randomized $n^{o(1)}$-time\nalgorithm solving the LCL can be transformed into a deterministic $O(\\log\nn)$-time algorithm. Together with a previous result, this establishes that on\ntrees, there are no natural deterministic complexities in the ranges\n$\\omega(\\log^* n)$---$o(\\log n)$ or $\\omega(\\log n)$---$n^{o(1)}$.\n3. We expose a gap in the randomized time hierarchy on general graphs. Any\nrandomized algorithm that solves an LCL problem in sublogarithmic time can be\nsped up to run in $O(T_{LLL})$ time, which is the complexity of the distributed\nLovasz local lemma problem, currently known to be $\\Omega(\\log\\log n)$ and\n$O(\\log n)$.\n",
"title": "A Time Hierarchy Theorem for the LOCAL Model"
}
| null | null | null | null | true | null |
4341
| null |
Default
| null | null |
null |
{
"abstract": " We study sound in Galilean invariant systems of one-dimensional fermions. At\nlow temperatures, we find a broad range of frequencies in which in addition to\nthe waves of density there is a second sound corresponding to ballistic\npropagation of heat in the system. The damping of the second sound mode is\nweak, provided the frequency is large compared to a relaxation rate that is\nexponentially small at low temperatures. At lower frequencies the second sound\nmode is damped, and the propagation of heat is diffusive.\n",
"title": "Second sound in systems of one-dimensional fermions"
}
| null | null | null | null | true | null |
4342
| null |
Default
| null | null |
null |
{
"abstract": " The natural join and the inner union operations combine relations of a\ndatabase. Tropashko and Spight [24] realized that these two operations are the\nmeet and join operations in a class of lattices, known by now as the relational\nlattices. They proposed then lattice theory as an algebraic approach to the\ntheory of databases, alternative to the relational algebra. Previous works [17,\n22] proved that the quasiequational theory of these lattices-that is, the set\nof definite Horn sentences valid in all the relational lattices-is undecidable,\neven when the signature is restricted to the pure lattice signature. We prove\nhere that the equational theory of relational lattices is decidable. That, is\nwe provide an algorithm to decide if two lattice theoretic terms t, s are made\nequal under all intepretations in some relational lattice. We achieve this goal\nby showing that if an inclusion t $\\le$ s fails in any of these lattices, then\nit fails in a relational lattice whose size is bound by a triple exponential\nfunction of the sizes of t and s.\n",
"title": "The equational theory of the natural join and inner union is decidable"
}
| null | null |
[
"Computer Science",
"Mathematics"
] | null | true | null |
4343
| null |
Validated
| null | null |
null |
{
"abstract": " Our predictions, based on density-functional calculations, reveal that\nsurface doping of ZnO nanowires with Bi leads to a linear-in-$k$ splitting of\nthe conduction-band states, through spin-orbit interaction, due to the lowering\nof the symmetry in the presence of the dopant. This finding implies that spin\npolarization of the conduction electrons in Bi-doped ZnO nanowires could be\ncontrolled with applied electric (as opposed to magnetic) fields, making them\ncandidate materials for spin-orbitronic applications. Our findings also show\nthat the degree of spin splitting could be tuned by adjusting the dopant\nconcentration. Defect calculations and ab initio molecular dynamics simulations\nindicate that stable doping configurations exhibiting the foregoing\nlinear-in-$k$ splitting could be realized under reasonable thermodynamic\nconditions.\n",
"title": "Doping-induced spin-orbit splitting in Bi-doped ZnO nanowires"
}
| null | null |
[
"Physics"
] | null | true | null |
4344
| null |
Validated
| null | null |
null |
{
"abstract": " The autonomous measurement of tree traits, such as branching structure,\nbranch diameters, branch lengths, and branch angles, is required for tasks such\nas robotic pruning of trees as well as structural phenotyping. We propose a\nrobotic vision system called the Robotic System for Tree Shape Estimation\n(RoTSE) to determine tree traits in field settings. The process is composed of\nthe following stages: image acquisition with a mobile robot unit, segmentation,\nreconstruction, curve skeletonization, conversion to a graph representation,\nand then computation of traits. Quantitative and qualitative results on apple\ntrees are shown in terms of accuracy, computation time, and robustness.\nCompared to ground truth measurements, the RoTSE produced the following\nestimates: branch diameter (mean-squared error $0.99$ mm), branch length\n(mean-squared error $45.64$ mm), and branch angle (mean-squared error $10.36$\ndegrees). The average run time was 8.47 minutes when the voxel resolution was\n$3$ mm$^3$.\n",
"title": "A robotic vision system to measure tree traits"
}
| null | null |
[
"Computer Science"
] | null | true | null |
4345
| null |
Validated
| null | null |
null |
{
"abstract": " Mechanical vibrations of components of the optical system is one of the\nsources of blurring of interference pattern in coherent imaging systems. The\nproblem is especially important in holography where the resolution of the\nreconstructed objects depends on the effective size of the hologram, that is on\nthe extent of the interference pattern, and on the contrast of the interference\nfringes. We discuss the mathematical relation between the vibrations, the\nhologram contrast and the reconstructed object. We show how vibrations can be\npost-filtered out from the hologram or from the reconstructed object assuming a\nGaussian distribution of the vibrations. We also provide a numerical example of\ncompensation for directional motion blur. We demonstrate our approach for light\noptical and electron holograms, acquired with both, plane- as well as\nspherical-waves. As a result of such hologram deblurring, the resolution of the\nreconstructed objects is enhanced by almost a factor of 2. We believe that our\napproach opens up a new venue of post-experimental resolution enhancement in\nin-line holography by adapting the rich database/catalogue of motion deblurring\nalgorithms developed for photography and image restoration applications.\n",
"title": "Resolution enhancement in in-line holography by numerical compensation of vibrations"
}
| null | null | null | null | true | null |
4346
| null |
Default
| null | null |
null |
{
"abstract": " We present a new AI task -- Embodied Question Answering (EmbodiedQA) -- where\nan agent is spawned at a random location in a 3D environment and asked a\nquestion (\"What color is the car?\"). In order to answer, the agent must first\nintelligently navigate to explore the environment, gather information through\nfirst-person (egocentric) vision, and then answer the question (\"orange\").\nThis challenging task requires a range of AI skills -- active perception,\nlanguage understanding, goal-driven navigation, commonsense reasoning, and\ngrounding of language into actions. In this work, we develop the environments,\nend-to-end-trained reinforcement learning agents, and evaluation protocols for\nEmbodiedQA.\n",
"title": "Embodied Question Answering"
}
| null | null |
[
"Computer Science"
] | null | true | null |
4347
| null |
Validated
| null | null |
null |
{
"abstract": " Controlled generation of text is of high practical use. Recent efforts have\nmade impressive progress in generating or editing sentences with given textual\nattributes (e.g., sentiment). This work studies a new practical setting of text\ncontent manipulation. Given a structured record, such as `(PLAYER: Lebron,\nPOINTS: 20, ASSISTS: 10)', and a reference sentence, such as `Kobe easily\ndropped 30 points', we aim to generate a sentence that accurately describes the\nfull content in the record, with the same writing style (e.g., wording,\ntransitions) of the reference. The problem is unsupervised due to lack of\nparallel data in practice, and is challenging to minimally yet effectively\nmanipulate the text (by rewriting/adding/deleting text portions) to ensure\nfidelity to the structured content. We derive a dataset from a basketball game\nreport corpus as our testbed, and develop a neural method with unsupervised\ncompeting objectives and explicit content coverage constraints. Automatic and\nhuman evaluations show superiority of our approach over competitive methods\nincluding a strong rule-based baseline and prior approaches designed for style\ntransfer.\n",
"title": "Toward Unsupervised Text Content Manipulation"
}
| null | null | null | null | true | null |
4348
| null |
Default
| null | null |
null |
{
"abstract": " Learning automatically the structure of object categories remains an\nimportant open problem in computer vision. In this paper, we propose a novel\nunsupervised approach that can discover and learn landmarks in object\ncategories, thus characterizing their structure. Our approach is based on\nfactorizing image deformations, as induced by a viewpoint change or an object\ndeformation, by learning a deep neural network that detects landmarks\nconsistently with such visual effects. Furthermore, we show that the learned\nlandmarks establish meaningful correspondences between different object\ninstances in a category without having to impose this requirement explicitly.\nWe assess the method qualitatively on a variety of object types, natural and\nman-made. We also show that our unsupervised landmarks are highly predictive of\nmanually-annotated landmarks in face benchmark datasets, and can be used to\nregress these with a high degree of accuracy.\n",
"title": "Unsupervised learning of object landmarks by factorized spatial embeddings"
}
| null | null | null | null | true | null |
4349
| null |
Default
| null | null |
null |
{
"abstract": " Design optimization techniques are often used at the beginning of the design\nprocess to explore the space of possible designs. In these domains illumination\nalgorithms, such as MAP-Elites, are promising alternatives to classic\noptimization algorithms because they produce diverse, high-quality solutions in\na single run, instead of only a single near-optimal solution. Unfortunately,\nthese algorithms currently require a large number of function evaluations,\nlimiting their applicability. In this article we introduce a new illumination\nalgorithm, Surrogate-Assisted Illumination (SAIL), that leverages surrogate\nmodeling techniques to create a map of the design space according to\nuser-defined features while minimizing the number of fitness evaluations. On a\n2-dimensional airfoil optimization problem SAIL produces hundreds of diverse\nbut high-performing designs with several orders of magnitude fewer evaluations\nthan MAP-Elites or CMA-ES. We demonstrate that SAIL is also capable of\nproducing maps of high-performing designs in realistic 3-dimensional\naerodynamic tasks with an accurate flow simulation. Data-efficient design\nexploration with SAIL can help designers understand what is possible, beyond\nwhat is optimal, by considering more than pure objective-based optimization.\n",
"title": "Data-Efficient Design Exploration through Surrogate-Assisted Illumination"
}
| null | null | null | null | true | null |
4350
| null |
Default
| null | null |
null |
{
"abstract": " We present an application of deep generative models in the context of\npartial-differential equation (PDE) constrained inverse problems. We combine a\ngenerative adversarial network (GAN) representing an a priori model that\ncreates subsurface geological structures and their petrophysical properties,\nwith the numerical solution of the PDE governing the propagation of acoustic\nwaves within the earth's interior. We perform Bayesian inversion using an\napproximate Metropolis-adjusted Langevin algorithm (MALA) to sample from the\nposterior given seismic observations. Gradients with respect to the model\nparameters governing the forward problem are obtained by solving the adjoint of\nthe acoustic wave equation. Gradients of the mismatch with respect to the\nlatent variables are obtained by leveraging the differentiable nature of the\ndeep neural network used to represent the generative model. We show that\napproximate MALA sampling allows efficient Bayesian inversion of model\nparameters obtained from a prior represented by a deep generative model,\nobtaining a diverse set of realizations that reflect the observed seismic\nresponse.\n",
"title": "Stochastic seismic waveform inversion using generative adversarial networks as a geological prior"
}
| null | null |
[
"Statistics"
] | null | true | null |
4351
| null |
Validated
| null | null |
null |
{
"abstract": " In 1996, Kirk Lancaster and David Siegel investigated the existence and\nbehavior of radial limits at a corner of the boundary of the domain of\nsolutions of capillary and other prescribed mean curvature problems with\ncontact angle boundary data. In Theorem 3, they provide an example of a\ncapillary surface in a unit disk $D$ which has no radial limits at\n$(0,0)\\in\\partial D.$ In their example, the contact angle ($\\gamma$) cannot be\nbounded away from zero and $\\pi.$\nHere we consider a domain $\\Omega$ with a convex corner at $(0,0)$ and find a\ncapillary surface $z=f(x,y)$ in $\\Omega\\times\\mathbb{R}$ which has no radial\nlimits at $(0,0)\\in\\partial\\Omega$ such that $\\gamma$ is bounded away from $0$\nand $\\pi.$\n",
"title": "A Capillary Surface with No Radial Limits"
}
| null | null | null | null | true | null |
4352
| null |
Default
| null | null |
null |
{
"abstract": " Brain computer interface (BCI) provides promising applications in\nneuroprosthesis and neurorehabilitation by controlling computers and robotic\ndevices based on the patient's intentions. Here, we have developed a novel BCI\nplatform that controls a personalized social robot using noninvasively acquired\nbrain signals. Scalp electroencephalogram (EEG) signals are collected from a\nuser in real-time during tasks of imaginary movements. The imagined body\nkinematics are decoded using a regression model to calculate the user-intended\nvelocity. Then, the decoded kinematic information is mapped to control the\ngestures of a social robot. The platform here may be utilized as a\nhuman-robot-interaction framework by combining with neurofeedback mechanisms to\nenhance the cognitive capability of persons with dementia.\n",
"title": "Brain Computer Interface for Gesture Control of a Social Robot: an Offline Study"
}
| null | null | null | null | true | null |
4353
| null |
Default
| null | null |
null |
{
"abstract": " This paper deals with motion planning for multiple agents by representing the\nproblem as a simultaneous optimization of every agent's trajectory. Each\ntrajectory is considered as a sample from a one-dimensional continuous-time\nGaussian process (GP) generated by a linear time-varying stochastic\ndifferential equation driven by white noise. By formulating the planning\nproblem as probabilistic inference on a factor graph, the structure of the\npertaining GP can be exploited to find the solution efficiently using numerical\noptimization. In contrast to planning each agent's trajectory individually,\nwhere only the current poses of other agents are taken into account, we propose\nsimultaneous planning of multiple trajectories that works in a predictive\nmanner. It takes into account the information about each agent's whereabouts at\nevery future time instant, since full trajectories of each agent are found\njointly during a single optimization procedure. We compare the proposed method\nto an individual trajectory planning approach, demonstrating significant\nimprovement in both success rate and computational efficiency.\n",
"title": "Multi-agent Gaussian Process Motion Planning via Probabilistic Inference"
}
| null | null | null | null | true | null |
4354
| null |
Default
| null | null |
null |
{
"abstract": " Graph games provide the foundation for modeling and synthesis of reactive\nprocesses. Such games are played over graphs where the vertices are controlled\nby two adversarial players. We consider graph games where the objective of the\nfirst player is the conjunction of a qualitative objective (specified as a\nparity condition) and a quantitative objective (specified as a mean-payoff\ncondition). There are two variants of the problem, namely, the threshold\nproblem where the quantitative goal is to ensure that the mean-payoff value is\nabove a threshold, and the value problem where the quantitative goal is to\nensure the optimal mean-payoff value; in both cases ensuring the qualitative\nparity objective. The previous best-known algorithms for game graphs with $n$\nvertices, $m$ edges, parity objectives with $d$ priorities, and maximal\nabsolute reward value $W$ for mean-payoff objectives, are as follows:\n$O(n^{d+1} \\cdot m \\cdot W)$ for the threshold problem, and $O(n^{d+2} \\cdot m\n\\cdot W)$ for the value problem. Our main contributions are faster algorithms,\nand the running times of our algorithms are as follows: $O(n^{d-1} \\cdot m\n\\cdot W)$ for the threshold problem, and $O(n^{d} \\cdot m \\cdot W \\cdot \\log\n(n\\cdot W))$ for the value problem. For mean-payoff parity objectives with two\npriorities, our algorithms match the best-known bounds of the algorithms for\nmean-payoff games (without conjunction with parity objectives). Our results are\nrelevant in synthesis of reactive systems with both functional requirement\n(given as a qualitative objective) and performance requirement (given as a\nquantitative objective).\n",
"title": "Faster Algorithms for Mean-Payoff Parity Games"
}
| null | null | null | null | true | null |
4355
| null |
Default
| null | null |
null |
{
"abstract": " Recent work in distance metric learning has focused on learning\ntransformations of data that best align with specified pairwise similarity and\ndissimilarity constraints, often supplied by a human observer. The learned\ntransformations lead to improved retrieval, classification, and clustering\nalgorithms due to the better adapted distance or similarity measures. Here, we\naddress the problem of learning these transformations when the underlying\nconstraint generation process is nonstationary. This nonstationarity can be due\nto changes in either the ground-truth clustering used to generate constraints\nor changes in the feature subspaces in which the class structure is apparent.\nWe propose Online Convex Ensemble StrongLy Adaptive Dynamic Learning (OCELAD),\na general adaptive, online approach for learning and tracking optimal metrics\nas they change over time that is highly robust to a variety of nonstationary\nbehaviors in the changing metric. We apply the OCELAD framework to an ensemble\nof online learners. Specifically, we create a retro-initialized composite\nobjective mirror descent (COMID) ensemble (RICE) consisting of a set of\nparallel COMID learners with different learning rates, and demonstrate\nparameter-free RICE-OCELAD metric learning on both synthetic data and a highly\nnonstationary Twitter dataset. We show significant performance improvements and\nincreased robustness to nonstationary effects relative to previously proposed\nbatch and online distance metric learning algorithms.\n",
"title": "Similarity Function Tracking using Pairwise Comparisons"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
4356
| null |
Validated
| null | null |
null |
{
"abstract": " Attenuation correction is an essential requirement of positron emission\ntomography (PET) image reconstruction to allow for accurate quantification.\nHowever, attenuation correction is particularly challenging for PET-MRI as\nneither PET nor magnetic resonance imaging (MRI) can directly image tissue\nattenuation properties. MRI-based computed tomography (CT) synthesis has been\nproposed as an alternative to physics based and segmentation-based approaches\nthat assign a population-based tissue density value in order to generate an\nattenuation map. We propose a novel deep fully convolutional neural network\nthat generates synthetic CTs in a recursive manner by gradually reducing the\nresiduals of the previous network, increasing the overall accuracy and\ngeneralisability, while keeping the number of trainable parameters within\nreasonable limits. The model is trained on a database of 20 pre-acquired MRI/CT\npairs and a four-fold random bootstrapped validation with a 80:20 split is\nperformed. Quantitative results show that the proposed framework outperforms a\nstate-of-the-art atlas-based approach decreasing the Mean Absolute Error (MAE)\nfrom 131HU to 68HU for the synthetic CTs and reducing the PET reconstruction\nerror from 14.3% to 7.2%.\n",
"title": "Deep Boosted Regression for MR to CT Synthesis"
}
| null | null | null | null | true | null |
4357
| null |
Default
| null | null |
null |
{
"abstract": " We introduce a general framework allowing to apply the theory of regularity\nstructures to discretisations of stochastic PDEs. The approach pursued in this\narticle is that we do not focus on any one specific discretisation procedure.\nInstead, we assume that we are given a scale $\\varepsilon > 0$ and a \"black\nbox\" describing the behaviour of our discretised objects at scales below\n$\\varepsilon $.\n",
"title": "Discretisation of regularity structures"
}
| null | null | null | null | true | null |
4358
| null |
Default
| null | null |
null |
{
"abstract": " This paper presents a rigorous optimization technique for wireless power\ntransfer (WPT) systems enhanced by passive elements, ranging from simple\nreflectors and intermedi- ate relays all the way to general electromagnetic\nguiding and focusing structures, such as metasurfaces and metamaterials. At its\ncore is a convex semidefinite relaxation formulation of the otherwise nonconvex\noptimization problem, of which tightness and optimality can be confirmed by a\nsimple test of its solutions. The resulting method is rigorous, versatile, and\ngeneral -- it does not rely on any assumptions. As shown in various examples,\nit is able to efficiently and reliably optimize such WPT systems in order to\nfind their physical limitations on performance, optimal operating parameters\nand inspect their working principles, even for a large number of active\ntransmitters and passive elements.\n",
"title": "Optimization of Wireless Power Transfer Systems Enhanced by Passive Elements and Metasurfaces"
}
| null | null | null | null | true | null |
4359
| null |
Default
| null | null |
null |
{
"abstract": " This paper proposes a new sharpened version of the Jensen's inequality. The\nproposed new bound is simple and insightful, is broadly applicable by imposing\nminimum assumptions, and provides fairly accurate result in spite of its simple\nform. Applications to the moment generating function, power mean inequalities,\nand Rao-Blackwell estimation are presented. This presentation can be\nincorporated in any calculus-based statistical course.\n",
"title": "Sharpening Jensen's Inequality"
}
| null | null | null | null | true | null |
4360
| null |
Default
| null | null |
null |
{
"abstract": " Knotted solutions to electromagnetism and fluid dynamics are investigated,\nbased on relations we find between the two subjects. We can write fluid\ndynamics in electromagnetism language, but only on an initial surface, or for\nlinear perturbations, and we use this map to find knotted fluid solutions, as\nwell as new electromagnetic solutions. We find that knotted solutions of\nMaxwell electromagnetism are also solutions of more general nonlinear theories,\nlike Born-Infeld, and including ones which contain quantum corrections from\ncouplings with other modes, like Euler-Heisenberg and string theory DBI. Null\nconfigurations in electromagnetism can be described as a null pressureless\nfluid, and from this map we can find null fluid knotted solutions. A type of\nnonrelativistic reduction of the relativistic fluid equations is described,\nwhich allows us to find also solutions of the (nonrelativistic) Euler's\nequations.\n",
"title": "Knotted solutions, from electromagnetism to fluid dynamics"
}
| null | null | null | null | true | null |
4361
| null |
Default
| null | null |
null |
{
"abstract": " Let $m>1$ be an integer, and let $I(\\mathbb{Z}_m)^*$ be the set of all\nnon-zero proper ideals of $\\mathbb{Z}_m$. The intersection graph of ideals of\n$\\mathbb{Z}_m$, denoted by $G(\\mathbb{Z}_m)$, is a graph with vertices\n$I(\\mathbb{Z}_m)^*$ and two distinct vertices $I,J\\in I(\\mathbb{Z}_m)^*$ are\nadjacent if and only if $I\\cap J\\neq 0$. Let $n>1$ be an integer and\n$\\mathbb{Z}_n$ be a $\\mathbb{Z}_m$-module. In this paper, we introduce and\nstudy a kind of graph structure of $\\mathbb{Z}_m$, denoted by\n$G_n(\\mathbb{Z}_m)$. It is the undirected graph with the vertex set\n$I(\\mathbb{Z}_m)^*$, and two distinct vertices $I$ and $J$ are adjacent if and\nonly if $I\\mathbb{Z}_n\\cap J\\mathbb{Z}_n\\neq 0$. Clearly,\n$G_m(\\mathbb{Z}_m)=G(\\mathbb{Z}_m)$. We obtain some graph theoretical\nproperties of $G_n(\\mathbb{Z}_m)$ and we compute some of its numerical\ninvariants, namely girth, independence number, domination number, maximum\ndegree and chromatic index. We also determine all integer numbers $n$ and $m$\nfor which $G_n(\\mathbb{Z}_m)$ is Eulerian.\n",
"title": "On the intersection graph of ideals of $\\mathbb{Z}_m$"
}
| null | null | null | null | true | null |
4362
| null |
Default
| null | null |
null |
{
"abstract": " In recent publications, we presented a novel formal symbolic process virtual\nmachine (FSPVM) framework that combined higher-order theorem proving and\nsymbolic execution for verifying the reliability and security of smart\ncontracts developed in the Ethereum blockchain system without suffering the\nstandard issues surrounding reusability, consistency, and automation. A\nspecific FSPVM, denoted as FSPVM-E, was developed in Coq based on a general,\nextensible, and reusable formal memory (GERM) framework, an extensible and\nuniversal formal intermediate programming language, denoted as Lolisa, which is\na large subset of the Solidity programming language that uses generalized\nalgebraic datatypes, and a corresponding formally verified interpreter for\nLolisa, denoted as FEther, which serves as a crucial component of FSPVM-E.\nHowever, our past work has demonstrated that the execution efficiency of the\nstandard development of FEther is extremely low. As a result, FSPVM-E fails to\nachieve its expected verification effect. The present work addresses this issue\nby first identifying three root causes of the low execution efficiency of\nformal interpreters. We then build abstract models of these causes, and present\nrespective optimization schemes for rectifying the identified conditions.\nFinally, we apply these optimization schemes to FEther, and demonstrate that\nits execution efficiency has been improved significantly.\n",
"title": "Optimization of Executable Formal Interpreters developed in Higher-order Theorem Proving Systems"
}
| null | null | null | null | true | null |
4363
| null |
Default
| null | null |
null |
{
"abstract": " Robots have the potential to be a game changer in healthcare: improving\nhealth and well-being, filling care gaps, supporting care givers, and aiding\nhealth care workers. However, before robots are able to be widely deployed, it\nis crucial that both the research and industrial communities work together to\nestablish a strong evidence-base for healthcare robotics, and surmount likely\nadoption barriers. This article presents a broad contextualization of robots in\nhealthcare by identifying key stakeholders, care settings, and tasks; reviewing\nrecent advances in healthcare robotics; and outlining major challenges and\nopportunities to their adoption.\n",
"title": "Healthcare Robotics"
}
| null | null | null | null | true | null |
4364
| null |
Default
| null | null |
null |
{
"abstract": " Social learning, i.e., students learning from each other through social\ninteractions, has the potential to significantly scale up instruction in online\neducation. In many cases, such as in massive open online courses (MOOCs),\nsocial learning is facilitated through discussion forums hosted by course\nproviders. In this paper, we propose a probabilistic model for the process of\nlearners posting on such forums, using point processes. Different from existing\nworks, our method integrates topic modeling of the post text, timescale\nmodeling of the decay in post activity over time, and learner topic interest\nmodeling into a single model, and infers this information from user data. Our\nmethod also varies the excitation levels induced by posts according to the\nthread structure, to reflect typical notification settings in discussion\nforums. We experimentally validate the proposed model on three real-world MOOC\ndatasets, with the largest one containing up to 6,000 learners making 40,000\nposts in 5,000 threads. Results show that our model excels at thread\nrecommendation, achieving significant improvement over a number of baselines,\nthus showing promise of being able to direct learners to threads that they are\ninterested in more efficiently. Moreover, we demonstrate analytics that our\nmodel parameters can provide, such as the timescales of different topic\ncategories in a course.\n",
"title": "Personalized Thread Recommendation for MOOC Discussion Forums"
}
| null | null | null | null | true | null |
4365
| null |
Default
| null | null |
null |
{
"abstract": " Background: Widespread adoption of electronic health records (EHRs) has\nenabled secondary use of EHR data for clinical research and healthcare\ndelivery. Natural language processing (NLP) techniques have shown promise in\ntheir capability to extract the embedded information in unstructured clinical\ndata, and information retrieval (IR) techniques provide flexible and scalable\nsolutions that can augment the NLP systems for retrieving and ranking relevant\nrecords. Methods: In this paper, we present the implementation of Cohort\nRetrieval Enhanced by Analysis of Text from EHRs (CREATE), a cohort retrieval\nsystem that can execute textual cohort selection queries on both structured and\nunstructured EHR data. CREATE is a proof-of-concept system that leverages a\ncombination of structured queries and IR techniques on NLP results to improve\ncohort retrieval performance while adopting the Observational Medical Outcomes\nPartnership (OMOP) Common Data Model (CDM) to enhance model portability. The\nNLP component empowered by cTAKES is used to extract CDM concepts from textual\nqueries. We design a hierarchical index in Elasticsearch to support CDM concept\nsearch utilizing IR techniques and frameworks. Results: Our case study on 5\ncohort identification queries evaluated using the IR metric, P@5 (Precision at\n5) at both the patient-level and document-level, demonstrates that CREATE\nachieves an average P@5 of 0.90, which outperforms systems using only\nstructured data or only unstructured data with average P@5s of 0.54 and 0.74,\nrespectively.\n",
"title": "CREATE: Cohort Retrieval Enhanced by Analysis of Text from Electronic Health Records using OMOP Common Data Model"
}
| null | null | null | null | true | null |
4366
| null |
Default
| null | null |
null |
{
"abstract": " Piscine orthoreovirus Strain PRV-1 is the causative agent of heart and\nskeletal muscle inflammation (HSMI) in Atlantic salmon (Salmo salar). Given its\nhigh prevalence in net pen salmon, debate has arisen on whether PRV poses a\nrisk to migratory salmon, especially in British Columbia (BC) where\ncommercially important wild Pacific salmon are in decline. Various strains of\nPRV have been associated with diseases in Pacific salmon, including\nerythrocytic inclusion body syndrome (EIBS), HSMI-like disease, and\njaundice/anemia in Japan, Norway, Chile and Canada. We examine the\ndevelopmental pathway of HSMI and jaundice/anemia associated with PRV-1 in\nfarmed Atlantic and Chinook (Oncorhynchus tshawytscha) salmon in BC,\nrespectively. In situ hybridization localized PRV-1 within developing lesions\nin both diseases. The two diseases showed dissimilar pathological pathways,\nwith inflammatory lesions in heart and skeletal muscle in Atlantic salmon, and\ndegenerative-necrotic lesions in kidney and liver in Chinook salmon, plausibly\nexplained by differences in PRV load tolerance in red blood cells. Viral genome\nsequencing revealed no consistent differences in PRV-1 variants intimately\ninvolved in the development of both diseases, suggesting that migratory Chinook\nsalmon may be at more than a minimal risk of disease from exposure to the high\nlevels of PRV occurring on salmon farms.\n",
"title": "The same strain of Piscine orthoreovirus (PRV-1) is involved with the development of different, but related, diseases in Atlantic and Pacific Salmon in British Columbia"
}
| null | null | null | null | true | null |
4367
| null |
Default
| null | null |
null |
{
"abstract": " Motivated by recent experiments, we use the $+U$ extension of the generalized\ngradient approximation to density functional theory to study superlattices\ncomposed of alternating layers of LaNiO$_3$ and LaMnO$_3$. For comparison we\nalso study a rocksalt ((111) double perovskite) structure and bulk LaNiO$_3$\nand LaMnO$_3$. A Wannier function analysis indicates that band parameters are\ntransferable from bulk to superlattice situations with the exception of the\ntransition metal d-level energy, which has a contribution from the change in\nd-shell occupancy. The charge transfer from Mn to Ni is found to be moderate in\nthe superlattice, indicating metallic behavior, in contrast to the insulating\nbehavior found in recent experiments, while the rocksalt structure is found to\nbe insulating with a large Mn-Ni charge transfer. We suggest a high density of\ncation antisite defects may account for the insulating behavior experimentally\nobserved in short-period superlattices.\n",
"title": "Charge transfer and metallicity in LaNiO$_3$/LaMnO$_3$ superlattices"
}
| null | null | null | null | true | null |
4368
| null |
Default
| null | null |
null |
{
"abstract": " We consider learning-based variants of the $c \\mu$ rule for scheduling in\nsingle and parallel server settings of multi-class queueing systems.\nIn the single server setting, the $c \\mu$ rule is known to minimize the\nexpected holding-cost (weighted queue-lengths summed over classes and a fixed\ntime horizon). We focus on the problem where the service rates $\\mu$ are\nunknown with the holding-cost regret (regret against the $c \\mu$ rule with\nknown $\\mu$) as our objective. We show that the greedy algorithm that uses\nempirically learned service rates results in a constant holding-cost regret\n(the regret is independent of the time horizon). This free exploration can be\nexplained in the single server setting by the fact that any work-conserving\npolicy obtains the same number of samples in a busy cycle.\nIn the parallel server setting, we show that the $c \\mu$ rule may result in\nunstable queues, even for arrival rates within the capacity region. We then\npresent sufficient conditions for geometric ergodicity under the $c \\mu$ rule.\nUsing these results, we propose an almost greedy algorithm that explores only\nwhen the number of samples falls below a threshold. We show that this algorithm\ndelivers constant holding-cost regret because a free exploration condition is\neventually satisfied.\n",
"title": "On Learning the $cμ$ Rule in Single and Parallel Server Networks"
}
| null | null | null | null | true | null |
4369
| null |
Default
| null | null |
null |
{
"abstract": " In this report, some cosmological correlation functions are used to evaluate\nthe differential performance between C2075 and P100 GPU cards. In the past, the\ncorrelation functions used in this work have been widely studied and exploited\non some previous GPU architectures. The analysis of the performance indicates\nthat a speedup in the range from 13 to 15 is achieved without any additional\noptimization process for the P100 card.\n",
"title": "Report: Performance comparison between C2075 and P100 GPU cards using cosmological correlation functions"
}
| null | null | null | null | true | null |
4370
| null |
Default
| null | null |
null |
{
"abstract": " Computational paralinguistic analysis is increasingly being used in a wide\nrange of cyber applications, including security-sensitive applications such as\nspeaker verification, deceptive speech detection, and medical diagnostics.\nWhile state-of-the-art machine learning techniques, such as deep neural\nnetworks, can provide robust and accurate speech analysis, they are susceptible\nto adversarial attacks. In this work, we propose an end-to-end scheme to\ngenerate adversarial examples for computational paralinguistic applications by\nperturbing directly the raw waveform of an audio recording rather than specific\nacoustic features. Our experiments show that the proposed adversarial\nperturbation can lead to a significant performance drop of state-of-the-art\ndeep neural networks, while only minimally impairing the audio quality.\n",
"title": "Crafting Adversarial Examples For Speech Paralinguistics Applications"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
4371
| null |
Validated
| null | null |
null |
{
"abstract": " The exchange of small molecular signals within microbial populations is\ngenerally referred to as quorum sensing (QS). QS is ubiquitous in nature and\nenables microorganisms to respond to fluctuations of living environments by\nworking together. In this work, a QS-based communication system within a\nmicrobial population in a two-dimensional (2D) environment is analytically\nmodeled. Notably, the diffusion and degradation of signaling molecules within\nthe population is characterized. Microorganisms are randomly distributed on a\n2D circle where each one releases molecules at random times. The number of\nmolecules observed at each randomly-distributed bacterium is analyzed. Using\nthis analysis and some approximation, the expected density of cooperating\nbacteria is derived. The analytical results are validated via a particle-based\nsimulation method. The model can be used to predict and control behavioral\ndynamics of microscopic populations that have imperfect signal propagation.\n",
"title": "On the Analysis of Bacterial Cooperation with a Characterization of 2D Signal Propagation"
}
| null | null | null | null | true | null |
4372
| null |
Default
| null | null |
null |
{
"abstract": " We explore how the polarization around controversial topics evolves on\nTwitter - over a long period of time (2011 to 2016), and also as a response to\nmajor external events that lead to increased related activity. We find that\nincreased activity is typically associated with increased polarization;\nhowever, we find no consistent long-term trend in polarization over time among\nthe topics we study.\n",
"title": "The Ebb and Flow of Controversial Debates on Social Media"
}
| null | null | null | null | true | null |
4373
| null |
Default
| null | null |
null |
{
"abstract": " The vanishing gradient problem was a major obstacle for the success of deep\nlearning. In recent years it was gradually alleviated through multiple\ndifferent techniques. However the problem was not really overcome in a\nfundamental way, since it is inherent to neural networks with activation\nfunctions based on dot products. In a series of papers, we are going to analyze\nalternative neural network structures which are not based on dot products. In\nthis first paper, we revisit neural networks built up of layers based on\ndistance measures and Gaussian activation functions. These kinds of networks\nwere only sparsely used in the past since they are hard to train when using\nplain stochastic gradient descent methods. We show that by using Root Mean\nSquare Propagation (RMSProp) it is possible to efficiently learn multi-layer\nneural networks. Furthermore we show that when appropriately initialized these\nkinds of neural networks suffer much less from the vanishing and exploding\ngradient problem than traditional neural networks even for deep networks.\n",
"title": "Training of Deep Neural Networks based on Distance Measures using RMSProp"
}
| null | null | null | null | true | null |
4374
| null |
Default
| null | null |
null |
{
"abstract": " In high dimension, it is customary to consider Lasso-type estimators to\nenforce sparsity. For standard Lasso theory to hold, the regularization\nparameter should be proportional to the noise level, yet the latter is\ngenerally unknown in practice. A possible remedy is to consider estimators,\nsuch as the Concomitant/Scaled Lasso, which jointly optimize over the\nregression coefficients as well as over the noise level, making the choice of\nthe regularization independent of the noise level. However, when data from\ndifferent sources are pooled to increase sample size, or when dealing with\nmultimodal datasets, noise levels typically differ and new dedicated estimators\nare needed. In this work we provide new statistical and computational solutions\nto deal with such heteroscedastic regression models, with an emphasis on\nfunctional brain imaging with combined magneto- and electroencephalographic\n(M/EEG) signals. Adopting the formulation of Concomitant Lasso-type estimators,\nwe propose a jointly convex formulation to estimate both the regression\ncoefficients and the (square root of the) noise covariance. When our framework\nis instantiated to de-correlated noise, it leads to an efficient algorithm\nwhose computational cost is not higher than for the Lasso and Concomitant\nLasso, while addressing more complex noise structures. Numerical experiments\ndemonstrate that our estimator yields improved prediction and support\nidentification while correctly estimating the noise (square root) covariance.\nResults on multimodal neuroimaging problems with M/EEG data are also reported.\n",
"title": "Generalized Concomitant Multi-Task Lasso for sparse multimodal regression"
}
| null | null |
[
"Mathematics",
"Statistics"
] | null | true | null |
4375
| null |
Validated
| null | null |
null |
{
"abstract": " We propose a novel automatic method for accurate segmentation of the prostate\nin T2-weighted magnetic resonance imaging (MRI). Our method is based on\nconvolutional neural networks (CNNs). Because of the large variability in the\nshape, size, and appearance of the prostate and the scarcity of annotated\ntraining data, we suggest training two separate CNNs. A global CNN will\ndetermine a prostate bounding box, which is then resampled and sent to a local\nCNN for accurate delineation of the prostate boundary. This way, the local CNN\ncan effectively learn to segment the fine details that distinguish the prostate\nfrom the surrounding tissue using the small amount of available training data.\nTo fully exploit the training data, we synthesize additional data by deforming\nthe training images and segmentations using a learned shape model. We apply the\nproposed method on the PROMISE12 challenge dataset and achieve state of the art\nresults. Our proposed method generates accurate, smooth, and artifact-free\nsegmentations. On the test images, we achieve an average Dice score of 90.6\nwith a small standard deviation of 2.2, which is superior to all previous\nmethods. Our two-step segmentation approach and data augmentation strategy may\nbe highly effective in segmentation of other organs from small amounts of\nannotated medical images.\n",
"title": "A deep learning-based method for prostate segmentation in T2-weighted magnetic resonance imaging"
}
| null | null | null | null | true | null |
4376
| null |
Default
| null | null |
null |
{
"abstract": " Using theorems of Eliashberg and McDuff, Etnyre [Et] proved that the\nintersection form of a symplectic filling of a contact 3-manifold supported by\nplanar open book is negative definite.\nIn this paper, we prove a signature formula for allowable Lefschetz\nfibrations over $D^2$ with planar fiber by computing Maslov index appearing in\nWall's non-additivity formula.\nThe signature formula leads to an alternative proof of Etnyre's theorem via\nworks of Niederkrüger and Wendl [NWe] and Wendl [We].\nConversely, Etnyre's theorem, together with the existence theorem of Stein\nstructures on Lefschetz fibrations over $D^2$ with bordered fiber by Loi and\nPiergallini [LP], implies the formula.\n",
"title": "A note on signature of Lefschetz fibrations with planar fiber"
}
| null | null | null | null | true | null |
4377
| null |
Default
| null | null |
null |
{
"abstract": " Split-plot or repeated measures designs are frequently used for planning\nexperiments in the life or social sciences. Typical examples include the\ncomparison of different treatments over time, where both factors may possess an\nadditional factorial structure. For such designs, the statistical analysis\nusually consists of several steps. If the global null is rejected, multiple\ncomparisons are usually performed. Usually, general factorial repeated measures\ndesigns are inferred by classical linear mixed models. Common underlying\nassumptions, such as normality or variance homogeneity are often not met in\nreal data. Furthermore, to deal even with, e.g., ordinal or ordered categorical\ndata, adequate effect sizes should be used. Here, multiple contrast tests and\nsimultaneous confidence intervals for general factorial split-plot designs are\ndeveloped and equipped with a novel asymptotically correct wild bootstrap\napproach.\nBecause the regulatory authorities typically require the calculation of\nconfidence intervals, this work also provides simultaneous confidence intervals\nfor single contrasts and for the ratio of different contrasts in meaningful\neffects. Extensive simulations are conducted to foster the theoretical\nfindings. Finally, two different datasets exemplify the applicability of the\nnovel procedure.\n",
"title": "Wild Bootstrapping Rank-Based Procedures: Multiple Testing in Nonparametric Split-Plot Designs"
}
| null | null | null | null | true | null |
4378
| null |
Default
| null | null |
null |
{
"abstract": " The syntax of modal graphs is defined in terms of the continuous cut and\nbroken cut following Charles Peirce's notation in the gamma part of his\ngraphical logic of existential graphs. Graphical calculi for normal modal\nlogics are developed based on a reformulation of the graphical calculus for\nclassical propositional logic. These graphical calculi are of the nature of\ndeep inference. The relationship between graphical calculi and sequent calculi\nfor modal logics is shown by translations between graphs and modal formulas.\n",
"title": "Graphical Sequent Calculi for Modal Logics"
}
| null | null | null | null | true | null |
4379
| null |
Default
| null | null |
null |
{
"abstract": " There has been great interest in realizing quantum simulators of charged\nparticles in artificial gauge fields. Here, we perform the first quantum\nsimulation explorations of the combination of artificial gauge fields and\ndisorder. Using synthetic lattice techniques based on parametrically-coupled\natomic momentum states, we engineer zigzag chains with a tunable homogeneous\nflux. The breaking of time-reversal symmetry by the applied flux leads to\nanalogs of spin-orbit coupling and spin-momentum locking, which we observe\ndirectly through the chiral dynamics of atoms initialized to single lattice\nsites. We additionally introduce precisely controlled disorder in the site\nenergy landscape, allowing us to explore the interplay of disorder and large\neffective magnetic fields. The combination of correlated disorder and\ncontrolled intra- and inter-row tunneling in this system naturally supports\nenergy-dependent localization, relating to a single-particle mobility edge. We\nmeasure the localization properties of the extremal eigenstates of this system,\nthe ground state and the most-excited state, and demonstrate clear evidence for\na flux-dependent mobility edge. These measurements constitute the first direct\nevidence for energy-dependent localization in a lower-dimensional system, as\nwell as the first explorations of the combined influence of artificial gauge\nfields and engineered disorder. Moreover, we provide direct evidence for\ninteraction shifts of the localization transitions for both low- and\nhigh-energy eigenstates in correlated disorder, relating to the presence of a\nmany-body mobility edge. The unique combination of strong interactions,\ncontrolled disorder, and tunable artificial gauge fields present in this\nsynthetic lattice system should enable myriad explorations into intriguing\ncorrelated transport phenomena.\n",
"title": "Engineering a flux-dependent mobility edge in disordered zigzag chains"
}
| null | null | null | null | true | null |
4380
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we are interested in the decomposition of the tensor product\nof two representations of a symmetrizable Kac-Moody Lie algebra $\\mathfrak g$.\nLet $P\\_+$ be the set of dominant integral weights. For $\\lambda\\in P\\_+$ ,\n$L(\\lambda)$ denotes the irreducible, integrable, highest weight representation\nof g with highest weight $\\lambda$. Let $P\\_{+,\\mathbb Q}$ be the rational\nconvex cone generated by $P\\_+$. Consider the tensor cone $\\Gamma(\\mathfrak g)\n:= \\{(\\lambda\\_1 ,\\lambda\\_2, \\mu) $\\in$ P\\_{+,\\mathbb Q}^3\\,| \\exists N\n\\textgreater{} 1 L(N\\mu) \\subset L(N \\lambda\\_1)\\otimes L(N \\lambda\\_2)\\}$. If\n$\\mathfrak g$ is finite dimensional, $\\Gamma(\\mathfrak g)$ is a polyhedral\nconvex cone described in 2006 by Belkale-Kumar by an explicit finite list of\ninequalities. In general, $\\Gamma(\\mathfrak g)$ is nor polyhedral, nor closed.\nIn this article we describe the closure of $\\Gamma(\\mathfrak g)$ by an explicit\ncountable family of linear inequalities, when $\\mathfrak g$ is untwisted\naffine. This solves a Brown-Kumar's conjecture in this case. We also obtain\nexplicit saturation factors for the semigroup of triples $(\\lambda\\_1,\n\\lambda\\_2 , \\mu) $\\in$ P\\_+^3$ such that $L(\\mu) $\\subset$ L(\\lambda\\_1)\n\\otimes L(\\lambda\\_2)$. Note that even the existence of such saturation factors\nis not obvious since the semigroup is not finitely generated. For example, in\ntype $A , we prove that any integer $d\\geq 2$ is a saturation factor,\ngeneralizing the case ${\\tilde A}\\_1$ shown by Brown-Kumar.\n",
"title": "On the tensor semigroup of affine kac-moody lie algebras"
}
| null | null | null | null | true | null |
4381
| null |
Default
| null | null |
null |
{
"abstract": " Development of new greenhouse gas scavengers is actively pursued nowadays.\nVolatility caused solvent consumption and significant regeneration costs\nassociated with the aqueous amine solutions motivate search for more\ntechnologically and economically advanced solutions. We hereby used hybrid\ndensity functional theory to characterize thermodynamics, structure, electronic\nand solvation properties of amino and carboxamido functionalized C60 fullerene.\nC60 is non-volatile and supports a large density of amino groups on its\nsurface. Attachment of polar groups to fullerene C60 adjusts its dipole moment\nand band gap quite substantially, ultimately resulting in systematically better\nhydration thermodynamics. Reaction of polyaminofullerenes with CO2 is favored\nenthalpically, but prohibited entropically at standard conditions. Free energy\nof the CO2 capture by polyaminofullerenes is non-sensitive to the number of\namino groups per fullerene. This result fosters consideration of\npolyaminofullerenes for CO2 fixation.\n",
"title": "Electronic and Thermodynamic Properties of the Amino- and Carboxamido-Functionalized C-60-Based Fullerenes: Towards Non-Volatile Carbon Dioxide Scavengers"
}
| null | null | null | null | true | null |
4382
| null |
Default
| null | null |
null |
{
"abstract": " Singing voice separation based on deep learning relies on the usage of\ntime-frequency masking. In many cases the masking process is not a learnable\nfunction or is not encapsulated into the deep learning optimization.\nConsequently, most of the existing methods rely on a post processing step using\nthe generalized Wiener filtering. This work proposes a method that learns and\noptimizes (during training) a source-dependent mask and does not need the\naforementioned post processing step. We introduce a recurrent inference\nalgorithm, a sparse transformation step to improve the mask generation process,\nand a learned denoising filter. Obtained results show an increase of 0.49 dB\nfor the signal to distortion ratio and 0.30 dB for the signal to interference\nratio, compared to previous state-of-the-art approaches for monaural singing\nvoice separation.\n",
"title": "Monaural Singing Voice Separation with Skip-Filtering Connections and Recurrent Inference of Time-Frequency Mask"
}
| null | null | null | null | true | null |
4383
| null |
Default
| null | null |
null |
{
"abstract": " We study high-dimensional linear models with error-in-variables. Such models\nare motivated by various applications in econometrics, finance and genetics.\nThese models are challenging because of the need to account for measurement\nerrors to avoid non-vanishing biases in addition to handle the high\ndimensionality of the parameters. A recent growing literature has proposed\nvarious estimators that achieve good rates of convergence. Our main\ncontribution complements this literature with the construction of simultaneous\nconfidence regions for the parameters of interest in such high-dimensional\nlinear models with error-in-variables.\nThese confidence regions are based on the construction of moment conditions\nthat have an additional orthogonal property with respect to nuisance\nparameters. We provide a construction that requires us to estimate an\nadditional high-dimensional linear model with error-in-variables for each\ncomponent of interest. We use a multiplier bootstrap to compute critical values\nfor simultaneous confidence intervals for a subset $S$ of the components. We\nshow its validity despite of possible model selection mistakes, and allowing\nfor the cardinality of $S$ to be larger than the sample size.\nWe apply and discuss the implications of our results to two examples and\nconduct Monte Carlo simulations to illustrate the performance of the proposed\nprocedure.\n",
"title": "Confidence Bands for Coefficients in High Dimensional Linear Models with Error-in-variables"
}
| null | null | null | null | true | null |
4384
| null |
Default
| null | null |
null |
{
"abstract": " We show that given an estimate $\\widehat{A}$ that is close to a general\nhigh-rank positive semi-definite (PSD) matrix $A$ in spectral norm (i.e.,\n$\\|\\widehat{A}-A\\|_2 \\leq \\delta$), the simple truncated SVD of $\\widehat{A}$\nproduces a multiplicative approximation of $A$ in Frobenius norm. This\nobservation leads to many interesting results on general high-rank matrix\nestimation problems, which we briefly summarize below ($A$ is an $n\\times n$\nhigh-rank PSD matrix and $A_k$ is the best rank-$k$ approximation of $A$):\n(1) High-rank matrix completion: By observing\n$\\Omega(\\frac{n\\max\\{\\epsilon^{-4},k^2\\}\\mu_0^2\\|A\\|_F^2\\log\nn}{\\sigma_{k+1}(A)^2})$ elements of $A$ where $\\sigma_{k+1}\\left(A\\right)$ is\nthe $\\left(k+1\\right)$-th singular value of $A$ and $\\mu_0$ is the incoherence,\nthe truncated SVD on a zero-filled matrix satisfies $\\|\\widehat{A}_k-A\\|_F \\leq\n(1+O(\\epsilon))\\|A-A_k\\|_F$ with high probability.\n(2)High-rank matrix de-noising: Let $\\widehat{A}=A+E$ where $E$ is a Gaussian\nrandom noise matrix with zero mean and $\\nu^2/n$ variance on each entry. Then\nthe truncated SVD of $\\widehat{A}$ satisfies $\\|\\widehat{A}_k-A\\|_F \\leq\n(1+O(\\sqrt{\\nu/\\sigma_{k+1}(A)}))\\|A-A_k\\|_F + O(\\sqrt{k}\\nu)$.\n(3) Low-rank Estimation of high-dimensional covariance: Given $N$\ni.i.d.~samples $X_1,\\cdots,X_N\\sim\\mathcal N_n(0,A)$, can we estimate $A$ with\na relative-error Frobenius norm bound? We show that if $N =\n\\Omega\\left(n\\max\\{\\epsilon^{-4},k^2\\}\\gamma_k(A)^2\\log N\\right)$ for\n$\\gamma_k(A)=\\sigma_1(A)/\\sigma_{k+1}(A)$, then $\\|\\widehat{A}_k-A\\|_F \\leq\n(1+O(\\epsilon))\\|A-A_k\\|_F$ with high probability, where\n$\\widehat{A}=\\frac{1}{N}\\sum_{i=1}^N{X_iX_i^\\top}$ is the sample covariance.\n",
"title": "On the Power of Truncated SVD for General High-rank Matrix Estimation Problems"
}
| null | null | null | null | true | null |
4385
| null |
Default
| null | null |
null |
{
"abstract": " When a human drives a car along a road for the first time, they later\nrecognize where they are on the return journey typically without needing to\nlook in their rear-view mirror or turn around to look back, despite significant\nviewpoint and appearance change. Such navigation capabilities are typically\nattributed to our semantic visual understanding of the environment [1] beyond\ngeometry to recognizing the types of places we are passing through such as\n\"passing a shop on the left\" or \"moving through a forested area\". Humans are in\neffect using place categorization [2] to perform specific place recognition\neven when the viewpoint is 180 degrees reversed. Recent advances in deep neural\nnetworks have enabled high-performance semantic understanding of visual places\nand scenes, opening up the possibility of emulating what humans do. In this\nwork, we develop a novel methodology for using the semantics-aware higher-order\nlayers of deep neural networks for recognizing specific places from within a\nreference database. To further improve the robustness to appearance change, we\ndevelop a descriptor normalization scheme that builds on the success of\nnormalization schemes for pure appearance-based techniques such as SeqSLAM [3].\nUsing two different datasets - one road-based, one pedestrian-based, we\nevaluate the performance of the system in performing place recognition on\nreverse traversals of a route with a limited field of view camera and no\nturn-back-and-look behaviours, and compare to existing state-of-the-art\ntechniques and vanilla off-the-shelf features. The results demonstrate\nsignificant improvements over the existing state of the art, especially for\nextreme perceptual challenges that involve both great viewpoint change and\nenvironmental appearance change. We also provide experimental analyses of the\ncontributions of the various system components.\n",
"title": "Don't Look Back: Robustifying Place Categorization for Viewpoint- and Condition-Invariant Place Recognition"
}
| null | null | null | null | true | null |
4386
| null |
Default
| null | null |
null |
{
"abstract": " On-chip twisted light emitters are essential components for orbital angular\nmomentum (OAM) communication devices, which could address the growing demand\nfor high-capacity communication systems by providing an additional degree of\nfreedom for wavelength/frequency division multiplexing (WDM/FDM). Although\nwhispering gallery mode enabled OAM emitters have been shown to possess some\nadvantages, such as being compact and phase accurate, their inherent narrow\nbandwidth prevents them from being compatible with WDM/FDM techniques. Here, we\ndemonstrate an ultra-broadband multiplexed OAM emitter that utilizes a novel\njoint path-resonance phase control concept. The emitter has a micron sized\nradius and nanometer sized features. Coaxial OAM beams are emitted across the\nentire telecommunication band from 1450 to 1650 nm. We applied the emitter for\nOAM communication with a data rate of 1.2 Tbit/s assisted by 30-channel optical\nfrequency combs (OFC). The emitter provides a new solution to further increase\nof the capacity in the OFC communication scenario.\n",
"title": "Ultra-broadband On-chip Twisted Light Emitter"
}
| null | null |
[
"Physics"
] | null | true | null |
4387
| null |
Validated
| null | null |
null |
{
"abstract": " We introduce LAMP: the Linear Additive Markov Process. Transitions in LAMP\nmay be influenced by states visited in the distant history of the process, but\nunlike higher-order Markov processes, LAMP retains an efficient\nparametrization. LAMP also allows the specific dependence on history to be\nlearned efficiently from data. We characterize some theoretical properties of\nLAMP, including its steady-state and mixing time. We then give an algorithm\nbased on alternating minimization to learn LAMP models from data. Finally, we\nperform a series of real-world experiments to show that LAMP is more powerful\nthan first-order Markov processes, and even holds its own against deep\nsequential models (LSTMs) with a negligible increase in parameter complexity.\n",
"title": "Linear Additive Markov Processes"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
4388
| null |
Validated
| null | null |
null |
{
"abstract": " We consider extended starlike networks where the hub node is coupled with\nseveral chains of nodes representing star rays. Assuming that nodes of the\nnetwork are occupied by nonidentical self-oscillators we study various forms of\ntheir cluster synchronization. Radial cluster emerges when the nodes are\nsynchronized along a ray, while circular cluster is formed by nodes without\nimmediate connections but located on identical distances to the hub. By its\nnature the circular synchronization is a new manifestation of so called remote\nsynchronization [Phys. Rev. E 85 (2012), 026208]. We report its long-range form\nwhen the synchronized nodes interact through at least three intermediate nodes.\nForms of long-range remote synchronization are elements of scenario of\ntransition to the total synchronization of the network. We observe that the far\nends of rays synchronize first. Then more circular clusters appear involving\ncloser to hub nodes. Subsequently the clusters merge and, finally, all network\nbecome synchronous. Behavior of the extended starlike networks is found to be\nstrongly determined by the ray length, while varying the number of rays\nbasically affects fine details of a dynamical picture. Symmetry of the star\nalso extensively influences the dynamics. In an asymmetric star circular\ncluster mainly vanish in favor of radial ones, however, long-range remote\nsynchronization survives.\n",
"title": "Radial and circular synchronization clusters in extended starlike network of van der Pol oscillators"
}
| null | null | null | null | true | null |
4389
| null |
Default
| null | null |
null |
{
"abstract": " Let $a(n)$ be the Fourier coefficients of a holomorphic cusp form of weight\n$\\kappa=2n\\geqslant12$ for the full modular group and\n$A(x)=\\sum\\limits_{n\\leqslant x}a(n)$. In this paper, we establish an\nasymptotic formula of the fourth power moment of $A(x)$ and prove that\n\\begin{equation*}\n\\int_1^TA^4(x)\\mathrm{d}x=\\frac{3}{64\\kappa\\pi^4}s_{4;2}(\\tilde{a})\nT^{2\\kappa}+O\\big(T^{2\\kappa-\\delta_4+\\varepsilon}\\big) \\end{equation*} with\n$\\delta_4=1/8$, which improves the previous result.\n",
"title": "On the Fourth Power Moment of Fourier Coefficients of Cusp Form"
}
| null | null | null | null | true | null |
4390
| null |
Default
| null | null |
null |
{
"abstract": " There is general consensus that learning representations is useful for a\nvariety of reasons, e.g. efficient use of labeled data (semi-supervised\nlearning), transfer learning and understanding hidden structure of data.\nPopular techniques for representation learning include clustering, manifold\nlearning, kernel-learning, autoencoders, Boltzmann machines, etc.\nTo study the relative merits of these techniques, it's essential to formalize\nthe definition and goals of representation learning, so that they are all\nbecome instances of the same definition. This paper introduces such a formal\nframework that also formalizes the utility of learning the representation. It\nis related to previous Bayesian notions, but with some new twists. We show the\nusefulness of our framework by exhibiting simple and natural settings -- linear\nmixture models and loglinear models, where the power of representation learning\ncan be formally shown. In these examples, representation learning can be\nperformed provably and efficiently under plausible assumptions (despite being\nNP-hard), and furthermore: (i) it greatly reduces the need for labeled data\n(semi-supervised learning) and (ii) it allows solving classification tasks when\nsimpler approaches like nearest neighbors require too much data (iii) it is\nmore powerful than manifold learning methods.\n",
"title": "Provable benefits of representation learning"
}
| null | null | null | null | true | null |
4391
| null |
Default
| null | null |
null |
{
"abstract": " In this work we offer a framework for reasoning about a wide class of\nexisting objectives in machine learning. We develop a formal correspondence\nbetween this work and thermodynamics and discuss its implications.\n",
"title": "TherML: Thermodynamics of Machine Learning"
}
| null | null | null | null | true | null |
4392
| null |
Default
| null | null |
null |
{
"abstract": " With the advent of Big Data, nowadays in many applications databases\ncontaining large quantities of similar time series are available. Forecasting\ntime series in these domains with traditional univariate forecasting procedures\nleaves great potentials for producing accurate forecasts untapped. Recurrent\nneural networks (RNNs), and in particular Long Short-Term Memory (LSTM)\nnetworks, have proven recently that they are able to outperform\nstate-of-the-art univariate time series forecasting methods in this context\nwhen trained across all available time series. However, if the time series\ndatabase is heterogeneous, accuracy may degenerate, so that on the way towards\nfully automatic forecasting methods in this space, a notion of similarity\nbetween the time series needs to be built into the methods. To this end, we\npresent a prediction model that can be used with different types of RNN models\non subgroups of similar time series, which are identified by time series\nclustering techniques. We assess our proposed methodology using LSTM networks,\na widely popular RNN variant. Our method achieves competitive results on\nbenchmarking datasets under competition evaluation procedures. In particular,\nin terms of mean sMAPE accuracy, it consistently outperforms the baseline LSTM\nmodel and outperforms all other methods on the CIF2016 forecasting competition\ndataset.\n",
"title": "Forecasting Across Time Series Databases using Recurrent Neural Networks on Groups of Similar Series: A Clustering Approach"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
4393
| null |
Validated
| null | null |
null |
{
"abstract": " We introduce a new virtual environment for simulating a card game known as\n\"Big 2\". This is a four-player game of imperfect information with a relatively\ncomplicated action space (being allowed to play 1,2,3,4 or 5 card combinations\nfrom an initial starting hand of 13 cards). As such it poses a challenge for\nmany current reinforcement learning methods. We then use the recently proposed\n\"Proximal Policy Optimization\" algorithm to train a deep neural network to play\nthe game, purely learning via self-play, and find that it is able to reach a\nlevel which outperforms amateur human players after only a relatively short\namount of training time and without needing to search a tree of future game\nstates.\n",
"title": "Application of Self-Play Reinforcement Learning to a Four-Player Game of Imperfect Information"
}
| null | null | null | null | true | null |
4394
| null |
Default
| null | null |
null |
{
"abstract": " We present Fashion-MNIST, a new dataset comprising of 28x28 grayscale images\nof 70,000 fashion products from 10 categories, with 7,000 images per category.\nThe training set has 60,000 images and the test set has 10,000 images.\nFashion-MNIST is intended to serve as a direct drop-in replacement for the\noriginal MNIST dataset for benchmarking machine learning algorithms, as it\nshares the same image size, data format and the structure of training and\ntesting splits. The dataset is freely available at\nthis https URL\n",
"title": "Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms"
}
| null | null | null | null | true | null |
4395
| null |
Default
| null | null |
null |
{
"abstract": " A $k$-page book drawing of a graph $G=(V,E)$ consists of a linear ordering of\nits vertices along a spine and an assignment of each edge to one of the $k$\npages, which are half-planes bounded by the spine. In a book drawing, two edges\ncross if and only if they are assigned to the same page and their vertices\nalternate along the spine. Crossing minimization in a $k$-page book drawing is\nNP-hard, yet book drawings have multiple applications in visualization and\nbeyond. Therefore several heuristic book drawing algorithms exist, but there is\nno broader comparative study on their relative performance. In this paper, we\npropose a comprehensive benchmark set of challenging graph classes for book\ndrawing algorithms and provide an extensive experimental study of the\nperformance of existing book drawing algorithms.\n",
"title": "Experimental Evaluation of Book Drawing Algorithms"
}
| null | null | null | null | true | null |
4396
| null |
Default
| null | null |
null |
{
"abstract": " We have measured X-ray magnetic circular dichroism (XMCD) spectra at the Pu\n$M_{4,5}$ absorption edges from a newly-prepared high-quality single crystal of\nthe heavy fermion superconductor $^{242}$PuCoGa$_{5}$, exhibiting a critical\ntemperature $T_{c} = 18.7~{\\rm K}$. The experiment probes the vortex phase\nbelow $T_{c}$ and shows that an external magnetic field induces a Pu 5$f$\nmagnetic moment at 2 K equal to the temperature-independent moment measured in\nthe normal phase up to 300 K by a SQUID device. This observation is in\nagreement with theoretical models claiming that the Pu atoms in PuCoGa$_{5}$\nhave a nonmagnetic singlet ground state resulting from the hybridization of the\nconduction electrons with the intermediate-valence 5$f$ electronic shell.\nUnexpectedly, XMCD spectra show that the orbital component of the $5f$ magnetic\nmoment increases significantly between 30 and 2 K; the antiparallel spin\ncomponent increases as well, leaving the total moment practically constant. We\nsuggest that this indicates a low-temperature breakdown of the complete\nKondo-like screening of the local 5$f$ moment.\n",
"title": "Probing magnetism in the vortex phase of PuCoGa$_5$ by X-ray magnetic circular dichroism"
}
| null | null | null | null | true | null |
4397
| null |
Default
| null | null |
null |
{
"abstract": " This paper studies the problem of secure communication over a K-transmitter\nmultiple access channel in the presence of an external eavesdropper, subject to\na joint secrecy constraint (i.e., information leakage rate from the collection\nof K messages to an eavesdropper is made vanishing). As a result, we establish\nthe joint secrecy achievable rate region. To this end, our results build upon\ntwo techniques in addition to the standard information-theoretic methods. The\nfirst is a generalization of Chia-El Gamal's lemma on entropy bound for a set\nof codewords given partial information. The second is to utilize a compact\nrepresentation of a list of sets that, together with properties of mutual\ninformation, leads to an efficient Fourier-Motzkin elimination. These two\napproaches could also be of independent interests in other contexts.\n",
"title": "Joint secrecy over the K-Transmitter Multiple Access Channel"
}
| null | null | null | null | true | null |
4398
| null |
Default
| null | null |
null |
{
"abstract": " We analytically derive the elastic, dielectric, piezoelectric, and the\nflexoelectric phenomenological coefficients as functions of microscopic model\nparameters such as ionic positions and spring constants in the two-dimensional\nsquare-lattice model with rock-salt-type ionic arrangement. Monte-Carlo\nsimulation reveals that a difference in the given elastic constants of the\ndiagonal springs, each of which connects the same cations or anions, is\nresponsible for the linear flexoelectric effect in the model. We show the\nquadratic flexoelectric effect is present only in non-centrosymmetric systems\nand it can overwhelm the linear effect in feasibly large strain gradients.\n",
"title": "Nonlinear Flexoelectricity in Non-centrosymmetric Crystals"
}
| null | null | null | null | true | null |
4399
| null |
Default
| null | null |
null |
{
"abstract": " We are concerned with multidimensional nonlinear stochastic transport\nequation driven by Brownian motions. For irregular fluxes, by using stochastic\nBGK approximations and commutator estimates, we gain the existence and\nuniqueness of stochastic entropy solutions. Besides, for $BV$ initial data, the\n$BV$ and Hölder regularities are also derived for the unique stochastic\nentropy solution. Particularly, for the transport equation, we gain a\nregularization result, i.e. while the existence fails for the transport\nequation, we prove that a multiplicative stochastic perturbation of Brownian\ntype is enough to render the equation well-posed. This seems to be another\nexplicit example (the first example is given in [22]) of a PDE of fluid\ndynamics that becomes well-posed under the influence of a multiplicative\nBrownian type noise.\n",
"title": "Well-posedness of nonlinear transport equation by stochastic perturbation"
}
| null | null |
[
"Mathematics"
] | null | true | null |
4400
| null |
Validated
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.