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": " In this paper, we are concerned with deterministic wave generation in a\nhydrodynamic laboratory. A linear wavemaker theory is developed based on the\nfully dispersive water wave equations. The governing field equation is the\nLaplace equation for potential flow with several boundary conditions: the\ndynamic and kinematic boundary condition at the free surface, the lateral\nboundary condition at the wavemaker and the bottom boundary condition. In this\nwork, we consider both single-flap and double-flap wavemakers. The velocity\npotential and surface wave elevation are derived, and the relation between the\npropagating wave height and wavemaker stroke is formulated. This formulation is\nthen used to find how to operate the wavemaker in an efficient way to generate\nthe desired propagating waves with minimal disturbances near the wavemaker.\n",
"title": "Linear theory for single and double flap wavemakers"
}
| null | null | null | null | true | null |
17801
| null |
Default
| null | null |
null |
{
"abstract": " Dual energy CT (DECT) enhances tissue characterization because it can produce\nimages of basis materials such as soft-tissue and bone. DECT is of great\ninterest in applications to medical imaging, security inspection and\nnondestructive testing. Theoretically, two materials with different linear\nattenuation coefficients can be accurately reconstructed using DECT technique.\nHowever, the ability to reconstruct three or more basis materials is clinically\nand industrially important. Under the assumption that there are at most three\nmaterials in each pixel, there are a few methods that estimate multiple\nmaterial images from DECT measurements by enforcing sum-to-one and a box\nconstraint ([0 1]) derived from both the volume and mass conservation\nassumption. The recently proposed image-domain multi-material decomposition\n(MMD) method introduces edge-preserving regularization for each material image\nwhich neglects the relations among material images, and enforced the assumption\nthat there are at most three materials in each pixel using a time-consuming\nloop over all possible material-triplet in each iteration of optimizing its\ncost function. We propose a new image-domain MMD method for DECT that considers\nthe prior information that different material images have common edges and\nencourages sparsity of material composition in each pixel using regularization.\n",
"title": "Image-domain multi-material decomposition for dual-energy CT based on correlation and sparsity of material images"
}
| null | null | null | null | true | null |
17802
| null |
Default
| null | null |
null |
{
"abstract": " In wind farms, wake interaction leads to losses in power capture and\naccelerated structural degradation when compared to freestanding turbines. One\nmethod to reduce wake losses is by misaligning the rotor with the incoming flow\nusing its yaw actuator, thereby laterally deflecting the wake away from\ndownstream turbines. However, this demands an accurate and computationally\ntractable model of the wind farm dynamics. This problem calls for a closed-loop\nsolution. This tutorial paper fills the scientific gap by demonstrating the\nfull closed-loop controller synthesis cycle using a steady-state surrogate\nmodel. Furthermore, a novel, computationally efficient and modular\ncommunication interface is presented that enables researchers to\nstraight-forwardly test their control algorithms in large-eddy simulations.\nHigh-fidelity simulations of a 9-turbine farm show a power production increase\nof up to 11% using the proposed closed-loop controller compared to traditional,\ngreedy wind farm operation.\n",
"title": "A tutorial on the synthesis and validation of a closed-loop wind farm controller using a steady-state surrogate model"
}
| null | null |
[
"Computer Science"
] | null | true | null |
17803
| null |
Validated
| null | null |
null |
{
"abstract": " We show how an ensemble of $Q^*$-functions can be leveraged for more\neffective exploration in deep reinforcement learning. We build on well\nestablished algorithms from the bandit setting, and adapt them to the\n$Q$-learning setting. We propose an exploration strategy based on\nupper-confidence bounds (UCB). Our experiments show significant gains on the\nAtari benchmark.\n",
"title": "UCB Exploration via Q-Ensembles"
}
| null | null | null | null | true | null |
17804
| null |
Default
| null | null |
null |
{
"abstract": " Two modest-sized symbolic corpora of post-tonal and post-metric keyboard\nmusic have been constructed, one algorithmic, the other improvised. Deep\nlearning models of each have been trained and largely optimised. Our purpose is\nto obtain a model with sufficient generalisation capacity that in response to a\nsmall quantity of separate fresh input seed material, it can generate outputs\nthat are distinctive, rather than recreative of the learned corpora or the seed\nmaterial. This objective has been first assessed statistically, and as judged\nby k-sample Anderson-Darling and Cramer tests, has been achieved. Music has\nbeen generated using the approach, and informal judgements place it roughly on\na par with algorithmic and composed music in related forms. Future work will\naim to enhance the model such that it can be evaluated in relation to\nexpression, meaning and utility in real-time performance.\n",
"title": "Towards a Deep Improviser: a prototype deep learning post-tonal free music generator"
}
| null | null | null | null | true | null |
17805
| null |
Default
| null | null |
null |
{
"abstract": " Existing urban boundaries are usually defined by government agencies for\nadministrative, economic, and political purposes. Defining urban boundaries\nthat consider socio-economic relationships and citizen commute patterns is\nimportant for many aspects of urban and regional planning. In this paper, we\ndescribe a method to delineate urban boundaries based upon human interactions\nwith physical space inferred from social media. Specifically, we depicted the\nurban boundaries of Great Britain using a mobility network of Twitter user\nspatial interactions, which was inferred from over 69 million geo-located\ntweets. We define the non-administrative anthropographic boundaries in a\nhierarchical fashion based on different physical movement ranges of users\nderived from the collective mobility patterns of Twitter users in Great\nBritain. The results of strongly connected urban regions in the form of\ncommunities in the network space yield geographically cohesive, non-overlapping\nurban areas, which provide a clear delineation of the non-administrative\nanthropographic urban boundaries of Great Britain. The method was applied to\nboth national (Great Britain) and municipal scales (the London metropolis).\nWhile our results corresponded well with the administrative boundaries, many\nunexpected and interesting boundaries were identified. Importantly, as the\ndepicted urban boundaries exhibited a strong instance of spatial proximity, we\nemployed a gravity model to understand the distance decay effects in shaping\nthe delineated urban boundaries. The model explains how geographical distances\nfound in the mobility patterns affect the interaction intensity among different\nnon-administrative anthropographic urban areas, which provides new insights\ninto human spatial interactions with urban space.\n",
"title": "Depicting urban boundaries from a mobility network of spatial interactions: A case study of Great Britain with geo-located Twitter data"
}
| null | null | null | null | true | null |
17806
| null |
Default
| null | null |
null |
{
"abstract": " Peer review is the foundation of scientific publication, and the task of\nreviewing has long been seen as a cornerstone of professional service. However,\nthe massive growth in the field of machine learning has put this community\nbenefit under stress, threatening both the sustainability of an effective\nreview process and the overall progress of the field. In this position paper,\nwe argue that a tragedy of the commons outcome may be avoided by emphasizing\nthe professional aspects of this service. In particular, we propose a rubric to\nhold reviewers to an objective standard for review quality. In turn, we also\npropose that reviewers be given appropriate incentive. As one possible such\nincentive, we explore the idea of financial compensation on a per-review basis.\nWe suggest reasonable funding models and thoughts on long term effects.\n",
"title": "Avoiding a Tragedy of the Commons in the Peer Review Process"
}
| null | null | null | null | true | null |
17807
| null |
Default
| null | null |
null |
{
"abstract": " Generating random variates from high-dimensional distributions is often done\napproximately using Markov chain Monte Carlo. In certain cases, perfect\nsimulation algorithms exist that allow one to draw exactly from the stationary\ndistribution, but most require $O(n \\ln(n))$ time, where $n$ measures the size\nof the input. In this work a new protocol for creating perfect simulation\nalgorithms that runs in $O(n)$ time for a wider range of parameters on several\nmodels (such as Strauss, Ising, and random cluster) than was known previously.\nThis work represents an extension of the popping algorithms due to Wilson.\n",
"title": "Partially Recursive Acceptance Rejection"
}
| null | null | null | null | true | null |
17808
| null |
Default
| null | null |
null |
{
"abstract": " We show (under mild topological assumptions) that small oscillation of the\nunit normal vector implies Reifenberg flatness. We then apply this observation\nto the study of chord-arc domains and to a quantitative version of a two-phase\nfree boundary problem for harmonic measure previously studied by Kenig-Toro.\n",
"title": "Reifenberg Flatness and Oscillation of the Unit Normal Vector"
}
| null | null | null | null | true | null |
17809
| null |
Default
| null | null |
null |
{
"abstract": " We present a systematic method for designing distributed generation and\ndemand control schemes for secondary frequency regulation in power networks\nsuch that stability and an economically optimal power allocation can be\nguaranteed. A dissipativity condition is imposed on net power supply variables\nto provide stability guarantees. Furthermore, economic optimality is achieved\nby explicit decentralized steady state conditions on the generation and\ncontrollable demand. We discuss how various classes of dynamics used in recent\nstudies fit within our framework and give examples of higher order generation\nand controllable demand dynamics that can be included within our analysis. In\ncase of linear dynamics, we discuss how the proposed dissipativity condition\ncan be efficiently verified using an appropriate linear matrix inequality.\nMoreover, it is shown how the addition of a suitable observer layer can relax\nthe requirement for demand measurements in the employed controller. The\nefficiency and practicality of the proposed results are demonstrated with a\nsimulation on the Northeast Power Coordinating Council (NPCC) 140-bus system.\n",
"title": "Stability and optimality of distributed secondary frequency control schemes in power networks"
}
| null | null | null | null | true | null |
17810
| null |
Default
| null | null |
null |
{
"abstract": " Terrorism has become one of the most tedious problems to deal with and a\nprominent threat to mankind. To enhance counter-terrorism, several research\nworks are developing efficient and precise systems, data mining is not an\nexception. Immense data is floating in our lives, though the scarce\navailability of authentic terrorist attack data in the public domain makes it\ncomplicated to fight terrorism. This manuscript focuses on data mining\nclassification techniques and discusses the role of United Nations in\ncounter-terrorism. It analyzes the performance of classifiers such as Lazy\nTree, Multilayer Perceptron, Multiclass and Naïve Bayes classifiers for\nobserving the trends for terrorist attacks around the world. The database for\nexperiment purpose is created from different public and open access sources for\nyears 1970-2015 comprising of 156,772 reported attacks causing massive losses\nof lives and property. This work enumerates the losses occurred, trends in\nattack frequency and places more prone to it, by considering the attack\nresponsibilities taken as evaluation class.\n",
"title": "A Conjoint Application of Data Mining Techniques for Analysis of Global Terrorist Attacks -- Prevention and Prediction for Combating Terrorism"
}
| null | null | null | null | true | null |
17811
| null |
Default
| null | null |
null |
{
"abstract": " I present here the first results from an ongoing pilot project with the 1.6 m\ntelescope at the OPD, Brasil, aimed at the detection of the OVI $\\lambda$6830\nline via linear polarization in symbiotic stars. The main goal is to\ndemonstrate that OVI imaging polarimetry is an efficient technique for\ndiscovering new symbiotic stars. The OVI $\\lambda$6830 line is found in 5 out\nof 9 known symbiotic stars, in which the OVI line has already been\nspectroscopically confirmed, with at least 3-$\\sigma$ detection. Three new\nsymbiotic star candidates have also been found.\n",
"title": "OVI 6830Å Imaging Polarimetry of Symbiotic Stars"
}
| null | null | null | null | true | null |
17812
| null |
Default
| null | null |
null |
{
"abstract": " It is well-known that GANs are difficult to train, and several different\ntechniques have been proposed in order to stabilize their training. In this\npaper, we propose a novel training method called manifold-matching, and a new\nGAN model called manifold-matching GAN (MMGAN). MMGAN finds two manifolds\nrepresenting the vector representations of real and fake images. If these two\nmanifolds match, it means that real and fake images are statistically\nidentical. To assist the manifold-matching task, we also use i) kernel tricks\nto find better manifold structures, ii) moving-averaged manifolds across\nmini-batches, and iii) a regularizer based on correlation matrix to suppress\nmode collapse.\nWe conduct in-depth experiments with three image datasets and compare with\nseveral state-of-the-art GAN models. 32.4% of images generated by the proposed\nMMGAN are recognized as fake images during our user study (16% enhancement\ncompared to other state-of-the-art model). MMGAN achieved an unsupervised\ninception score of 7.8 for CIFAR-10.\n",
"title": "MMGAN: Manifold Matching Generative Adversarial Network"
}
| null | null |
[
"Computer Science"
] | null | true | null |
17813
| null |
Validated
| null | null |
null |
{
"abstract": " Acute respiratory infections have epidemic and pandemic potential and thus\nare being studied worldwide, albeit in many different contexts and study\nformats. Predicting infection from symptom data is critical, though using\nsymptom data from varied studies in aggregate is challenging because the data\nis collected in different ways. Accordingly, different symptom profiles could\nbe more predictive in certain studies, or even symptoms of the same name could\nhave different meanings in different contexts. We assess state-of-the-art\ntransfer learning methods for improving prediction of infection from symptom\ndata in multiple types of health care data ranging from clinical, to home-visit\nas well as crowdsourced studies. We show interesting characteristics regarding\nsix different study types and their feature domains. Further, we demonstrate\nthat it is possible to use data collected from one study to predict infection\nin another, at close to or better than using a single dataset for prediction on\nitself. We also investigate in which conditions specific transfer learning and\ndomain adaptation methods may perform better on symptom data. This work has the\npotential for broad applicability as we show how it is possible to transfer\nlearning from one public health study design to another, and data collected\nfrom one study may be used for prediction of labels for another, even collected\nthrough different study designs, populations and contexts.\n",
"title": "Domain Adaptation for Infection Prediction from Symptoms Based on Data from Different Study Designs and Contexts"
}
| null | null |
[
"Statistics",
"Quantitative Biology"
] | null | true | null |
17814
| null |
Validated
| null | null |
null |
{
"abstract": " Under model uncertainty, empirical Bayes (EB) procedures can have undesirable\nproperties such as extreme estimates of inclusion probabilities (Scott &\nBerger, 2010) or inconsistency under the null model (Liang et al., 2008). To\navoid these issues, we define empirical Bayes priors with constraints that\nensure that the estimates of the hyperparameters are at least as \"vague\" as\nthose of proper default priors. In our examples, we observe that constrained EB\nprocedures are better behaved than their unconstrained counterparts and that\nthe Bayesian Information Criterion (BIC) is similar to an intuitively appealing\nconstrained EB procedure.\n",
"title": "Constrained empirical Bayes priors on regression coefficients"
}
| null | null | null | null | true | null |
17815
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we introduce a finite field analogue for the Appell series F_3\nand give some reduction formulae and certain generating functions for this\nfunction over finite fields.\n",
"title": "A finite field analogue for Appell series F_3"
}
| null | null |
[
"Mathematics"
] | null | true | null |
17816
| null |
Validated
| null | null |
null |
{
"abstract": " In this article we consider two-way two-tape (alternating) automata accepting\npairs of words and we study some closure properties of this model. Our main\nresult is that such alternating automata are not closed under complementation\nfor non-unary alphabets. This improves a similar result of Kari and Moore for\npicture languages. We also show that these deterministic, non-deterministic and\nalternating automata are not closed under composition.\n",
"title": "Two-way Two-tape Automata"
}
| null | null | null | null | true | null |
17817
| null |
Default
| null | null |
null |
{
"abstract": " 360$^{\\circ}$ video requires human viewers to actively control \"where\" to\nlook while watching the video. Although it provides a more immersive experience\nof the visual content, it also introduces additional burden for viewers;\nawkward interfaces to navigate the video lead to suboptimal viewing\nexperiences. Virtual cinematography is an appealing direction to remedy these\nproblems, but conventional methods are limited to virtual environments or rely\non hand-crafted heuristics. We propose a new algorithm for virtual\ncinematography that automatically controls a virtual camera within a\n360$^{\\circ}$ video. Compared to the state of the art, our algorithm allows\nmore general camera control, avoids redundant outputs, and extracts its output\nvideos substantially more efficiently. Experimental results on over 7 hours of\nreal \"in the wild\" video show that our generalized camera control is crucial\nfor viewing 360$^{\\circ}$ video, while the proposed efficient algorithm is\nessential for making the generalized control computationally tractable.\n",
"title": "Making 360$^{\\circ}$ Video Watchable in 2D: Learning Videography for Click Free Viewing"
}
| null | null | null | null | true | null |
17818
| null |
Default
| null | null |
null |
{
"abstract": " Zero-shot learning (ZSL) is a challenging task aiming at recognizing novel\nclasses without any training instances. In this paper we present a simple but\nhigh-performance ZSL approach by generating pseudo feature representations\n(GPFR). Given the dataset of seen classes and side information of unseen\nclasses (e.g. attributes), we synthesize feature-level pseudo representations\nfor novel concepts, which allows us access to the formulation of unseen class\npredictor. Firstly we design a Joint Attribute Feature Extractor (JAFE) to\nacquire understandings about attributes, then construct a cognitive repository\nof attributes filtered by confidence margins, and finally generate pseudo\nfeature representations using a probability based sampling strategy to\nfacilitate subsequent training process of class predictor. We demonstrate the\neffectiveness in ZSL settings and the extensibility in supervised recognition\nscenario of our method on a synthetic colored MNIST dataset (C-MNIST). For\nseveral popular ZSL benchmark datasets, our approach also shows compelling\nresults on zero-shot recognition task, especially leading to tremendous\nimprovement to state-of-the-art mAP on zero-shot retrieval task.\n",
"title": "Zero-Shot Learning by Generating Pseudo Feature Representations"
}
| null | null | null | null | true | null |
17819
| null |
Default
| null | null |
null |
{
"abstract": " Speaker change detection (SCD) is an important task in dialog modeling. Our\npaper addresses the problem of text-based SCD, which differs from existing\naudio-based studies and is useful in various scenarios, for example, processing\ndialog transcripts where speaker identities are missing (e.g., OpenSubtitle),\nand enhancing audio SCD with textual information. We formulate text-based SCD\nas a matching problem of utterances before and after a certain decision point;\nwe propose a hierarchical recurrent neural network (RNN) with static\nsentence-level attention. Experimental results show that neural networks\nconsistently achieve better performance than feature-based approaches, and that\nour attention-based model significantly outperforms non-attention neural\nnetworks.\n",
"title": "Hierarchical RNN with Static Sentence-Level Attention for Text-Based Speaker Change Detection"
}
| null | null | null | null | true | null |
17820
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we develop a bivariate discrete generalized exponential\ndistribution, whose marginals are discrete generalized exponential distribution\nas proposed by Nekoukhou, Alamatsaz and Bidram (\"Discrete generalized\nexponential distribution of a second type\", Statistics, 47, 876 - 887, 2013).\nIt is observed that the proposed bivariate distribution is a very flexible\ndistribution and the bivariate geometric distribution can be obtained as a\nspecial case of this distribution. The proposed distribution can be seen as a\nnatural discrete analogue of the bivariate generalized exponential distribution\nproposed by Kundu and Gupta (\"Bivariate generalized exponential distribution\",\nJournal of Multivariate Analysis, 100, 581 - 593, 2009). We study different\nproperties of this distribution and explore its dependence structures. We\npropose a new EM algorithm to compute the maximum likelihood estimators of the\nunknown parameters which can be implemented very efficiently, and discuss some\ninferential issues also. The analysis of one data set has been performed to\nshow the effectiveness of the proposed model. Finally we propose some open\nproblems and conclude the paper.\n",
"title": "Bivariate Discrete Generalized Exponential Distribution"
}
| null | null | null | null | true | null |
17821
| null |
Default
| null | null |
null |
{
"abstract": " Over 150,000 new people in the United States are diagnosed with colorectal\ncancer each year. Nearly a third die from it (American Cancer Society). The\nonly approved noninvasive diagnosis tools currently involve fecal blood count\ntests (FOBTs) or stool DNA tests. Fecal blood count tests take only five\nminutes and are available over the counter for as low as \\$15. They are highly\nspecific, yet not nearly as sensitive, yielding a high percentage (25%) of\nfalse negatives (Colon Cancer Alliance). Moreover, FOBT results are far too\ngeneralized, meaning that a positive result could mean much more than just\ncolorectal cancer, and could just as easily mean hemorrhoids, anal fissure,\nproctitis, Crohn's disease, diverticulosis, ulcerative colitis, rectal ulcer,\nrectal prolapse, ischemic colitis, angiodysplasia, rectal trauma, proctitis\nfrom radiation therapy, and others. Stool DNA tests, the modern benchmark for\nCRC screening, have a much higher sensitivity and specificity, but also cost\n\\$600, take two weeks to process, and are not for high-risk individuals or\npeople with a history of polyps. To yield a cheap and effective CRC screening\nalternative, a unique ensemble-based classification algorithm is put in place\nthat considers the FIT result, BMI, smoking history, and diabetic status of\npatients. This method is tested under ten-fold cross validation to have a .95\nAUC, 92% specificity, 89% sensitivity, .88 F1, and 90% precision. Once\nclinically validated, this test promises to be cheaper, faster, and potentially\nmore accurate when compared to a stool DNA test.\n",
"title": "Optimization of Ensemble Supervised Learning Algorithms for Increased Sensitivity, Specificity, and AUC of Population-Based Colorectal Cancer Screenings"
}
| null | null | null | null | true | null |
17822
| null |
Default
| null | null |
null |
{
"abstract": " Users of Online Social Networks (OSNs) interact with each other more than\never. In the context of a public discussion group, people receive, read, and\nwrite comments in response to articles and postings. In the absence of access\ncontrol mechanisms, OSNs are a great environment for attackers to influence\nothers, from spreading phishing URLs, to posting fake news. Moreover, OSN user\nbehavior can be predicted by social science concepts which include conformity\nand the bandwagon effect. In this paper, we show how social recommendation\nsystems affect the occurrence of malicious URLs on Facebook. We exploit\ntemporal features to build a prediction framework, having greater than 75%\naccuracy, to predict whether the following group users' behavior will increase\nor not. Included in this work, we demarcate classes of URLs, including those\nmalicious URLs classified as creating critical damage, as well as those of a\nlesser nature which only inflict light damage such as aggressive commercial\nadvertisements and spam content. It is our hope that the data and analyses in\nthis paper provide a better understanding of OSN user reactions to different\ncategories of malicious URLs, thereby providing a way to mitigate the influence\nof these malicious URL attacks.\n",
"title": "More or Less? Predict the Social Influence of Malicious URLs on Social Media"
}
| null | null | null | null | true | null |
17823
| null |
Default
| null | null |
null |
{
"abstract": " We present an approach to path following using so-called control funnel\nfunctions. Synthesizing controllers to \"robustly\" follow a reference trajectory\nis a fundamental problem for autonomous vehicles. Robustness, in this context,\nrequires our controllers to handle a specified amount of deviation from the\ndesired trajectory. Our approach considers a timing law that describes how fast\nto move along a given reference trajectory and a control feedback law for\nreducing deviations from the reference. We synthesize both feedback laws using\n\"control funnel functions\" that jointly encode the control law as well as its\ncorrectness argument over a mathematical model of the vehicle dynamics. We\nadapt a previously described demonstration-based learning algorithm to\nsynthesize a control funnel function as well as the associated feedback law. We\nimplement this law on top of a 1/8th scale autonomous vehicle called the\nParkour car. We compare the performance of our path following approach against\na trajectory tracking approach by specifying trajectories of varying lengths\nand curvatures. Our experiments demonstrate the improved robustness obtained\nfrom the use of control funnel functions.\n",
"title": "Path-Following through Control Funnel Functions"
}
| null | null | null | null | true | null |
17824
| null |
Default
| null | null |
null |
{
"abstract": " We theoretically study scattering process and superconducting triplet\ncorrelations in a graphene junction comprised of ferromagnet-RSO-superconductor\nin which RSO stands for a region with Rashba spin orbit interaction. Our\nresults reveal spin-polarized subgap transport through the system due to an\nanomalous equal-spin Andreev reflection in addition to conventional back\nscatterings. We calculate equal- and opposite-spin pair correlations near the\nF-RSO interface and demonstrate direct link of the anomalous Andreev reflection\nand equal-spin pairings arised due to the proximity effect in the presence of\nRSO interaction. Moreover, we show that the amplitude of anomalous Andreev\nreflection, and thus the triplet pairings, are experimentally controllable when\nincorporating the influences of both tunable strain and Fermi level in the\nnonsuperconducting region. Our findings can be confirmed by a conductance\nspectroscopy experiment and provide better insights into the proximity-induced\nRSO coupling in graphene layers reported in recent experiments.\n",
"title": "Tunable Anomalous Andreev Reflection and Triplet Pairings in Spin Orbit Coupled Graphene"
}
| null | null | null | null | true | null |
17825
| null |
Default
| null | null |
null |
{
"abstract": " Model selection in mixed models based on the conditional distribution is\nappropriate for many practical applications and has been a focus of recent\nstatistical research. In this paper we introduce the R-package cAIC4 that\nallows for the computation of the conditional Akaike Information Criterion\n(cAIC). Computation of the conditional AIC needs to take into account the\nuncertainty of the random effects variance and is therefore not\nstraightforward. We introduce a fast and stable implementation for the\ncalculation of the cAIC for linear mixed models estimated with lme4 and\nadditive mixed models estimated with gamm4 . Furthermore, cAIC4 offers a\nstepwise function that allows for a fully automated stepwise selection scheme\nfor mixed models based on the conditional AIC. Examples of many possible\napplications are presented to illustrate the practical impact and easy handling\nof the package.\n",
"title": "Conditional Model Selection in Mixed-Effects Models with cAIC4"
}
| null | null | null | null | true | null |
17826
| null |
Default
| null | null |
null |
{
"abstract": " We study the ground state energy of the Neumann magnetic Laplacian on planar\ndomains. For a constant magnetic field we consider the question whether, under\nan assumption of fixed area, the disc maximizes this eigenvalue. More\ngenerally, we discuss old and new bounds obtained on this problem.\n",
"title": "Inequalities for the lowest magnetic Neumann eigenvalue"
}
| null | null | null | null | true | null |
17827
| null |
Default
| null | null |
null |
{
"abstract": " We develop and implement automated methods for optimizing quantum circuits of\nthe size and type expected in quantum computations that outperform classical\ncomputers. We show how to handle continuous gate parameters and report a\ncollection of fast algorithms capable of optimizing large-scale quantum\ncircuits. For the suite of benchmarks considered, we obtain substantial\nreductions in gate counts. In particular, we provide better optimization in\nsignificantly less time than previous approaches, while making minimal\nstructural changes so as to preserve the basic layout of the underlying quantum\nalgorithms. Our results help bridge the gap between the computations that can\nbe run on existing hardware and those that are expected to outperform classical\ncomputers.\n",
"title": "Automated optimization of large quantum circuits with continuous parameters"
}
| null | null | null | null | true | null |
17828
| null |
Default
| null | null |
null |
{
"abstract": " Testing the simplifying assumption in high-dimensional vine copulas is a\ndifficult task because tests must be based on estimated observations and amount\nto checking constraints on high-dimensional distributions. So far,\ncorresponding tests have been limited to single conditional copulas with a\nlow-dimensional set of conditioning variables. We propose a novel testing\nprocedure that is computationally feasible for high-dimensional data sets and\nthat exhibits a power that decreases only slightly with the dimension. By\ndiscretizing the support of the conditioning variables and incorporating a\npenalty in the test statistic, we mitigate the curse of dimensions by looking\nfor the possibly strongest deviation from the simplifying assumption. The use\nof a decision tree renders the test computationally feasible for large\ndimensions. We derive the asymptotic distribution of the test and analyze its\nfinite sample performance in an extensive simulation study. The utility of the\ntest is demonstrated by its application to 10 data sets with up to 49\ndimensions.\n",
"title": "Testing the simplifying assumption in high-dimensional vine copulas"
}
| null | null | null | null | true | null |
17829
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we study a natural special case of the Traveling Salesman\nProblem (TSP) with point-locational-uncertainty which we will call the {\\em\nadversarial TSP} problem (ATSP). Given a metric space $(X, d)$ and a set of\nsubsets $R = \\{R_1, R_2, ... , R_n\\} : R_i \\subseteq X$, the goal is to devise\nan ordering of the regions, $\\sigma_R$, that the tour will visit such that when\na single point is chosen from each region, the induced tour over those points\nin the ordering prescribed by $\\sigma_R$ is as short as possible. Unlike the\nclassical locational-uncertainty-TSP problem, which focuses on minimizing the\nexpected length of such a tour when the point within each region is chosen\naccording to some probability distribution, here, we focus on the {\\em\nadversarial model} in which once the choice of $\\sigma_R$ is announced, an\nadversary selects a point from each region in order to make the resulting tour\nas long as possible. In other words, we consider an offline problem in which\nthe goal is to determine an ordering of the regions $R$ that is optimal with\nrespect to the \"worst\" point possible within each region being chosen by an\nadversary, who knows the chosen ordering. We give a $3$-approximation when $R$\nis a set of arbitrary regions/sets of points in a metric space. We show how\ngeometry leads to improved constant factor approximations when regions are\nparallel line segments of the same lengths, and a polynomial-time approximation\nscheme (PTAS) for the important special case in which $R$ is a set of disjoint\nunit disks in the plane.\n",
"title": "TSP With Locational Uncertainty: The Adversarial Model"
}
| null | null | null | null | true | null |
17830
| null |
Default
| null | null |
null |
{
"abstract": " It is shown that using the similarity transformations, a set of\nthree-dimensional p-q nonlinear Schrodinger (NLS) equations with inhomogeneous\ncoefficients can be reduced to one-dimensional stationary NLS equation with\nconstant or varying coefficients, thus allowing for obtaining exact localized\nand periodic wave solutions. In the suggested reduction the original\ncoordinates in the (1+3)-space are mapped into a set of one-parametric\ncoordinate surfaces, whose parameter plays the role of the coordinate of the\none-dimensional equation. We describe the algorithm of finding solutions and\nconcentrate on power (linear and nonlinear) potentials presenting a number of\ncase examples. Generalizations of the method are also discussed.\n",
"title": "Exact solutions to three-dimensional generalized nonlinear Schrodinger equations with varying potential and nonlinearities"
}
| null | null | null | null | true | null |
17831
| null |
Default
| null | null |
null |
{
"abstract": " Recent studies have shown that reinforcement learning (RL) is an effective\napproach for improving the performance of neural machine translation (NMT)\nsystem. However, due to its instability, successfully RL training is\nchallenging, especially in real-world systems where deep models and large\ndatasets are leveraged. In this paper, taking several large-scale translation\ntasks as testbeds, we conduct a systematic study on how to train better NMT\nmodels using reinforcement learning. We provide a comprehensive comparison of\nseveral important factors (e.g., baseline reward, reward shaping) in RL\ntraining. Furthermore, to fill in the gap that it remains unclear whether RL is\nstill beneficial when monolingual data is used, we propose a new method to\nleverage RL to further boost the performance of NMT systems trained with\nsource/target monolingual data. By integrating all our findings, we obtain\ncompetitive results on WMT14 English- German, WMT17 English-Chinese, and WMT17\nChinese-English translation tasks, especially setting a state-of-the-art\nperformance on WMT17 Chinese-English translation task.\n",
"title": "A Study of Reinforcement Learning for Neural Machine Translation"
}
| null | null | null | null | true | null |
17832
| null |
Default
| null | null |
null |
{
"abstract": " Artificial neural network (ANN) is a very useful tool in solving learning\nproblems. Boosting the performances of ANN can be mainly concluded from two\naspects: optimizing the architecture of ANN and normalizing the raw data for\nANN. In this paper, a novel method which improves the effects of ANN by\npreprocessing the raw data is proposed. It totally leverages the fact that\ndifferent features should play different roles. The raw data set is firstly\npreprocessed by principle component analysis (PCA), and then its principle\ncomponents are weighted by their corresponding eigenvalues. Several aspects of\nanalysis are carried out to analyze its theory and the applicable occasions.\nThree classification problems are launched by an active learning algorithm to\nverify the proposed method. From the empirical results, conclusion comes to the\nfact that the proposed method can significantly improve the performance of ANN.\n",
"title": "Weighted Data Normalization Based on Eigenvalues for Artificial Neural Network Classification"
}
| null | null | null | null | true | null |
17833
| null |
Default
| null | null |
null |
{
"abstract": " Determinantal point processes (DPPs) have wide-ranging applications in\nmachine learning, where they are used to enforce the notion of diversity in\nsubset selection problems. Many estimators have been proposed, but surprisingly\nthe basic properties of the maximum likelihood estimator (MLE) have received\nlittle attention. The difficulty is that it is a non-concave maximization\nproblem, and such functions are notoriously difficult to understand in high\ndimensions, despite their importance in modern machine learning. Here we study\nboth the local and global geometry of the expected log-likelihood function. We\nprove several rates of convergence for the MLE and give a complete\ncharacterization of the case where these are parametric. We also exhibit a\npotential curse of dimensionality where the asymptotic variance of the MLE\nscales exponentially with the dimension of the problem. Moreover, we exhibit an\nexponential number of saddle points, and give evidence that these may be the\nonly critical points.\n",
"title": "Maximum likelihood estimation of determinantal point processes"
}
| null | null | null | null | true | null |
17834
| null |
Default
| null | null |
null |
{
"abstract": " Double machine learning provides $\\sqrt{n}$-consistent estimates of\nparameters of interest even when high-dimensional or nonparametric nuisance\nparameters are estimated at an $n^{-1/4}$ rate. The key is to employ\nNeyman-orthogonal moment equations which are first-order insensitive to\nperturbations in the nuisance parameters. We show that the $n^{-1/4}$\nrequirement can be improved to $n^{-1/(2k+2)}$ by employing a $k$-th order\nnotion of orthogonality that grants robustness to more complex or\nhigher-dimensional nuisance parameters. In the partially linear regression\nsetting popular in causal inference, we show that we can construct second-order\northogonal moments if and only if the treatment residual is not normally\ndistributed. Our proof relies on Stein's lemma and may be of independent\ninterest. We conclude by demonstrating the robustness benefits of an explicit\ndoubly-orthogonal estimation procedure for treatment effect.\n",
"title": "Orthogonal Machine Learning: Power and Limitations"
}
| null | null |
[
"Computer Science",
"Mathematics",
"Statistics"
] | null | true | null |
17835
| null |
Validated
| null | null |
null |
{
"abstract": " We perform direct numerical simulations of shock-wave/boundary-layer\ninteractions (SBLI) at Mach number M = 1.7 to investigate the influence of the\nstate of the incoming boundary layer on the interaction properties. We\nreproduce and extend the flow conditions of the experiments performed by\nGiepman et al., in which a spatially evolving laminar boundary layer over a\nflat plate is initially tripped by an array of distributed roughness elements\nand impinged further downstream by an oblique shock wave. Four SBLI cases are\nconsidered, based on two different shock impingement locations along the\nstreamwise direction, corresponding to transitional and turbulent interactions,\nand two different shock strengths, corresponding to flow deflection angles 3\ndegreees and 6 degrees. We find that, for all flow cases, shock induced\nseparation is not observed, the boundary layer remains attached for the 3\ndegrees case and close to incipient separation for the 6 degrees case,\nindependent of the state of the incoming boundary layer. The findings of this\nwork suggest that a transitional interaction might be the optimal solution for\npractical SBLI applications, as it removes the large separation bubble typical\nof laminar interactions and reduces the extent of the high-friction region\nassociated with an incoming turbulent boundary layer.\n",
"title": "Numerical investigation of supersonic shock-wave/boundary-layer interaction in transitional and turbulent regime"
}
| null | null | null | null | true | null |
17836
| null |
Default
| null | null |
null |
{
"abstract": " Latent factor models for recommender systems represent users and items as low\ndimensional vectors. Privacy risks of such systems have previously been studied\nmostly in the context of recovery of personal information in the form of usage\nrecords from the training data. However, the user representations themselves\nmay be used together with external data to recover private user information\nsuch as gender and age. In this paper we show that user vectors calculated by a\ncommon recommender system can be exploited in this way. We propose the\nprivacy-adversarial framework to eliminate such leakage of private information,\nand study the trade-off between recommender performance and leakage both\ntheoretically and empirically using a benchmark dataset. An advantage of the\nproposed method is that it also helps guarantee fairness of results, since all\nimplicit knowledge of a set of attributes is scrubbed from the representations\nused by the model, and thus can't enter into the decision making. We discuss\nfurther applications of this method towards the generation of deeper and more\ninsightful recommendations.\n",
"title": "Privacy and Fairness in Recommender Systems via Adversarial Training of User Representations"
}
| null | null | null | null | true | null |
17837
| null |
Default
| null | null |
null |
{
"abstract": " We establish large sample approximations for an arbitray number of bilinear\nforms of the sample variance-covariance matrix of a high-dimensional vector\ntime series using $ \\ell_1$-bounded and small $\\ell_2$-bounded weighting\nvectors. Estimation of the asymptotic covariance structure is also discussed.\nThe results hold true without any constraint on the dimension, the number of\nforms and the sample size or their ratios. Concrete and potential applications\nare widespread and cover high-dimensional data science problems such as tests\nfor large numbers of covariances, sparse portfolio optimization and projections\nonto sparse principal components or more general spanning sets as frequently\nconsidered, e.g. in classification and dictionary learning. As two specific\napplications of our results, we study in greater detail the asymptotics of the\ntrace functional and shrinkage estimation of covariance matrices. In shrinkage\nestimation, it turns out that the asymptotics differs for weighting vectors\nbounded away from orthogonaliy and nearly orthogonal ones in the sense that\ntheir inner product converges to 0.\n",
"title": "Asymptotics for high-dimensional covariance matrices and quadratic forms with applications to the trace functional and shrinkage"
}
| null | null | null | null | true | null |
17838
| null |
Default
| null | null |
null |
{
"abstract": " Word2vec (Mikolov et al., 2013) has proven to be successful in natural\nlanguage processing by capturing the semantic relationships between different\nwords. Built on top of single-word embeddings, paragraph vectors (Le and\nMikolov, 2014) find fixed-length representations for pieces of text with\narbitrary lengths, such as documents, paragraphs, and sentences. In this work,\nwe propose a novel interpretation for neural-network-based paragraph vectors by\ndeveloping an unsupervised generative model whose maximum likelihood solution\ncorresponds to traditional paragraph vectors. This probabilistic formulation\nallows us to go beyond point estimates of parameters and to perform Bayesian\nposterior inference. We find that the entropy of paragraph vectors decreases\nwith the length of documents, and that information about posterior uncertainty\nimproves performance in supervised learning tasks such as sentiment analysis\nand paraphrase detection.\n",
"title": "Bayesian Paragraph Vectors"
}
| null | null | null | null | true | null |
17839
| null |
Default
| null | null |
null |
{
"abstract": " We study extremal and algorithmic questions of subset and careful\nsynchronization in monotonic automata. We show that several synchronization\nproblems that are hard in general automata can be solved in polynomial time in\nmonotonic automata, even without knowing a linear order of the states preserved\nby the transitions. We provide asymptotically tight bounds on the maximum\nlength of a shortest word synchronizing a subset of states in a monotonic\nautomaton and a shortest word carefully synchronizing a partial monotonic\nautomaton. We provide a complexity framework for dealing with problems for\nmonotonic weakly acyclic automata over a three-letter alphabet, and use it to\nprove NP-completeness and inapproximability of problems such as {\\sc Finite\nAutomata Intersection} and the problem of computing the rank of a subset of\nstates in this class. We also show that checking whether a monotonic partial\nautomaton over a four-letter alphabet is carefully synchronizing is NP-hard.\nFinally, we give a simple necessary and sufficient condition when a strongly\nconnected digraph with a selected subset of vertices can be transformed into a\ndeterministic automaton where the corresponding subset of states is\nsynchronizing.\n",
"title": "Subset Synchronization in Monotonic Automata"
}
| null | null | null | null | true | null |
17840
| null |
Default
| null | null |
null |
{
"abstract": " The selection of West Java governor is one event that seizes the attention of\nthe public is no exception to social media users. Public opinion on a\nprospective regional leader can help predict electability and tendency of\nvoters. Data that can be used by the opinion mining process can be obtained\nfrom Twitter. Because the data is very varied form and very unstructured, it\nmust be managed and uninformed using data pre-processing techniques into\nsemi-structured data. This semi-structured information is followed by a\nclassification stage to categorize the opinion into negative or positive\nopinions. The research methodology uses a literature study where the research\nwill examine previous research on a similar topic. The purpose of this study is\nto find the right architecture to develop it into the application of twitter\nopinion mining to know public sentiments toward the election of the governor of\nwest java. The result of this research is that Twitter opinion mining is part\nof text mining where opinions in Twitter if they want to be classified, must go\nthrough the preprocessing text stage first. The preprocessing step required\nfrom twitter data is cleansing, case folding, POS Tagging and stemming. The\nresulting text mining architecture is an architecture that can be used for text\nmining research with different topics.\n",
"title": "Architecture of Text Mining Application in Analyzing Public Sentiments of West Java Governor Election using Naive Bayes Classification"
}
| null | null |
[
"Computer Science"
] | null | true | null |
17841
| null |
Validated
| null | null |
null |
{
"abstract": " The next generation of AI applications will continuously interact with the\nenvironment and learn from these interactions. These applications impose new\nand demanding systems requirements, both in terms of performance and\nflexibility. In this paper, we consider these requirements and present Ray---a\ndistributed system to address them. Ray implements a unified interface that can\nexpress both task-parallel and actor-based computations, supported by a single\ndynamic execution engine. To meet the performance requirements, Ray employs a\ndistributed scheduler and a distributed and fault-tolerant store to manage the\nsystem's control state. In our experiments, we demonstrate scaling beyond 1.8\nmillion tasks per second and better performance than existing specialized\nsystems for several challenging reinforcement learning applications.\n",
"title": "Ray: A Distributed Framework for Emerging AI Applications"
}
| null | null | null | null | true | null |
17842
| null |
Default
| null | null |
null |
{
"abstract": " Temporary work is an employment situation useful and suitable in all\noccasions in which business needs to adjust more easily and quickly to workload\nfluctuations or maintain staffing flexibility. Temporary workers play therefore\nan important role in many companies, but this kind of activity is subject to a\nspecial form of legal protections and many aspects and risks must be taken into\naccount both employers and employees. In this work we propose a\nblockchain-based system that aims to ensure respect for the rights for all\nactors involved in a temporary employment, in order to provide employees with\nthe fair and legal remuneration (including taxes) of work performances and a\nprotection in the case employer becomes insolvent. At the same time, our system\nwants to assist the employer in processing contracts with a fully automated and\nfast procedure. To resolve these problems we propose the D-ES (Decentralized\nEmployment System). We first model the employment relationship as a state\nsystem. Then we describe the enabling technology that makes us able to realize\nthe D-ES. In facts, we propose the implementation of a DLT (Decentralized\nLedger Technology) based system, consisting in a blockchain system and of a\nweb-based environment. Thanks the decentralized application platforms that\nmakes us able to develop smart contracts, we define a discrete event control\nsystem that works inside the blockchain. In addition, we discuss the temporary\nwork in agriculture as a interesting case of study.\n",
"title": "A blockchain-based Decentralized System for proper handling of temporary Employment contracts"
}
| null | null |
[
"Computer Science"
] | null | true | null |
17843
| null |
Validated
| null | null |
null |
{
"abstract": " Transition metal dichalcogenides represent an ideal testbed to study\nexcitonic effects, spin-related phenomena and fundamental light-matter coupling\nin nanoscopic condensed matter systems. In particular, the valley degree of\nfreedom, which is unique to such direct band gap monolayers with broken\ninversion symmetry, adds fundamental interest in these materials. Here, we\nimplement a Tamm-plasmon structure with an embedded MoSe2 monolayer and study\nthe formation of polaritonic quasi-particles. Strong coupling conditions\nbetween the Tamm-mode and the trion resonance of MoSe2 are established,\nyielding bright luminescence from the polaritonic ground state under\nnon-resonant optical excitation. We demonstrate, that tailoring the\nelectrodynamic environment of the monolayer results in a significantly\nincreased valley polarization. This enhancement can be related to change in\nrecombination dynamics shown in time-resolved photoluminescence measurements.\nWe furthermore observe strong upconversion luminescence from resonantly excited\npolariton states in the lower polariton branch. This upconverted polariton\nluminescence is shown to preserve the valley polarization of the\ntrion-polariton, which paves the way towards combining spin-valley physics and\nexciton scattering experiments.\n",
"title": "Valley polarized relaxation and upconversion luminescence from Tamm-Plasmon Trion-Polaritons with a MoSe2 monolayer"
}
| null | null | null | null | true | null |
17844
| null |
Default
| null | null |
null |
{
"abstract": " We demonstrate identification of position, material, orientation and shape of\nobjects imaged by an $^{85}$Rb atomic magnetometer performing electromagnetic\ninduction imaging supported by machine learning. Machine learning maximizes the\ninformation extracted from the images created by the magnetometer,\ndemonstrating the use of hidden data. Localization 2.6 times better than the\nspatial resolution of the imaging system and successful classification up to\n97$\\%$ are obtained. This circumvents the need of solving the inverse problem,\nand demonstrates the extension of machine learning to diffusive systems such as\nlow-frequency electrodynamics in media. Automated collection of task-relevant\ninformation from quantum-based electromagnetic imaging will have a relevant\nimpact from biomedicine to security.\n",
"title": "Machine learning based localization and classification with atomic magnetometers"
}
| null | null |
[
"Physics"
] | null | true | null |
17845
| null |
Validated
| null | null |
null |
{
"abstract": " We prove that the set of symplectic lattices in the Siegel space\n$\\mathfrak{h}_g$ whose systoles generate a subspace of dimension at least 3 in\n$\\mathbb{R}^{2g}$ does not contain any $\\mathrm{Sp}(2g,\\mathbb{Z})$-equivariant\ndeformation retract of $\\mathfrak{h}_g$.\n",
"title": "On the difficulty of finding spines"
}
| null | null | null | null | true | null |
17846
| null |
Default
| null | null |
null |
{
"abstract": " A Viterbi-like decoding algorithm is proposed in this paper for generalized\nconvolutional network error correction coding. Different from classical Viterbi\nalgorithm, our decoding algorithm is based on minimum error weight rather than\nthe shortest Hamming distance between received and sent sequences. Network\nerrors may disperse or neutralize due to network transmission and convolutional\nnetwork coding. Therefore, classical decoding algorithm cannot be employed any\nmore. Source decoding was proposed by multiplying the inverse of network\ntransmission matrix, where the inverse is hard to compute. Starting from the\nMaximum A Posteriori (MAP) decoding criterion, we find that it is equivalent to\nthe minimum error weight under our model. Inspired by Viterbi algorithm, we\npropose a Viterbi-like decoding algorithm based on minimum error weight of\ncombined error vectors, which can be carried out directly at sink nodes and can\ncorrect any network errors within the capability of convolutional network error\ncorrection codes (CNECC). Under certain situations, the proposed algorithm can\nrealize the distributed decoding of CNECC.\n",
"title": "Distributed Decoding of Convolutional Network Error Correction Codes"
}
| null | null |
[
"Computer Science"
] | null | true | null |
17847
| null |
Validated
| null | null |
null |
{
"abstract": " In multiferroic BiFeO$_3$ a cycloidal antiferromagnetic structure is coupled\nto a large electric polarization at room temperature, giving rise to\nmagnetoelectric functionality that may be exploited in novel multiferroic-based\ndevices. In this paper, we demonstrate that by substituting samarium for 10% of\nthe bismuth ions the periodicity of the room temperature cycloid is increased,\nand by cooling below $\\sim15$ K the magnetic structure tends towards a simple\nG-type antiferromagnet, which is fully established at 1.5 K. We show that this\ntransition results from $f-d$ exchange coupling, which induces a local\nanisotropy on the iron magnetic moments that destroys the cycloidal order - a\nresult of general significance regarding the stability of non-collinear\nmagnetic structures in the presence of multiple magnetic sublattices.\n",
"title": "Temperature induced phase transition from cycloidal to collinear antiferromagnetism in multiferroic Bi$_{0.9}$Sm$_{0.1}$FeO$_3$ driven by $f$-$d$ induced magnetic anisotropy"
}
| null | null | null | null | true | null |
17848
| null |
Default
| null | null |
null |
{
"abstract": " This project proposes to reuse the DAFNE accelerator complex for producing a\nhigh intensity (up to 10^10), high-quality beam of high-energy (up to 500 MeV)\npositrons for HEP experiments, mainly - but not only - motivated by light dark\nparticles searches. Such a facility would provide a unique source of\nultra-relativistic, narrow-band and low-emittance positrons, with a high duty\nfactor, without employing a cold technology, that would be an ideal facility\nfor exploring the existence of light dark matter particles, produced in\npositron-on-target annihilations into a photon+missing mass, and using the\nbump-hunt technique. The PADME experiment, that will use the extracted beam\nfrom the DAFNE BTF, is indeed limited by the low duty-factor (10^-5=200 ns/20\nms). The idea is to use a variant of the third of integer resonant extraction,\nwith the aim of getting a <10^-6 m rad emittance and, at the same time,\ntailoring the scheme to the peculiar optics of the DAFNE machine. In\nalternative, the possibility of kicking the positrons by means of channelling\neffects in crystals can be evaluated. This would not only increase the\nextraction efficiency but also improve the beam quality, thanks to the high\ncollimation of channelled particles.\n",
"title": "POSEYDON - Converting the DAFNE Collider into a double Positron Facility: a High Duty-Cycle pulse stretcher and a storage ring"
}
| null | null | null | null | true | null |
17849
| null |
Default
| null | null |
null |
{
"abstract": " Understanding planetary interiors is directly linked to our ability of\nsimulating exotic quantum mechanical systems such as hydrogen (H) and\nhydrogen-helium (H-He) mixtures at high pressures and temperatures. Equations\nof State (EOSs) tables based on Density Functional Theory (DFT), are commonly\nused by planetary scientists, although this method allows only for a\nqualitative description of the phase diagram, due to an incomplete treatment of\nelectronic interactions. Here we report Quantum Monte Carlo (QMC) molecular\ndynamics simulations of pure H and H-He mixture. We calculate the first QMC EOS\nat 6000 K for an H-He mixture of a proto-solar composition, and show the\ncrucial influence of He on the H metallization pressure. Our results can be\nused to calibrate other EOS calculations and are very timely given the accurate\ndetermination of Jupiter's gravitational field from the NASA Juno mission and\nthe effort to determine its structure.\n",
"title": "Phase diagram of hydrogen and a hydrogen-helium mixture at planetary conditions by Quantum Monte Carlo simulations"
}
| null | null | null | null | true | null |
17850
| null |
Default
| null | null |
null |
{
"abstract": " The occurrence of new events in a system is typically driven by external\ncauses and by previous events taking place inside the system. This is a general\nstatement, applying to a range of situations including, more recently, to the\nactivity of users in Online social networks (OSNs). Here we develop a method\nfor extracting from a series of posting times the relative contributions of\nexogenous, e.g. news media, and endogenous, e.g. information cascade. The\nmethod is based on the fitting of a generalized linear model (GLM) equipped\nwith a self-excitation mechanism. We test the method with synthetic data\ngenerated by a nonlinear Hawkes process, and apply it to a real time series of\ntweets with a given hashtag. In the empirical dataset, the estimated\ncontributions of exogenous and endogenous volumes are close to the amounts of\noriginal tweets and retweets respectively. We conclude by discussing the\npossible applications of the method, for instance in online marketing.\n",
"title": "Identifying exogenous and endogenous activity in social media"
}
| null | null |
[
"Computer Science"
] | null | true | null |
17851
| null |
Validated
| null | null |
null |
{
"abstract": " In this work the conduction of ion-water solution through two discrete\nbundles of armchair carbon and silicon carbide nanotubes, as useful membranes\nfor water desalination, is studied. In order that studies on different types of\nnanotubes be comparable, the chiral vectors of C and Si-C nanotubes are\nselected as (7,7) and (5,5), respectively, so that a similar volume of fluid is\ninvestigated flowing through two similar dimension membranes. Different\nhydrostatic pressures are applied and the flow rates of water and ions are\ncalculated through molecular dynamics simulations. Consequently, according to\nconductance of water per each nanotube, per nanosecond, it is perceived that at\nlower pressures (below 150 MPa) the Si-C nanotubes seem to be more applicable,\nwhile higher hydrostatic pressures make carbon nanotube membranes more suitable\nfor water desalination.\n",
"title": "Water flow in Carbon and Silicon Carbide nanotubes"
}
| null | null | null | null | true | null |
17852
| null |
Default
| null | null |
null |
{
"abstract": " Extreme values modeling has attracting the attention of researchers in\ndiverse areas such as the environment, engineering, or finance. Multivariate\nextreme value distributions are particularly suitable to model the tails of\nmultidimensional phenomena. The analysis of the dependence among multivariate\nmaxima is useful to evaluate risk. Here we present new multivariate extreme\nvalue models, as well as, coefficients to assess multivariate extremal\ndependence.\n",
"title": "Multidimensional extremal dependence coefficients"
}
| null | null | null | null | true | null |
17853
| null |
Default
| null | null |
null |
{
"abstract": " Systems subject to uncertain inputs produce uncertain responses. Uncertainty\nquantification (UQ) deals with the estimation of statistics of the system\nresponse, given a computational model of the system and a probabilistic model\nof its inputs. In engineering applications it is common to assume that the\ninputs are mutually independent or coupled by a Gaussian or elliptical\ndependence structure (copula). In this paper we overcome such limitations by\nmodelling the dependence structure of multivariate inputs as vine copulas. Vine\ncopulas are models of multivariate dependence built from simpler pair-copulas.\nThe vine representation is flexible enough to capture complex dependencies.\nThis paper formalises the framework needed to build vine copula models of\nmultivariate inputs and to combine them with virtually any UQ method. The\nframework allows for a fully automated, data-driven inference of the\nprobabilistic input model on available input data. The procedure is exemplified\non two finite element models of truss structures, both subject to inputs with\nnon-Gaussian dependence structures. For each case, we analyse the moments of\nthe model response (using polynomial chaos expansions), and perform a\nstructural reliability analysis to calculate the probability of failure of the\nsystem (using the first order reliability method and importance sampling).\nReference solutions are obtained by Monte Carlo simulation. The results show\nthat, while the Gaussian assumption yields biased statistics, the vine copula\nrepresentation achieves significantly more precise estimates, even when its\nstructure needs to be fully inferred from a limited amount of observations.\n",
"title": "A general framework for data-driven uncertainty quantification under complex input dependencies using vine copulas"
}
| null | null | null | null | true | null |
17854
| null |
Default
| null | null |
null |
{
"abstract": " The combination of large open data sources with machine learning approaches\npresents a potentially powerful way to predict events such as protest or social\nunrest. However, accounting for uncertainty in such models, particularly when\nusing diverse, unstructured datasets such as social media, is essential to\nguarantee the appropriate use of such methods. Here we develop a Bayesian\nmethod for predicting social unrest events in Australia using social media\ndata. This method uses machine learning methods to classify individual postings\nto social media as being relevant, and an empirical Bayesian approach to\ncalculate posterior event probabilities. We use the method to predict events in\nAustralian cities over a period in 2017/18.\n",
"title": "Pachinko Prediction: A Bayesian method for event prediction from social media data"
}
| null | null | null | null | true | null |
17855
| null |
Default
| null | null |
null |
{
"abstract": " The discrete cosine transform (DCT) is the key step in many image and video\ncoding standards. The 8-point DCT is an important special case, possessing\nseveral low-complexity approximations widely investigated. However, 16-point\nDCT transform has energy compaction advantages. In this sense, this paper\npresents a new 16-point DCT approximation with null multiplicative complexity.\nThe proposed transform matrix is orthogonal and contains only zeros and ones.\nThe proposed transform outperforms the well-know Walsh-Hadamard transform and\nthe current state-of-the-art 16-point approximation. A fast algorithm for the\nproposed transform is also introduced. This fast algorithm is experimentally\nvalidated using hardware implementations that are physically realized and\nverified on a 40 nm CMOS Xilinx Virtex-6 XC6VLX240T FPGA chip for a maximum\nclock rate of 342 MHz. Rapid prototypes on FPGA for 8-bit input word size shows\nsignificant improvement in compressed image quality by up to 1-2 dB at the cost\nof only eight adders compared to the state-of-art 16-point DCT approximation\nalgorithm in the literature [S. Bouguezel, M. O. Ahmad, and M. N. S. Swamy. A\nnovel transform for image compression. In {\\em Proceedings of the 53rd IEEE\nInternational Midwest Symposium on Circuits and Systems (MWSCAS)}, 2010].\n",
"title": "A Digital Hardware Fast Algorithm and FPGA-based Prototype for a Novel 16-point Approximate DCT for Image Compression Applications"
}
| null | null | null | null | true | null |
17856
| null |
Default
| null | null |
null |
{
"abstract": " It is well known that the initialization of weights in deep neural networks\ncan have a dramatic impact on learning speed. For example, ensuring the mean\nsquared singular value of a network's input-output Jacobian is $O(1)$ is\nessential for avoiding the exponential vanishing or explosion of gradients. The\nstronger condition that all singular values of the Jacobian concentrate near\n$1$ is a property known as dynamical isometry. For deep linear networks,\ndynamical isometry can be achieved through orthogonal weight initialization and\nhas been shown to dramatically speed up learning; however, it has remained\nunclear how to extend these results to the nonlinear setting. We address this\nquestion by employing powerful tools from free probability theory to compute\nanalytically the entire singular value distribution of a deep network's\ninput-output Jacobian. We explore the dependence of the singular value\ndistribution on the depth of the network, the weight initialization, and the\nchoice of nonlinearity. Intriguingly, we find that ReLU networks are incapable\nof dynamical isometry. On the other hand, sigmoidal networks can achieve\nisometry, but only with orthogonal weight initialization. Moreover, we\ndemonstrate empirically that deep nonlinear networks achieving dynamical\nisometry learn orders of magnitude faster than networks that do not. Indeed, we\nshow that properly-initialized deep sigmoidal networks consistently outperform\ndeep ReLU networks. Overall, our analysis reveals that controlling the entire\ndistribution of Jacobian singular values is an important design consideration\nin deep learning.\n",
"title": "Resurrecting the sigmoid in deep learning through dynamical isometry: theory and practice"
}
| null | null | null | null | true | null |
17857
| null |
Default
| null | null |
null |
{
"abstract": " Boundary value problem for complete second order elliptic equation is\nconsidered in Banach space. The equation and boundary conditions involve a\nsmall and spectral parameter. The uniform L_{p}-regularity properties with\nrespect to space variable and parameters are established. Here, the explicit\nformula for the solution is given and behavior of solution is derived when the\nsmall parameter approaches zero. It used to obtain singular perturbation result\nfor abstract elliptic equation\n",
"title": "Singular perturbation for abstract elliptic equations and application"
}
| null | null | null | null | true | null |
17858
| null |
Default
| null | null |
null |
{
"abstract": " Change point analysis is a statistical tool to identify homogeneity within\ntime series data. We propose a pruning approach for approximate nonparametric\nestimation of multiple change points. This general purpose change point\ndetection procedure `cp3o' applies a pruning routine within a dynamic program\nto greatly reduce the search space and computational costs. Existing\ngoodness-of-fit change point objectives can immediately be utilized within the\nframework. We further propose novel change point algorithms by applying cp3o to\ntwo popular nonparametric goodness of fit measures: `e-cp3o' uses E-statistics,\nand `ks-cp3o' uses Kolmogorov-Smirnov statistics. Simulation studies highlight\nthe performance of these algorithms in comparison with parametric and other\nnonparametric change point methods. Finally, we illustrate these approaches\nwith climatological and financial applications.\n",
"title": "Pruning and Nonparametric Multiple Change Point Detection"
}
| null | null | null | null | true | null |
17859
| null |
Default
| null | null |
null |
{
"abstract": " Real-time monitoring of functional tissue parameters, such as local blood\noxygenation, based on optical imaging could provide groundbreaking advances in\nthe diagnosis and interventional therapy of various diseases. While\nphotoacoustic (PA) imaging is a novel modality with great potential to measure\noptical absorption deep inside tissue, quantification of the measurements\nremains a major challenge. In this paper, we introduce the first machine\nlearning based approach to quantitative PA imaging (qPAI), which relies on\nlearning the fluence in a voxel to deduce the corresponding optical absorption.\nThe method encodes relevant information of the measured signal and the\ncharacteristics of the imaging system in voxel-based feature vectors, which\nallow the generation of thousands of training samples from a single simulated\nPA image. Comprehensive in silico experiments suggest that context encoding\n(CE)-qPAI enables highly accurate and robust quantification of the local\nfluence and thereby the optical absorption from PA images.\n",
"title": "Context encoding enables machine learning-based quantitative photoacoustics"
}
| null | null |
[
"Computer Science",
"Physics"
] | null | true | null |
17860
| null |
Validated
| null | null |
null |
{
"abstract": " The interaction that occurs between a light solid object and a horizontal\nsoap film of a bamboo foam contained in a cylindrical tube is simulated in 3D.\nWe vary the shape of the falling object from a sphere to a cube by changing a\nsingle shape parameter as well as varying the initial orientation and position\nof the object. We investigate in detail how the soap film deforms in all these\ncases, and determine the network and pressure forces that a foam exerts on a\nfalling object, due to surface tension and bubble pressure respectively. We\nshow that a cubic particle in a particular orientation experiences the largest\ndrag force, and that this orientation is also the most likely outcome of\ndropping a cube from an arbitrary orientation through a bamboo foam.\n",
"title": "Simulating the interaction between a falling super-quadric object and a soap film"
}
| null | null | null | null | true | null |
17861
| null |
Default
| null | null |
null |
{
"abstract": " Compared with relational database (RDB), graph database (GDB) is a more\nintuitive expression of the real world. Each node in the GDB is a both storage\nand logic unit. Since it is connected to its neighboring nodes through edges,\nand its neighboring information could be easily obtained in one-step graph\ntraversal. It is able to conduct local computation independently and all nodes\ncan do their local work in parallel. Then the whole system can be maximally\nanalyzed and assessed in parallel to largely improve the computation\nperformance without sacrificing the precision of final results. This paper\nfirstly introduces graph database, power system graph modeling and potential\ngraph computing applications in power systems. Two iterative methods based on\ngraph database and PageRank are presented and their convergence are discussed.\nVertex contraction is proposed to improve the performance by eliminating\nzero-impedance branch. A combination of the two iterative methods is proposed\nto make use of their advantages. Testing results based on a provincial 1425-bus\nsystem demonstrate that the proposed comprehensive approach is a good candidate\nfor power flow analysis.\n",
"title": "Power Flow Analysis Using Graph based Combination of Iterative Methods and Vertex Contraction Approach"
}
| null | null | null | null | true | null |
17862
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we present an algorithm for the sparse signal recovery problem\nthat incorporates damped Gaussian generalized approximate message passing\n(GGAMP) into Expectation-Maximization (EM)-based sparse Bayesian learning\n(SBL). In particular, GGAMP is used to implement the E-step in SBL in place of\nmatrix inversion, leveraging the fact that GGAMP is guaranteed to converge with\nappropriate damping. The resulting GGAMP-SBL algorithm is much more robust to\narbitrary measurement matrix $\\boldsymbol{A}$ than the standard damped GAMP\nalgorithm while being much lower complexity than the standard SBL algorithm. We\nthen extend the approach from the single measurement vector (SMV) case to the\ntemporally correlated multiple measurement vector (MMV) case, leading to the\nGGAMP-TSBL algorithm. We verify the robustness and computational advantages of\nthe proposed algorithms through numerical experiments.\n",
"title": "A GAMP Based Low Complexity Sparse Bayesian Learning Algorithm"
}
| null | null | null | null | true | null |
17863
| null |
Default
| null | null |
null |
{
"abstract": " Two-dimensional (2D) materials are among the most promising candidates for\nnext-generation electronics due to their atomic thinness, allowing for flexible\ntransparent electronics and ultimate length scaling. Thus far, atomically-thin\np-n junctions, metal-semiconductor contacts, and metal-insulator barriers have\nbeen demonstrated. While 2D materials achieve the thinnest possible devices,\nprecise nanoscale control over the lateral dimensions is also necessary. Here,\nwe report the direct synthesis of sub-nanometer-wide 1D MoS2 channels embedded\nwithin WSe2 monolayers, using a dislocation-catalyzed approach. The 1D channels\nhave edges free of misfit dislocations and dangling bonds, forming a coherent\ninterface with the embedding 2D matrix. Periodic dislocation arrays produce 2D\nsuperlattices of coherent MoS2 1D channels in WSe2. Using molecular dynamics\nsimulations, we have identified other combinations of 2D materials where 1D\nchannels can also be formed. The electronic band structure of these 1D channels\noffer the promise of carrier confinement in a direct-gap material and charge\nseparation needed to access the ultimate length scales necessary for future\nelectronic applications.\n",
"title": "Sub-Nanometer Channels Embedded in Two-Dimensional Materials"
}
| null | null |
[
"Physics"
] | null | true | null |
17864
| null |
Validated
| null | null |
null |
{
"abstract": " This article provides the first survey of computational models of emotion in\nreinforcement learning (RL) agents. The survey focuses on agent/robot emotions,\nand mostly ignores human user emotions. Emotions are recognized as functional\nin decision-making by influencing motivation and action selection. Therefore,\ncomputational emotion models are usually grounded in the agent's decision\nmaking architecture, of which RL is an important subclass. Studying emotions in\nRL-based agents is useful for three research fields. For machine learning (ML)\nresearchers, emotion models may improve learning efficiency. For the\ninteractive ML and human-robot interaction (HRI) community, emotions can\ncommunicate state and enhance user investment. Lastly, it allows affective\nmodelling (AM) researchers to investigate their emotion theories in a\nsuccessful AI agent class. This survey provides background on emotion theory\nand RL. It systematically addresses 1) from what underlying dimensions (e.g.,\nhomeostasis, appraisal) emotions can be derived and how these can be modelled\nin RL-agents, 2) what types of emotions have been derived from these\ndimensions, and 3) how these emotions may either influence the learning\nefficiency of the agent or be useful as social signals. We also systematically\ncompare evaluation criteria, and draw connections to important RL sub-domains\nlike (intrinsic) motivation and model-based RL. In short, this survey provides\nboth a practical overview for engineers wanting to implement emotions in their\nRL agents, and identifies challenges and directions for future emotion-RL\nresearch.\n",
"title": "Emotion in Reinforcement Learning Agents and Robots: A Survey"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
17865
| null |
Validated
| null | null |
null |
{
"abstract": " Tensor network methods are taking a central role in modern quantum physics\nand beyond. They can provide an efficient approximation to certain classes of\nquantum states, and the associated graphical language makes it easy to describe\nand pictorially reason about quantum circuits, channels, protocols, open\nsystems and more. Our goal is to explain tensor networks and some associated\nmethods as quickly and as painlessly as possible. Beginning with the key\ndefinitions, the graphical tensor network language is presented through\nexamples. We then provide an introduction to matrix product states. We conclude\nthe tutorial with tensor contractions evaluating combinatorial counting\nproblems. The first one counts the number of solutions for Boolean formulae,\nwhereas the second is Penrose's tensor contraction algorithm, returning the\nnumber of $3$-edge-colorings of $3$-regular planar graphs.\n",
"title": "Tensor Networks in a Nutshell"
}
| null | null | null | null | true | null |
17866
| null |
Default
| null | null |
null |
{
"abstract": " We present four logic puzzles and after that their solutions. Joseph Yeo\ndesigned 'Cheryl's Birthday'. Mike Hartley came up with a novel solution for\n'One Hundred Prisoners and a Light Bulb'. Jonathan Welton designed 'A Blind\nGuess' and 'Abby's Birthday'. Hans van Ditmarsch and Barteld Kooi authored the\npuzzlebook 'One Hundred Prisoners and a Light Bulb' that contains other\nknowledge puzzles, and that can also be found on the webpage\nthis http URL dedicated to the book.\n",
"title": "Cheryl's Birthday"
}
| null | null |
[
"Computer Science"
] | null | true | null |
17867
| null |
Validated
| null | null |
null |
{
"abstract": " Exploration bonus derived from the novelty of the states in an environment\nhas become a popular approach to motivate exploration for deep reinforcement\nlearning agents in the past few years. Recent methods such as curiosity-driven\nexploration usually estimate the novelty of new observations by the prediction\nerrors of their system dynamics models. Due to the capacity limitation of the\nmodels and difficulty of performing next-frame prediction, however, these\nmethods typically fail to balance between exploration and exploitation in\nhigh-dimensional observation tasks, resulting in the agents forgetting the\nvisited paths and exploring those states repeatedly. Such inefficient\nexploration behavior causes significant performance drops, especially in large\nenvironments with sparse reward signals. In this paper, we propose to introduce\nthe concept of optical flow estimation from the field of computer vision to\ndeal with the above issue. We propose to employ optical flow estimation errors\nto examine the novelty of new observations, such that agents are able to\nmemorize and understand the visited states in a more comprehensive fashion. We\ncompare our method against the previous approaches in a number of experimental\nexperiments. Our results indicate that the proposed method appears to deliver\nsuperior and long-lasting performance than the previous methods. We further\nprovide a set of comprehensive ablative analysis of the proposed method, and\ninvestigate the impact of optical flow estimation on the learning curves of the\nDRL agents.\n",
"title": "Never Forget: Balancing Exploration and Exploitation via Learning Optical Flow"
}
| null | null | null | null | true | null |
17868
| null |
Default
| null | null |
null |
{
"abstract": " Much of the success of single agent deep reinforcement learning (DRL) in\nrecent years can be attributed to the use of experience replay memories (ERM),\nwhich allow Deep Q-Networks (DQNs) to be trained efficiently through sampling\nstored state transitions. However, care is required when using ERMs for\nmulti-agent deep reinforcement learning (MA-DRL), as stored transitions can\nbecome outdated because agents update their policies in parallel [11]. In this\nwork we apply leniency [23] to MA-DRL. Lenient agents map state-action pairs to\ndecaying temperature values that control the amount of leniency applied towards\nnegative policy updates that are sampled from the ERM. This introduces optimism\nin the value-function update, and has been shown to facilitate cooperation in\ntabular fully-cooperative multi-agent reinforcement learning problems. We\nevaluate our Lenient-DQN (LDQN) empirically against the related Hysteretic-DQN\n(HDQN) algorithm [22] as well as a modified version we call scheduled-HDQN,\nthat uses average reward learning near terminal states. Evaluations take place\nin extended variations of the Coordinated Multi-Agent Object Transportation\nProblem (CMOTP) [8] which include fully-cooperative sub-tasks and stochastic\nrewards. We find that LDQN agents are more likely to converge to the optimal\npolicy in a stochastic reward CMOTP compared to standard and scheduled-HDQN\nagents.\n",
"title": "Lenient Multi-Agent Deep Reinforcement Learning"
}
| null | null | null | null | true | null |
17869
| null |
Default
| null | null |
null |
{
"abstract": " Synthetic aperture imaging systems achieve constant azimuth resolution by\ncoherently summating the observations acquired along the aperture path. At this\naim, their locations have to be known with subwavelength accuracy. In\nunderwater Synthetic Aperture Sonar (SAS), the nature of propagation and\nnavigation in water makes the retrieval of this information challenging.\nInertial sensors have to be employed in combination with signal processing\ntechniques, which are usually referred to as micronavigation. In this paper we\npropose a novel micronavigation approach based on the minimization of an error\nfunction between two contiguous pings having some mutual information. This\nerror is obtained by comparing the vector space intersections between the pings\northogonal projectors. The effectiveness and generality of the proposed\napproach is demonstrated by means of simulations and by means of an experiment\nperformed in a controlled environment.\n",
"title": "A New Framework for Synthetic Aperture Sonar Micronavigation"
}
| null | null | null | null | true | null |
17870
| null |
Default
| null | null |
null |
{
"abstract": " In designing most software applications, much effort is placed upon the\nfunctional goals, which make a software system useful. However, the failure to\nconsider emotional goals, which make a software system pleasurable to use, can\nresult in disappointment and system rejection even if utilitarian goals are\nwell implemented. Although several studies have emphasized the importance of\npeople's emotional goals in developing software, there is little advice on how\nto address these goals in the software system development process. This paper\nproposes a theoretically-sound and practical method by combining the theories\nand techniques of software engineering, requirements engineering, and decision\nmaking. The outcome of this study is the Emotional Goal Systematic Analysis\nTechnique (EG-SAT), which facilitates the process of finding software system\ncapabilities to address emotional goals in software design. EG-SAT is easy to\nlearn and easy to use technique that helps analysts to gain insights into how\nto address people's emotional goals. To demonstrate the method in use, a\ntwo-part evaluation is conducted. First, EG-SAT is used to analyze the\nemotional goals of potential users of a mobile learning application that\nprovides information about low carbon living for tradespeople and professionals\nin the building industry in Australia. The results of using EG-SAT in this case\nstudy are compared with a professionally-developed baseline. Second, we ran a\nsemi-controlled experiment in which 12 participants were asked to apply EG-SAT\nand another technique on part of our case study. The outcomes show that EG-SAT\nhelped participants to both analyse emotional goals and gain valuable insights\nabout the functional and non-functional goals for addressing people's emotional\ngoals.\n",
"title": "Emotionalism within People-Oriented Software Design"
}
| null | null |
[
"Computer Science"
] | null | true | null |
17871
| null |
Validated
| null | null |
null |
{
"abstract": " Principal component analysis is an important pattern recognition and\ndimensionality reduction tool in many applications. Principal components are\ncomputed as eigenvectors of a maximum likelihood covariance $\\widehat{\\Sigma}$\nthat approximates a population covariance $\\Sigma$, and these eigenvectors are\noften used to extract structural information about the variables (or\nattributes) of the studied population. Since PCA is based on the\neigendecomposition of the proxy covariance $\\widehat{\\Sigma}$ rather than the\nground-truth $\\Sigma$, it is important to understand the approximation error in\neach individual eigenvector as a function of the number of available samples.\nThe recent results of Kolchinskii and Lounici yield such bounds. In the present\npaper we sharpen these bounds and show that eigenvectors can often be\nreconstructed to a required accuracy from a sample of strictly smaller size\norder.\n",
"title": "Quantifying the Estimation Error of Principal Components"
}
| null | null | null | null | true | null |
17872
| null |
Default
| null | null |
null |
{
"abstract": " In recent years, several convex programming relaxations have been proposed to\nestimate the permanent of a non-negative matrix, notably in the works of\nGurvits and Samorodnitsky. However, the origins of these relaxations and their\nrelationships to each other have remained somewhat mysterious. We present a\nconceptual framework, implicit in the belief propagation literature, to\nsystematically arrive at these convex programming relaxations for estimating\nthe permanent -- as approximations to an exponential-sized max-entropy convex\nprogram for computing the permanent. Further, using standard convex programming\ntechniques such as duality, we establish equivalence of these aforementioned\nrelaxations to those based on capacity-like quantities studied by Gurvits and\nAnari et al.\n",
"title": "On Convex Programming Relaxations for the Permanent"
}
| null | null | null | null | true | null |
17873
| null |
Default
| null | null |
null |
{
"abstract": " Building on recent work of Bhargava--Elkies--Schnidman and Kriz--Li, we\nproduce infinitely many smooth cubic surfaces defined over the field of\nrational numbers that contain rational points.\n",
"title": "Many cubic surfaces contain rational points"
}
| null | null | null | null | true | null |
17874
| null |
Default
| null | null |
null |
{
"abstract": " Drug repositioning (DR) refers to identification of novel indications for the\napproved drugs. The requirement of huge investment of time as well as money and\nrisk of failure in clinical trials have led to surge in interest in drug\nrepositioning. DR exploits two major aspects associated with drugs and\ndiseases: existence of similarity among drugs and among diseases due to their\nshared involved genes or pathways or common biological effects. Existing\nmethods of identifying drug-disease association majorly rely on the information\navailable in the structured databases only. On the other hand, abundant\ninformation available in form of free texts in biomedical research articles are\nnot being fully exploited. Word-embedding or obtaining vector representation of\nwords from a large corpora of free texts using neural network methods have been\nshown to give significant performance for several natural language processing\ntasks. In this work we propose a novel way of representation learning to obtain\nfeatures of drugs and diseases by combining complementary information available\nin unstructured texts and structured datasets. Next we use matrix completion\napproach on these feature vectors to learn projection matrix between drug and\ndisease vector spaces. The proposed method has shown competitive performance\nwith state-of-the-art methods. Further, the case studies on Alzheimer's and\nHypertension diseases have shown that the predicted associations are matching\nwith the existing knowledge.\n",
"title": "Representation learning of drug and disease terms for drug repositioning"
}
| null | null | null | null | true | null |
17875
| null |
Default
| null | null |
null |
{
"abstract": " In this work we investigate the dynamics of the nonlinear DDE\n(delay-differential equation)\nx''(t)+x(t-T)+x(t)^3=0\nwhere T is the delay. For T=0 this system is conservative and exhibits no\nlimit cycles. For T>0, no matter how small, an infinite number of limit cycles\nexist, their amplitudes going to infinity in the limit as T approaches zero.\nWe investigate this situation in three ways: 1) Harmonic Balance, 2)\nMelnikov's integral, and 3) Adding damping to regularize the singularity.\n",
"title": "Analysis of a remarkable singularity in a nonlinear DDE"
}
| null | null |
[
"Physics",
"Mathematics"
] | null | true | null |
17876
| null |
Validated
| null | null |
null |
{
"abstract": " Our aim is to characterize the Lipschitz functions by variable exponent\nLebesgue spaces. We give some characterizations of the boundedness of the\nmaximal or nonlinear commutators of the Hardy-Littlewood maximal function and\nsharp maximal function in variable exponent Lebesgue spaces when the symbols\n$b$ belong to the Lipschitz spaces, by which some new characterizations of\nLipschitz spaces and nonnegative Lipschitz functions are obtained. Some\nequivalent relations between the Lipschitz norm and the variable exponent\nLebesgue norm are also given.\n",
"title": "Characterization of Lipschitz functions in terms of variable exponent Lebesgue spaces"
}
| null | null | null | null | true | null |
17877
| null |
Default
| null | null |
null |
{
"abstract": " We consider a thin normal metal sandwiched between two ferromagnetic\ninsulators. At the interfaces, the exchange coupling causes electrons within\nthe metal to interact with magnons in the insulators. This electron-magnon\ninteraction induces electron-electron interactions, which, in turn, can result\nin p-wave superconductivity. In the weak-coupling limit, we solve the gap\nequation numerically and estimate the critical temperature. In YIG-Au-YIG\ntrilayers, superconductivity sets in at temperatures somewhere in the interval\nbetween 1 and 10 K. EuO-Au-EuO trilayers require a lower temperature, in the\nrange from 0.01 to 1 K.\n",
"title": "Superconductivity Induced by Interfacial Coupling to Magnons"
}
| null | null | null | null | true | null |
17878
| null |
Default
| null | null |
null |
{
"abstract": " We consider the privacy implications of public release of a de-identified\ndataset of Opal card transactions. The data was recently published at\nthis https URL. It\nconsists of tap-on and tap-off counts for NSW's four modes of public transport,\ncollected over two separate week-long periods. The data has been further\ntreated to improve privacy by removing small counts, aggregating some stops and\nroutes, and perturbing the counts. This is a summary of our findings.\n",
"title": "Privacy Assessment of De-identified Opal Data: A report for Transport for NSW"
}
| null | null | null | null | true | null |
17879
| null |
Default
| null | null |
null |
{
"abstract": " The Nash equilibrium paradigm, and Rational Choice Theory in general, rely on\nagents acting independently from each other. This note shows how this\nassumption is crucial in the definition of Rational Choice Theory. It explains\nhow a consistent Alternate Rational Choice Theory, as suggested by Jean-Pierre\nDupuy, can be built on the exact opposite assumption, and how it provides a\nviable account for alternate, actually observed behavior of rational agents\nthat is based on correlations between their decisions.\nThe end goal of this note is three-fold: (i) to motivate that the Perfect\nPrediction Equilibrium, implementing Dupuy's notion of projected time and\npreviously called \"projected equilibrium\", is a reasonable approach in certain\nreal situations and a meaningful complement to the Nash paradigm, (ii) to\nsummarize common misconceptions about this equilibrium, and (iii) to give a\nconcise motivation for future research on non-Nashian game theory.\n",
"title": "On the Importance of Correlations in Rational Choice: A Case for Non-Nashian Game Theory"
}
| null | null | null | null | true | null |
17880
| null |
Default
| null | null |
null |
{
"abstract": " The predictions of parameteric property models and their uncertainties are\nsensitive to systematic errors such as inconsistent reference data, parametric\nmodel assumptions, or inadequate computational methods. Here, we discuss the\ncalibration of property models in the light of bootstrapping, a sampling method\nakin to Bayesian inference that can be employed for identifying systematic\nerrors and for reliable estimation of the prediction uncertainty. We apply\nbootstrapping to assess a linear property model linking the 57Fe Moessbauer\nisomer shift to the contact electron density at the iron nucleus for a diverse\nset of 44 molecular iron compounds. The contact electron density is calculated\nwith twelve density functionals across Jacob's ladder (PWLDA, BP86, BLYP, PW91,\nPBE, M06-L, TPSS, B3LYP, B3PW91, PBE0, M06, TPSSh). We provide systematic-error\ndiagnostics and reliable, locally resolved uncertainties for isomer-shift\npredictions. Pure and hybrid density functionals yield average prediction\nuncertainties of 0.06-0.08 mm/s and 0.04-0.05 mm/s, respectively, the latter\nbeing close to the average experimental uncertainty of 0.02 mm/s. Furthermore,\nwe show that both model parameters and prediction uncertainty depend\nsignificantly on the composition and number of reference data points.\nAccordingly, we suggest that rankings of density functionals based on\nperformance measures (e.g., the coefficient of correlation, r2, or the\nroot-mean-square error, RMSE) should not be inferred from a single data set.\nThis study presents the first statistically rigorous calibration analysis for\ntheoretical Moessbauer spectroscopy, which is of general applicability for\nphysico-chemical property models and not restricted to isomer-shift\npredictions. We provide the statistically meaningful reference data set MIS39\nand a new calibration of the isomer shift based on the PBE0 functional.\n",
"title": "Reliable estimation of prediction uncertainty for physico-chemical property models"
}
| null | null | null | null | true | null |
17881
| null |
Default
| null | null |
null |
{
"abstract": " Split manufacturing is a promising technique to defend against fab-based\nmalicious activities such as IP piracy, overbuilding, and insertion of hardware\nTrojans. However, a network flow-based proximity attack, proposed by Wang et\nal. (DAC'16) [1], has demonstrated that most prior art on split manufacturing\nis highly vulnerable. Here in this work, we present two practical layout\ntechniques towards secure split manufacturing: (i) gate-level graph coloring\nand (ii) clustering of same-type gates. Our approach shows promising results\nagainst the advanced proximity attack, lowering its success rate by 5.27x,\n3.19x, and 1.73x on average compared to the unprotected layouts when splitting\nat metal layers M1, M2, and M3, respectively. Also, it largely outperforms\nprevious defense efforts; we observe on average 8x higher resilience when\ncompared to representative prior art. At the same time, extensive simulations\non ISCAS'85 and MCNC benchmarks reveal that our techniques incur an acceptable\nlayout overhead. Apart from this empirical study, we provide---for the first\ntime---a theoretical framework for quantifying the layout-level resilience\nagainst any proximity-induced information leakage. Towards this end, we\nleverage the notion of mutual information and provide extensive results to\nvalidate our model.\n",
"title": "Rethinking Split Manufacturing: An Information-Theoretic Approach with Secure Layout Techniques"
}
| null | null | null | null | true | null |
17882
| null |
Default
| null | null |
null |
{
"abstract": " Autonomous sorting is a crucial task in industrial robotics which can be very\nchallenging depending on the expected amount of automation. Usually, to decide\nwhere to sort an object, the system needs to solve either an instance retrieval\n(known object) or a supervised classification (predefined set of classes)\nproblem. In this paper, we introduce a new decision making module, where the\nrobotic system chooses how to sort the objects in an unsupervised way. We call\nthis problem Unsupervised Robotic Sorting (URS) and propose an implementation\non an industrial robotic system, using deep CNN feature extraction and standard\nclustering algorithms. We carry out extensive experiments on various standard\ndatasets to demonstrate the efficiency of the proposed image clustering\npipeline. To evaluate the robustness of our URS implementation, we also\nintroduce a complex real world dataset containing images of objects under\nvarious background and lighting conditions. This dataset is used to fine tune\nthe design choices (CNN and clustering algorithm) for URS. Finally, we propose\na method combining our pipeline with ensemble clustering to use multiple images\nof each object. This redundancy of information about the objects is shown to\nincrease the clustering results.\n",
"title": "Unsupervised robotic sorting: Towards autonomous decision making robots"
}
| null | null | null | null | true | null |
17883
| null |
Default
| null | null |
null |
{
"abstract": " The Whittle likelihood is widely used for Bayesian nonparametric estimation\nof the spectral density of stationary time series. However, the loss of\nefficiency for non-Gaussian time series can be substantial. On the other hand,\nparametric methods are more powerful if the model is well-specified, but may\nfail entirely otherwise. Therefore, we suggest a nonparametric correction of a\nparametric likelihood taking advantage of the efficiency of parametric models\nwhile mitigating sensitivities through a nonparametric amendment. Using a\nBernstein-Dirichlet prior for the nonparametric spectral correction, we show\nposterior consistency and illustrate the performance of our procedure in a\nsimulation study and with LIGO gravitational wave data.\n",
"title": "Beyond Whittle: Nonparametric correction of a parametric likelihood with a focus on Bayesian time series analysis"
}
| null | null | null | null | true | null |
17884
| null |
Default
| null | null |
null |
{
"abstract": " Geophysical model domains typically contain irregular, complex fractal-like\nboundaries and physical processes that act over a wide range of scales.\nConstructing geographically constrained boundary-conforming spatial\ndiscretizations of these domains with flexible use of anisotropically, fully\nunstructured meshes is a challenge. The problem contains a wide range of scales\nand a relatively large, heterogeneous constraint parameter space. Approaches\nare commonly ad hoc, model or application specific and insufficiently\ndescribed. Development of new spatial domains is frequently time-consuming,\nhard to repeat, error prone and difficult to ensure consistent due to the\nsignificant human input required. As a consequence, it is difficult to\nreproduce simulations, ensure a provenance in model data handling and\ninitialization, and a challenge to conduct model intercomparisons rigorously.\nMoreover, for flexible unstructured meshes, there is additionally a greater\npotential for inconsistencies in model initialization and forcing parameters.\nThis paper introduces a consistent approach to unstructured mesh generation for\ngeophysical models, that is automated, quick-to-draft and repeat, and provides\na rigorous and robust approach that is consistent to the source data\nthroughout. The approach is enabling further new research in complex\nmulti-scale domains, difficult or not possible to achieve with existing\nmethods. Examples being actively pursued in a range of geophysical modeling\nefforts are presented alongside the approach, together with the implementation\nlibrary Shingle and a selection of its verification test cases.\n",
"title": "A consistent approach to unstructured mesh generation for geophysical models"
}
| null | null | null | null | true | null |
17885
| null |
Default
| null | null |
null |
{
"abstract": " A map merging component is crucial for the proper functionality of a\nmulti-robot system performing exploration, since it provides the means to\nintegrate and distribute the most important information carried by the agents:\nthe explored-covered space and its exact (depending on the SLAM accuracy)\nmorphology. Map merging is a prerequisite for an intelligent multi-robot team\naiming to deploy a smart exploration technique. In the current work, a metric\nmap merging approach based on environmental information is proposed, in\nconjunction with spatially scattered RFID tags localization. This approach is\ndivided into the following parts: the maps approximate rotation calculation via\nthe obstacles poses and localized RFID tags, the translation employing the best\nlocalized common RFID tag and finally the transformation refinement using an\nICP algorithm.\n",
"title": "Metric Map Merging using RFID Tags & Topological Information"
}
| null | null | null | null | true | null |
17886
| null |
Default
| null | null |
null |
{
"abstract": " This paper introduces a family of local feature aggregation functions and a\nnovel method to estimate their parameters, such that they generate optimal\nrepresentations for classification (or any task that can be expressed as a cost\nfunction minimization problem). To achieve that, we compose the local feature\naggregation function with the classifier cost function and we backpropagate the\ngradient of this cost function in order to update the local feature aggregation\nfunction parameters. Experiments on synthetic datasets indicate that our method\ndiscovers parameters that model the class-relevant information in addition to\nthe local feature space. Further experiments on a variety of motion and visual\ndescriptors, both on image and video datasets, show that our method outperforms\nother state-of-the-art local feature aggregation functions, such as Bag of\nWords, Fisher Vectors and VLAD, by a large margin.\n",
"title": "Learning Local Feature Aggregation Functions with Backpropagation"
}
| null | null | null | null | true | null |
17887
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we demonstrate sub-harmonic injection locking (SHIL) in\nmechanical metronomes. To do so, we first formulate metronome's physical\ncompact model, focusing on its nonlinear terms for friction and the escapement\nmechanism. Then we analyze metronomes using phase-macromodel-based techniques\nand show that the phase of their oscillation is in fact very immune to periodic\nperturbation at twice its natural frequency, making SHIL difficult. Guided by\nthe phase-macromodel-based analysis, we are able to modify the escapement\nmechanism of metronomes such that SHIL can happen more easily. Then we verify\nthe occurrence of SHIL in experiments. To our knowledge, this is the first\ndemonstration of SHIL in metronomes; As such, it provides many valuable\ninsights into the modelling, simulation, analysis and design of nonlinear\noscillators. The demonstration is also suitable to use for teaching the subject\nof injection locking and SHIL.\n",
"title": "Sub-harmonic Injection Locking in Metronomes"
}
| null | null | null | null | true | null |
17888
| null |
Default
| null | null |
null |
{
"abstract": " We have studied the longitudinal spin Seebeck effect in a polar\nantiferromagnet $\\alpha$-Cu$_{2}$V$_{2}$O$_{7}$ in contact with a Pt film.\nBelow the antiferromagnetic transition temperature of\n$\\alpha$-Cu$_{2}$V$_{2}$O$_{7}$, spin Seebeck voltages whose magnetic field\ndependence is similar to that reported in antiferromagnetic MnF$_{2}$$\\mid$Pt\nbilayers are observed. Though a small weak-ferromagnetic moment appears owing\nto the Dzyaloshinskii-Moriya interaction in $\\alpha$-Cu$_{2}$V$_{2}$O$_{7}$,\nthe magnetic field dependence of spin Seebeck voltages is found to be\nirrelevant to the weak ferromagnetic moments. The dependences of the spin\nSeebeck voltages on magnetic fields and temperature are analyzed by a magnon\nspin current theory. The numerical calculation of spin Seebeck voltages using\nmagnetic parameters of $\\alpha$-Cu$_{2}$V$_{2}$O$_{7}$ determined by previous\nneutron scattering studies reveals that the magnetic-field and temperature\ndependences of the spin Seebeck voltages for\n$\\alpha$-Cu$_{2}$V$_{2}$O$_{7}$$\\mid$Pt are governed by the changes in magnon\nlifetimes with magnetic fields and temperature.\n",
"title": "Spin Seebeck effect in a polar antiferromagnet $α$-Cu$_{2}$V$_{2}$O$_{7}$"
}
| null | null | null | null | true | null |
17889
| null |
Default
| null | null |
null |
{
"abstract": " We complete the picture available in the literature by showing that the\nintegral Mackey algebra is Gorenstein if and only if the group order is\nsquare-free, in which case it must have Gorenstein dimension one. We illustrate\nthis result by looking in details at the examples of the cyclic group of order\nfour and the Klein four group.\n",
"title": "Mackey algebras which are Gorenstein"
}
| null | null |
[
"Mathematics"
] | null | true | null |
17890
| null |
Validated
| null | null |
null |
{
"abstract": " Approximate Bayesian computation (ABC) is a method for Bayesian inference\nwhen the likelihood is unavailable but simulating from the model is possible.\nHowever, many ABC algorithms require a large number of simulations, which can\nbe costly. To reduce the computational cost, Bayesian optimisation (BO) and\nsurrogate models such as Gaussian processes have been proposed. Bayesian\noptimisation enables one to intelligently decide where to evaluate the model\nnext but common BO strategies are not designed for the goal of estimating the\nposterior distribution. Our paper addresses this gap in the literature. We\npropose to compute the uncertainty in the ABC posterior density, which is due\nto a lack of simulations to estimate this quantity accurately, and define a\nloss function that measures this uncertainty. We then propose to select the\nnext evaluation location to minimise the expected loss. Experiments show that\nthe proposed method often produces the most accurate approximations as compared\nto common BO strategies.\n",
"title": "Efficient acquisition rules for model-based approximate Bayesian computation"
}
| null | null |
[
"Statistics"
] | null | true | null |
17891
| null |
Validated
| null | null |
null |
{
"abstract": " Regression problems are pervasive in real-world applications. Generally a\nsubstantial amount of labeled samples are needed to build a regression model\nwith good generalization ability. However, many times it is relatively easy to\ncollect a large number of unlabeled samples, but time-consuming or expensive to\nlabel them. Active learning for regression (ALR) is a methodology to reduce the\nnumber of labeled samples, by selecting the most beneficial ones to label,\ninstead of random selection. This paper proposes two new ALR approaches based\non greedy sampling (GS). The first approach (GSy) selects new samples to\nincrease the diversity in the output space, and the second (iGS) selects new\nsamples to increase the diversity in both input and output spaces. Extensive\nexperiments on 12 UCI and CMU StatLib datasets from various domains, and on 15\nsubjects on EEG-based driver drowsiness estimation, verified their\neffectiveness and robustness.\n",
"title": "Active Learning for Regression Using Greedy Sampling"
}
| null | null | null | null | true | null |
17892
| null |
Default
| null | null |
null |
{
"abstract": " We define a new method to estimate centroid for text classification based on\nthe symmetric KL-divergence between the distribution of words in training\ndocuments and their class centroids. Experiments on several standard data sets\nindicate that the new method achieves substantial improvements over the\ntraditional classifiers.\n",
"title": "Centroid estimation based on symmetric KL divergence for Multinomial text classification problem"
}
| null | null | null | null | true | null |
17893
| null |
Default
| null | null |
null |
{
"abstract": " Calculation of near-neighbor interactions among high dimensional, irregularly\ndistributed data points is a fundamental task to many graph-based or\nkernel-based machine learning algorithms and applications. Such calculations,\ninvolving large, sparse interaction matrices, expose the limitation of\nconventional data-and-computation reordering techniques for improving space and\ntime locality on modern computer memory hierarchies. We introduce a novel\nmethod for obtaining a matrix permutation that renders a desirable sparsity\nprofile. The method is distinguished by the guiding principle to obtain a\nprofile that is block-sparse with dense blocks. Our profile model and measure\ncapture the essential properties affecting space and time locality, and permit\nvariation in sparsity profile without imposing a restriction to a fixed\npattern. The second distinction lies in an efficient algorithm for obtaining a\ndesirable profile, via exploring and exploiting multi-scale cluster structure\nhidden in but intrinsic to the data. The algorithm accomplishes its task with\nkey components for lower-dimensional embedding with data-specific principal\nfeature axes, hierarchical data clustering, multi-level matrix compression\nstorage, and multi-level interaction computations. We provide experimental\nresults from case studies with two important data analysis algorithms. The\nresulting performance is remarkably comparable to the BLAS performance for the\nbest-case interaction governed by a regularly banded matrix with the same\nsparsity.\n",
"title": "Rapid Near-Neighbor Interaction of High-dimensional Data via Hierarchical Clustering"
}
| null | null |
[
"Computer Science"
] | null | true | null |
17894
| null |
Validated
| null | null |
null |
{
"abstract": " The evolution and the present status of the gaseous photon detectors\ntechnologies are reviewed. The most recent developments in several branches of\nthe field are described, in particular the installation and commissioning of\nthe first large area MPGD-based detectors of single photons on COMPASS RICH-1.\nInvestigation of novel detector architectures, different materials and various\napplications are reported, and the quest for visible light gaseous photon\ndetectors is discussed. The progress on the use of gaseous photon detector\nrelated techniques in the field of cryogenic applications and gaseous or liquid\nscintillation imaging are presented.\n",
"title": "Evolution and Recent Developments of the Gaseous Photon Detectors Technologies"
}
| null | null | null | null | true | null |
17895
| null |
Default
| null | null |
null |
{
"abstract": " We propose a hybrid quantum system, where an $LC$ resonator inductively\ninteracts with a flux qubit and is capacitively coupled to a Rydberg atom.\nVarying the external magnetic flux bias controls the flux-qubit flipping and\nthe flux qubit-resonator interface. The atomic spectrum is tuned via an\nelectrostatic field, manipulating the qubit-state transition of atom and the\natom-resonator coupling. Different types of entanglement of superconducting,\nphotonic, and atomic qubits can be prepared via simply tuning the flux bias and\nelectrostatic field, leading to the implementation of three-qubit Toffoli logic\ngate.\n",
"title": "Superconducting Qubit-Resonator-Atom Hybrid System"
}
| null | null |
[
"Physics"
] | null | true | null |
17896
| null |
Validated
| null | null |
null |
{
"abstract": " Video games and the playing thereof have been a fixture of American culture\nsince their introduction in the arcades of the 1980s. However, it was not until\nthe recent proliferation of broadband connections robust and fast enough to\nhandle live video streaming that players of video games have transitioned from\na content consumer role to a content producer role. Simultaneously, the rise of\nsocial media has revealed how interpersonal connections drive user engagement\nand interest. In this work, we discuss the recent proliferation of video game\nstreaming, particularly on Twitch.tv, analyze trends and patterns in video game\nviewing, and develop predictive models for determining if a new game will have\nsubstantial impact on the streaming ecosystem.\n",
"title": "Heroes and Zeroes: Predicting the Impact of New Video Games on Twitch.tv"
}
| null | null | null | null | true | null |
17897
| null |
Default
| null | null |
null |
{
"abstract": " Learning the latent representation of data in unsupervised fashion is a very\ninteresting process that provides relevant features for enhancing the\nperformance of a classifier. For speech emotion recognition tasks, generating\neffective features is crucial. Currently, handcrafted features are mostly used\nfor speech emotion recognition, however, features learned automatically using\ndeep learning have shown strong success in many problems, especially in image\nprocessing. In particular, deep generative models such as Variational\nAutoencoders (VAEs) have gained enormous success for generating features for\nnatural images. Inspired by this, we propose VAEs for deriving the latent\nrepresentation of speech signals and use this representation to classify\nemotions. To the best of our knowledge, we are the first to propose VAEs for\nspeech emotion classification. Evaluations on the IEMOCAP dataset demonstrate\nthat features learned by VAEs can produce state-of-the-art results for speech\nemotion classification.\n",
"title": "Variational Autoencoders for Learning Latent Representations of Speech Emotion: A Preliminary Study"
}
| null | null | null | null | true | null |
17898
| null |
Default
| null | null |
null |
{
"abstract": " Sensor selection refers to the problem of intelligently selecting a small\nsubset of a collection of available sensors to reduce the sensing cost while\npreserving signal acquisition performance. The majority of sensor selection\nalgorithms find the subset of sensors that best recovers an arbitrary signal\nfrom a number of linear measurements that is larger than the dimension of the\nsignal. In this paper, we develop a new sensor selection algorithm for sparse\n(or near sparse) signals that finds a subset of sensors that best recovers such\nsignals from a number of measurements that is much smaller than the dimension\nof the signal. Existing sensor selection algorithms cannot be applied in such\nsituations. Our proposed Incoherent Sensor Selection (Insense) algorithm\nminimizes a coherence-based cost function that is adapted from recent results\nin sparse recovery theory. Using six datasets, including two real-world\ndatasets on microbial diagnostics and structural health monitoring, we\ndemonstrate the superior performance of Insense for sparse-signal sensor\nselection.\n",
"title": "Insense: Incoherent Sensor Selection for Sparse Signals"
}
| null | null | null | null | true | null |
17899
| null |
Default
| null | null |
null |
{
"abstract": " We introduce the discrete distribution of a Wiener process range. Rather than\nfinding some basic distributional properties including hazard rate function,\nmoments, Stress-strength parameter and order statistics of this distribution,\nthis work studies some basic properties of the truncated version of this\ndistribution. The effectiveness of this distribution is established using a\ndata set.\n",
"title": "Discrete Distribution for a Wiener Process Range and its Properties"
}
| null | null | null | null | true | null |
17900
| null |
Default
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.