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 the study of extensions of polytopes of combinatorial optimization\nproblems, a notorious open question is that for the size of the smallest\nextended formulation of the Minimum Spanning Tree problem on a complete graph\nwith $n$ nodes. The best known lower bound is $\\Omega(n^2)$, the best known\nupper bound is $O(n^3)$.\nIn this note we show that the venerable fooling set method cannot be used to\nimprove the lower bound: every fooling set for the Spanning Tree polytope has\nsize $O(n^2)$.\n",
"title": "Fooling Sets and the Spanning Tree Polytope"
} | null | null | null | null | true | null | 20401 | null | Default | null | null |
null | {
"abstract": " Given a finite honest time, we derive representations for the additive and\nmultiplicative decomposition of it's Azéma supermartingale in terms of\noptional supermartingales and its running supremum. We then extend the notion\nof semimartingales of class-$(\\Sigma)$ to optional semimartingales with jumps\nin its finite variation part, allowing one to establish formulas similar to the\nMadan-Roynette-Yor option pricing formulas for larger class of processes.\nFinally, we introduce the optional multiplicative systems associated with\npositive submartingales and apply them to construct random times with given\nAzéma supermartingale.\n",
"title": "From Azéma supermartingales of finite honest times to optional semimartingales of class-($Σ$)"
} | null | null | null | null | true | null | 20402 | null | Default | null | null |
null | {
"abstract": " In this paper we study the implications for conference program committees of\nusing single-blind reviewing, in which committee members are aware of the names\nand affiliations of paper authors, versus double-blind reviewing, in which this\ninformation is not visible to committee members. WSDM 2017, the 10th ACM\nInternational ACM Conference on Web Search and Data Mining, performed a\ncontrolled experiment in which each paper was reviewed by four committee\nmembers. Two of these four reviewers were chosen from a pool of committee\nmembers who had access to author information; the other two were chosen from a\ndisjoint pool who did not have access to this information. This information\nasymmetry persisted through the process of bidding for papers, reviewing\npapers, and entering scores. Reviewers in the single-blind condition typically\nbid for 22% fewer papers, and preferentially bid for papers from top\ninstitutions. Once papers were allocated to reviewers, single-blind reviewers\nwere significantly more likely than their double-blind counterparts to\nrecommend for acceptance papers from famous authors and top institutions. The\nestimated odds multipliers are 1.63 for famous authors and 1.58 and 2.10 for\ntop universities and companies respectively, so the result is tangible. For\nfemale authors, the associated odds multiplier of 0.78 is not statistically\nsignificant in our study. However, a meta-analysis places this value in line\nwith that of other experiments, and in the context of this larger aggregate the\ngender effect is also statistically significant.\n",
"title": "Single versus Double Blind Reviewing at WSDM 2017"
} | null | null | null | null | true | null | 20403 | null | Default | null | null |
null | {
"abstract": " Real-world networks are difficult to characterize because of the variation of\ntopological scales, the non-dyadic complex interactions, and the fluctuations.\nHere, we propose a general framework to address these problems via a\nmethodology grounded on topology data analysis. By observing the diffusion\nprocess in a network at a single specified timescale, we can map the network\nnodes to a point cloud, which contains the topological information of the\nnetwork at a single scale. We then calculate the point clouds constructed over\nvariable timescales, which provide scale-variant topological information and\nenable a deep understanding of the network structure and functionality.\nExperiments on synthetic and real-world data demonstrate the effectiveness of\nour framework in identifying network models, classifying real-world networks\nand detecting transition points in time-evolving networks. Our work presents a\nunified analysis that is potentially applicable to more complicated network\nstructures such as multilayer and multiplex networks.\n",
"title": "Scale-variant Topological Information for Characterizing Complex Networks"
} | null | null | null | null | true | null | 20404 | null | Default | null | null |
null | {
"abstract": " Topological superfluid is an exotic state of quantum matter that possesses a\nnodeless superfluid gap in the bulk and Andreev edge modes at the boundary of a\nfinite system. Here, we study a multi-orbital superfluid driven by attractive\ns-wave interaction in a rotating optical lattice. Interestingly, we find that\nthe rotation induces the inter- orbital hybridization and drives the system\ninto topological orbital superfluid in accordance with intrinsically chiral\nd-wave pairing characteristics. Thanks to the conservation of spin, the\ntopological orbital superfluid supports four rather than two chiral Andreev\nedge modes at the boundary of the lattice. Moreover, we find that the intrinsic\nharmonic confining potential forms a circular spatial barrier which accumulates\natoms and supports a mass current under injection of small angular momentum as\nexternal driving force. This feature provides an experimentally detectable\nphenomenon to verify the topological orbital superfluid with chiral d-wave\norder in a rotating optical lattice.\n",
"title": "Topological orbital superfluid with chiral d-wave order in a rotating optical lattice"
} | null | null | null | null | true | null | 20405 | null | Default | null | null |
null | {
"abstract": " The process of liquidity provision in financial markets can result in\nprolonged exposure to illiquid instruments for market makers. In this case,\nwhere a proprietary position is not desired, pro-actively targeting the right\nclient who is likely to be interested can be an effective means to offset this\nposition, rather than relying on commensurate interest arising through natural\ndemand. In this paper, we consider the inference of a client profile for the\npurpose of corporate bond recommendation, based on typical recorded information\navailable to the market maker. Given a historical record of corporate bond\ntransactions and bond meta-data, we use a topic-modelling analogy to develop a\nprobabilistic technique for compiling a curated list of client recommendations\nfor a particular bond that needs to be traded, ranked by probability of\ninterest. We show that a model based on Latent Dirichlet Allocation offers\npromising performance to deliver relevant recommendations for sales traders.\n",
"title": "Optimal client recommendation for market makers in illiquid financial products"
} | null | null | null | null | true | null | 20406 | null | Default | null | null |
null | {
"abstract": " This paper describes the design and implementation of an audio interface for\nthe Patmos processor, which runs on an Altera DE2-115 FPGA board. This board\nhas an audio codec included, the WM8731. The interface described in this work\nallows to receive and send audio from and to the WM8731, and to synthesize,\nstore or manipulate audio signals writing C programs for Patmos. The audio\ninterface described in this paper is intended to be used with the Patmos\nprocessor. Patmos is an open source RISC ISAs with a load-store architecture,\nthat is optimized for Real-Time Systems. Patmos is part of a project founded by\nthe European Union called T-CREST (Time-predictable Multi-Core Architecture for\nEmbedded Systems).[5] The structure of this project is integrated with the\nPatmos project: new hardware modules have been added as IOs, which allow the\ncommunication between the processor and the audio codec. These modules include\na clock generator for the audio chip, ADC and DAC modules for the audio\nconversion from analog to digital and vice versa, and an I2C module which\nallows setting configuration parameters on the audio codec. Moreover, a top\nmodule has been created, which connects all the modules previously mentioned\nbetween them, to Patmos and to the WM8731, using the external pins of the FPGA.\n",
"title": "Design of an Audio Interface for Patmos"
} | null | null | null | null | true | null | 20407 | null | Default | null | null |
null | {
"abstract": " The variational autoencoder (VAE) is a popular probabilistic generative\nmodel. However, one shortcoming of VAEs is that the latent variables cannot be\ndiscrete, which makes it difficult to generate data from different modes of a\ndistribution. Here, we propose an extension of the VAE framework that\nincorporates a classifier to infer the discrete class of the modeled data. To\nmodel sequential data, we can combine our Classifying VAE with a recurrent\nneural network such as an LSTM. We apply this model to algorithmic music\ngeneration, where our model learns to generate musical sequences in different\nkeys. Most previous work in this area avoids modeling key by transposing data\ninto only one or two keys, as opposed to the 10+ different keys in the original\nmusic. We show that our Classifying VAE and Classifying VAE+LSTM models\noutperform the corresponding non-classifying models in generating musical\nsamples that stay in key. This benefit is especially apparent when trained on\nuntransposed music data in the original keys.\n",
"title": "A Classifying Variational Autoencoder with Application to Polyphonic Music Generation"
} | null | null | null | null | true | null | 20408 | null | Default | null | null |
null | {
"abstract": " Consider a sequence of real data points $X_1,\\ldots, X_n$ with underlying\nmeans $\\theta^*_1,\\dots,\\theta^*_n$. This paper starts from studying the\nsetting that $\\theta^*_i$ is both piecewise constant and monotone as a function\nof the index $i$. For this, we establish the exact minimax rate of estimating\nsuch monotone functions, and thus give a non-trivial answer to an open problem\nin the shape-constrained analysis literature. The minimax rate involves an\ninteresting iterated logarithmic dependence on the dimension, a phenomenon that\nis revealed through characterizing the interplay between the isotonic shape\nconstraint and model selection complexity. We then develop a penalized\nleast-squares procedure for estimating the vector\n$\\theta^*=(\\theta^*_1,\\dots,\\theta^*_n)^T$. This estimator is shown to achieve\nthe derived minimax rate adaptively. For the proposed estimator, we further\nallow the model to be misspecified and derive oracle inequalities with the\noptimal rates, and show there exists a computationally efficient algorithm to\ncompute the exact solution.\n",
"title": "On Estimation of Isotonic Piecewise Constant Signals"
} | null | null | null | null | true | null | 20409 | null | Default | null | null |
null | {
"abstract": " The concept of stochastic configuration networks (SCNs) others a solid\nframework for fast implementation of feedforward neural networks through\nrandomized learning. Unlike conventional randomized approaches, SCNs provide an\navenue to select appropriate scope of random parameters to ensure the universal\napproximation property. In this paper, a deep version of stochastic\nconfiguration networks, namely deep stacked stochastic configuration network\n(DSSCN), is proposed for modeling non-stationary data streams. As an extension\nof evolving stochastic connfiguration networks (eSCNs), this work contributes a\nway to grow and shrink the structure of deep stochastic configuration networks\nautonomously from data streams. The performance of DSSCN is evaluated by six\nbenchmark datasets. Simulation results, compared with prominent data stream\nalgorithms, show that the proposed method is capable of achieving comparable\naccuracy and evolving compact and parsimonious deep stacked network\narchitecture.\n",
"title": "Deep Stacked Stochastic Configuration Networks for Non-Stationary Data Streams"
} | null | null | [
"Statistics"
]
| null | true | null | 20410 | null | Validated | null | null |
null | {
"abstract": " Many database columns contain string or numerical data that conforms to a\npattern, such as phone numbers, dates, addresses, product identifiers, and\nemployee ids. These patterns are useful in a number of data processing\napplications, including understanding what a specific field represents when\nfield names are ambiguous, identifying outlier values, and finding similar\nfields across data sets. One way to express such patterns would be to learn\nregular expressions for each field in the database. Unfortunately, exist- ing\ntechniques on regular expression learning are slow, taking hundreds of seconds\nfor columns of just a few thousand values. In contrast, we develop XSystem, an\nefficient method to learn patterns over database columns in significantly less\ntime. We show that these patterns can not only be built quickly, but are\nexpressive enough to capture a number of key applications, including detecting\noutliers, measuring column similarity, and assigning semantic labels to columns\n(based on a library of regular expressions). We evaluate these applications\nwith datasets that range from chemical databases (based on a collaboration with\na pharmaceutical company), our university data warehouse, and open data from\nMassData.gov.\n",
"title": "Extracting Syntactic Patterns from Databases"
} | null | null | null | null | true | null | 20411 | null | Default | null | null |
null | {
"abstract": " We address the problem of multi-class classification in the case where the\nnumber of classes is very large. We propose a double sampling strategy on top\nof a multi-class to binary reduction strategy, which transforms the original\nmulti-class problem into a binary classification problem over pairs of\nexamples. The aim of the sampling strategy is to overcome the curse of\nlong-tailed class distributions exhibited in majority of large-scale\nmulti-class classification problems and to reduce the number of pairs of\nexamples in the expanded data. We show that this strategy does not alter the\nconsistency of the empirical risk minimization principle defined over the\ndouble sample reduction. Experiments are carried out on DMOZ and Wikipedia\ncollections with 10,000 to 100,000 classes where we show the efficiency of the\nproposed approach in terms of training and prediction time, memory consumption,\nand predictive performance with respect to state-of-the-art approaches.\n",
"title": "Aggressive Sampling for Multi-class to Binary Reduction with Applications to Text Classification"
} | null | null | null | null | true | null | 20412 | null | Default | null | null |
null | {
"abstract": " Travel time on a route varies substantially by time of day and from day to\nday. It is critical to understand to what extent this variation is correlated\nwith various factors, such as weather, incidents, events or travel demand level\nin the context of dynamic networks. This helps a better decision making for\ninfrastructure planning and real-time traffic operation. We propose a\ndata-driven approach to understand and predict highway travel time using\nspatio-temporal features of those factors, all of which are acquired from\nmultiple data sources. The prediction model holistically selects the most\nrelated features from a high-dimensional feature space by correlation analysis,\nprinciple component analysis and LASSO. We test and compare the performance of\nseveral regression models in predicting travel time 30 min in advance via two\ncase studies: (1) a 6-mile highway corridor of I-270N in D.C. region, and (2) a\n2.3-mile corridor of I-376E in Pittsburgh region. We found that some\nbottlenecks scattered in the network can imply congestion on those corridors at\nleast 30 minutes in advance, including those on the alternative route to the\ncorridors of study. In addition, real-time travel time is statistically related\nto incidents on some specific locations, morning/afternoon travel demand,\nvisibility, precipitation, wind speed/gust and the weather type. All those\nspatio-temporal information together help improve prediction accuracy,\ncomparing to using only speed data. In both case studies, random forest shows\nthe most promise, reaching a root-mean-squared error of 16.6\\% and 17.0\\%\nrespectively in afternoon peak hours for the entire year of 2014.\n",
"title": "Understanding and predicting travel time with spatio-temporal features of network traffic flow, weather and incidents"
} | null | null | [
"Statistics"
]
| null | true | null | 20413 | null | Validated | null | null |
null | {
"abstract": " We study planted problems---finding hidden structures in random noisy\ninputs---through the lens of the sum-of-squares semidefinite programming\nhierarchy (SoS). This family of powerful semidefinite programs has recently\nyielded many new algorithms for planted problems, often achieving the best\nknown polynomial-time guarantees in terms of accuracy of recovered solutions\nand robustness to noise. One theme in recent work is the design of spectral\nalgorithms which match the guarantees of SoS algorithms for planted problems.\nClassical spectral algorithms are often unable to accomplish this: the twist in\nthese new spectral algorithms is the use of spectral structure of matrices\nwhose entries are low-degree polynomials of the input variables. We prove that\nfor a wide class of planted problems, including refuting random constraint\nsatisfaction problems, tensor and sparse PCA, densest-k-subgraph, community\ndetection in stochastic block models, planted clique, and others, eigenvalues\nof degree-d matrix polynomials are as powerful as SoS semidefinite programs of\nroughly degree d. For such problems it is therefore always possible to match\nthe guarantees of SoS without solving a large semidefinite program. Using\nrelated ideas on SoS algorithms and low-degree matrix polynomials (and inspired\nby recent work on SoS and the planted clique problem by Barak et al.), we prove\nnew nearly-tight SoS lower bounds for the tensor and sparse principal component\nanalysis problems. Our lower bounds for sparse principal component analysis are\nthe first to suggest that going beyond existing algorithms for this problem may\nrequire sub-exponential time.\n",
"title": "The power of sum-of-squares for detecting hidden structures"
} | null | null | null | null | true | null | 20414 | null | Default | null | null |
null | {
"abstract": " Polynomial chaos expansions (PCE) have seen widespread use in the context of\nuncertainty quantification. However, their application to structural\nreliability problems has been hindered by the limited performance of PCE in the\ntails of the model response and due to the lack of local metamodel error\nestimates. We propose a new method to provide local metamodel error estimates\nbased on bootstrap resampling and sparse PCE. An initial experimental design is\niteratively updated based on the current estimation of the limit-state surface\nin an active learning algorithm. The greedy algorithm uses the bootstrap-based\nlocal error estimates for the polynomial chaos predictor to identify the best\ncandidate set of points to enrich the experimental design. We demonstrate the\neffectiveness of this approach on a well-known analytical benchmark\nrepresenting a series system, on a truss structure and on a complex realistic\nframe structure problem.\n",
"title": "An active-learning algorithm that combines sparse polynomial chaos expansions and bootstrap for structural reliability analysis"
} | null | null | null | null | true | null | 20415 | null | Default | null | null |
null | {
"abstract": " We currently harness technologies that could shed new light on old\nphilosophical questions, such as whether our mind entails anything beyond our\nbody or whether our moral values reflect universal truth.\n",
"title": "Experimental Tests of Spirituality"
} | null | null | null | null | true | null | 20416 | null | Default | null | null |
null | {
"abstract": " We prove that the Gromov--Witten theory (GWT) of a projective bundle can be\ndetermined by the Chern classes and the GWT of the base. It completely answers\na question raised in a previous paper (arXiv:1607.00740). Its consequences\ninclude that the GWT of the blow-up of X at a smooth subvariety Z is uniquely\ndetermined by GWT of X, Z plus some topological data.\n",
"title": "Chern classes and Gromov--Witten theory of projective bundles"
} | null | null | null | null | true | null | 20417 | null | Default | null | null |
null | {
"abstract": " We introduce the concept of numerical Gaussian processes, which we define as\nGaussian processes with covariance functions resulting from temporal\ndiscretization of time-dependent partial differential equations. Numerical\nGaussian processes, by construction, are designed to deal with cases where: (1)\nall we observe are noisy data on black-box initial conditions, and (2) we are\ninterested in quantifying the uncertainty associated with such noisy data in\nour solutions to time-dependent partial differential equations. Our method\ncircumvents the need for spatial discretization of the differential operators\nby proper placement of Gaussian process priors. This is an attempt to construct\nstructured and data-efficient learning machines, which are explicitly informed\nby the underlying physics that possibly generated the observed data. The\neffectiveness of the proposed approach is demonstrated through several\nbenchmark problems involving linear and nonlinear time-dependent operators. In\nall examples, we are able to recover accurate approximations of the latent\nsolutions, and consistently propagate uncertainty, even in cases involving very\nlong time integration.\n",
"title": "Numerical Gaussian Processes for Time-dependent and Non-linear Partial Differential Equations"
} | null | null | null | null | true | null | 20418 | null | Default | null | null |
null | {
"abstract": " The quality of experience (QoE) is known to be subjective and\ncontext-dependent. Identifying and calculating the factors that affect QoE is\nindeed a difficult task. Recently, a lot of effort has been devoted to estimate\nthe users QoE in order to improve video delivery. In the literature, most of\nthe QoE-driven optimization schemes that realize trade-offs among different\nquality metrics have been addressed under the assumption of homogenous\npopulations. Nevertheless, people perceptions on a given video quality may not\nbe the same, which makes the QoE optimization harder. This paper aims at taking\na step further in order to address this limitation and meet users profiles. To\ndo so, we propose a closed-loop control framework based on the\nusers(subjective) feedbacks to learn the QoE function and optimize it at the\nsame time. Our simulation results show that our system converges to a steady\nstate, where the resulting QoE function noticeably improves the users\nfeedbacks.\n",
"title": "Learning from Experience: A Dynamic Closed-Loop QoE Optimization for Video Adaptation and Delivery"
} | null | null | null | null | true | null | 20419 | null | Default | null | null |
null | {
"abstract": " Deep learning is still not a very common tool in speaker verification field.\nWe study deep convolutional neural network performance in the text-prompted\nspeaker verification task. The prompted passphrase is segmented into word\nstates - i.e. digits -to test each digit utterance separately. We train a\nsingle high-level feature extractor for all states and use cosine similarity\nmetric for scoring. The key feature of our network is the Max-Feature-Map\nactivation function, which acts as an embedded feature selector. By using\nmultitask learning scheme to train the high-level feature extractor we were\nable to surpass the classic baseline systems in terms of quality and achieved\nimpressive results for such a novice approach, getting 2.85% EER on the RSR2015\nevaluation set. Fusion of the proposed and the baseline systems improves this\nresult.\n",
"title": "Deep CNN based feature extractor for text-prompted speaker recognition"
} | null | null | [
"Statistics"
]
| null | true | null | 20420 | null | Validated | null | null |
null | {
"abstract": " Consider a stochastic process being controlled across a communication\nchannel. The control signal that is transmitted across the control channel can\nbe replaced by a malicious attacker. The controller is allowed to implement any\narbitrary detection algorithm to detect if an attacker is present. This work\ncharacterizes some fundamental limitations of when such an attack can be\ndetected, and quantifies the performance degradation that an attacker that\nseeks to be undetected or stealthy can introduce.\n",
"title": "Data-Injection Attacks in Stochastic Control Systems: Detectability and Performance Tradeoffs"
} | null | null | [
"Computer Science",
"Mathematics"
]
| null | true | null | 20421 | null | Validated | null | null |
null | {
"abstract": " Discrminative trackers, employ a classification approach to separate the\ntarget from its background. To cope with variations of the target shape and\nappearance, the classifier is updated online with different samples of the\ntarget and the background. Sample selection, labeling and updating the\nclassifier is prone to various sources of errors that drift the tracker. We\nintroduce the use of an efficient version space shrinking strategy to reduce\nthe labeling errors and enhance its sampling strategy by measuring the\nuncertainty of the tracker about the samples. The proposed tracker, utilize an\nensemble of classifiers that represents different hypotheses about the target,\ndiversify them using boosting to provide a larger and more consistent coverage\nof the version-space and tune the classifiers' weights in voting. The proposed\nsystem adjusts the model update rate by promoting the co-training of the\nshort-memory ensemble with a long-memory oracle. The proposed tracker\noutperformed state-of-the-art trackers on different sequences bearing various\ntracking challenges.\n",
"title": "Efficient Version-Space Reduction for Visual Tracking"
} | null | null | null | null | true | null | 20422 | null | Default | null | null |
null | {
"abstract": " Most of the JavaScript code deployed in the wild has been minified, a process\nin which identifier names are replaced with short, arbitrary and meaningless\nnames. Minified code occupies less space, but also makes the code extremely\ndifficult to manually inspect and understand. This paper presents Context2Name,\na deep learningbased technique that partially reverses the effect of\nminification by predicting natural identifier names for minified names. The\ncore idea is to predict from the usage context of a variable a name that\ncaptures the meaning of the variable. The approach combines a lightweight,\ntoken-based static analysis with an auto-encoder neural network that summarizes\nusage contexts and a recurrent neural network that predict natural names for a\ngiven usage context. We evaluate Context2Name with a large corpus of real-world\nJavaScript code and show that it successfully predicts 47.5% of all minified\nidentifiers while taking only 2.9 milliseconds on average to predict a name. A\ncomparison with the state-of-the-art tools JSNice and JSNaughty shows that our\napproach performs comparably in terms of accuracy while improving in terms of\nefficiency. Moreover, Context2Name complements the state-of-the-art by\npredicting 5.3% additional identifiers that are missed by both existing tools.\n",
"title": "Context2Name: A Deep Learning-Based Approach to Infer Natural Variable Names from Usage Contexts"
} | null | null | null | null | true | null | 20423 | null | Default | null | null |
null | {
"abstract": " In the theory of the Navier-Stokes equations, the viscous fluid in\nincompressible flow is modelled as a homogeneous and dense assemblage of\nconstituent \"fluid particles\" with viscous stress proportional to rate of\nstrain. The crucial concept of fluid flow is the velocity of the particle that\nis accelerated by the pressure and viscous interaction around it. In this\npaper, by virtue of the alternative constituent \"micro-finite element\", we\nintroduce a set of new intrinsic quantities, called the vortex fields, to\ncharacterise the relative orientation between elements and the feature of\nmicro-eddies in the element, while the description of viscous interaction in\nfluid returns to the initial intuition that the interlayer friction is\nproportional to the slip strength. Such a framework enables us to reconstruct\nthe dynamics theory of viscous fluid, in which the flowing fluid can be\nmodelled as a finite covering of elements and consequently indicated by a\nspace-time differential manifold that admits complex topological evolution.\n",
"title": "Reconstructing fluid dynamics with micro-finite element"
} | null | null | null | null | true | null | 20424 | null | Default | null | null |
null | {
"abstract": " In this paper, we make an important step towards the black-box machine\nteaching by considering the cross-space machine teaching, where the teacher and\nthe learner use different feature representations and the teacher can not fully\nobserve the learner's model. In such scenario, we study how the teacher is\nstill able to teach the learner to achieve faster convergence rate than the\ntraditional passive learning. We propose an active teacher model that can\nactively query the learner (i.e., make the learner take exams) for estimating\nthe learner's status and provably guide the learner to achieve faster\nconvergence. The sample complexities for both teaching and query are provided.\nIn the experiments, we compare the proposed active teacher with the omniscient\nteacher and verify the effectiveness of the active teacher model.\n",
"title": "Towards Black-box Iterative Machine Teaching"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20425 | null | Validated | null | null |
null | {
"abstract": " Why do large neural network generalize so well on complex tasks such as image\nclassification or speech recognition? What exactly is the role regularization\nfor them? These are arguably among the most important open questions in machine\nlearning today. In a recent and thought provoking paper [C. Zhang et al.]\nseveral authors performed a number of numerical experiments that hint at the\nneed for novel theoretical concepts to account for this phenomenon. The paper\nstirred quit a lot of excitement among the machine learning community but at\nthe same time it created some confusion as discussions on OpenReview.net\ntestifies. The aim of this pedagogical paper is to make this debate accessible\nto a wider audience of data scientists without advanced theoretical knowledge\nin statistical learning. The focus here is on explicit mathematical definitions\nand on a discussion of relevant concepts, not on proofs for which we provide\nreferences.\n",
"title": "On Generalization and Regularization in Deep Learning"
} | null | null | null | null | true | null | 20426 | null | Default | null | null |
null | {
"abstract": " We would like to learn latent representations that are low-dimensional and\nhighly interpretable. A model that has these characteristics is the Gaussian\nProcess Latent Variable Model. The benefits and negative of the GP-LVM are\ncomplementary to the Variational Autoencoder, the former provides interpretable\nlow-dimensional latent representations while the latter is able to handle large\namounts of data and can use non-Gaussian likelihoods. Our inspiration for this\npaper is to marry these two approaches and reap the benefits of both. In order\nto do so we will introduce a novel approximate inference scheme inspired by the\nGP-LVM and the VAE. We show experimentally that the approximation allows the\ncapacity of the generative bottle-neck (Z) of the VAE to be arbitrarily large\nwithout losing a highly interpretable representation, allowing reconstruction\nquality to be unlimited by Z at the same time as a low-dimensional space can be\nused to perform ancestral sampling from as well as a means to reason about the\nembedded data.\n",
"title": "Nonparametric Inference for Auto-Encoding Variational Bayes"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20427 | null | Validated | null | null |
null | {
"abstract": " Deep convolutional neural networks (CNNs) can be applied to malware binary\ndetection through images classification. The performance, however, is degraded\ndue to the imbalance of malware families (classes). To mitigate this issue, we\npropose a simple yet effective weighted softmax loss which can be employed as\nthe final layer of deep CNNs. The original softmax loss is weighted, and the\nweight value can be determined according to class size. A scaling parameter is\nalso included in computing the weight. Proper selection of this parameter has\nbeen studied and an empirical option is given. The weighted loss aims at\nalleviating the impact of data imbalance in an end-to-end learning fashion. To\nvalidate the efficacy, we deploy the proposed weighted loss in a pre-trained\ndeep CNN model and fine-tune it to achieve promising results on malware images\nclassification. Extensive experiments also indicate that the new loss function\ncan fit other typical CNNs with an improved classification performance.\n",
"title": "Imbalanced Malware Images Classification: a CNN based Approach"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20428 | null | Validated | null | null |
null | {
"abstract": " Mechanical oscillators are at the heart of many sensor applications. Recently\nseveral groups have developed oscillators that are probed optically, fabricated\nfrom high-stress silicon nitride films. They exhibit outstanding force\nsensitivities of a few aN/Hz$^{1/2}$ and can also be made highly reflective,\nfor efficient detection. The optical read-out usually requires complex\nexperimental setups, including positioning stages and bulky cavities, making\nthem impractical for real applications. In this paper we propose a novel way of\nbuilding fully integrated all-optical force sensors based on low-loss silicon\nnitride mechanical resonators with a photonic crystal reflector. We can\ncircumvent previous limitations in stability and complexity by simulating a\nsuspended focusing photonic crystal, purely made of silicon nitride. Our design\nallows for an all integrated sensor, built out of a single block that\nintegrates a full Fabry-Pérot cavity, without the need for assembly or\nalignment. The presented simulations will allow for a radical simplification of\nsensors based on high-Q silicon nitride membranes. Our results comprise, to the\nbest of our knowledge, the first simulations of a focusing mirror made from a\nmechanically suspended flat membrane with subwavelength thickness. Cavity\nlengths between a few hundred $\\mu$m and mm should be directly realizable.\n",
"title": "Integrated optical force sensors using focusing photonic crystal arrays"
} | null | null | null | null | true | null | 20429 | null | Default | null | null |
null | {
"abstract": " IP networks became the most dominant type of information networks nowadays.\nIt provides a number of services and makes it easy for users to be connected.\nIP networks provide an efficient way with a large number of services compared\nto other ways of voice communication. This leads to the migration to make voice\ncalls via IP networks. Despite the wide range of IP networks services,\navailability, and its capabilities, there still a large number of security\nthreats that affect IP networks and for sure affecting other services based on\nit and voice is one of them. This paper discusses reasons of migration from\nmaking voice calls via IP networks and leaving legacy networks, requirements to\nbe available in IP networks to support voice transport, and concentrating on\nSPIT attack and its detection methods. Experiments took place to compare the\ndifferent approaches used to detect spam over VoIP networks.\n",
"title": "Comparison of Signaling and Media Approaches to Detect VoIP SPIT Attack"
} | null | null | null | null | true | null | 20430 | null | Default | null | null |
null | {
"abstract": " The dynamical systems found in Nature are rarely isolated. Instead they\ninteract and influence each other. The coupling functions that connect them\ncontain detailed information about the functional mechanisms underlying the\ninteractions and prescribe the physical rule specifying how an interaction\noccurs. Here, we aim to present a coherent and comprehensive review\nencompassing the rapid progress made recently in the analysis, understanding\nand applications of coupling functions. The basic concepts and characteristics\nof coupling functions are presented through demonstrative examples of different\ndomains, revealing the mechanisms and emphasizing their multivariate nature.\nThe theory of coupling functions is discussed through gradually increasing\ncomplexity from strong and weak interactions to globally-coupled systems and\nnetworks. A variety of methods that have been developed for the detection and\nreconstruction of coupling functions from measured data is described. These\nmethods are based on different statistical techniques for dynamical inference.\nStemming from physics, such methods are being applied in diverse areas of\nscience and technology, including chemistry, biology, physiology, neuroscience,\nsocial sciences, mechanics and secure communications. This breadth of\napplication illustrates the universality of coupling functions for studying the\ninteraction mechanisms of coupled dynamical systems.\n",
"title": "Coupling functions: Universal insights into dynamical interaction mechanisms"
} | null | null | null | null | true | null | 20431 | null | Default | null | null |
null | {
"abstract": " Uncertainty quantification is a critical missing component in radio\ninterferometric imaging that will only become increasingly important as the\nbig-data era of radio interferometry emerges. Statistical sampling approaches\nto perform Bayesian inference, like Markov Chain Monte Carlo (MCMC) sampling,\ncan in principle recover the full posterior distribution of the image, from\nwhich uncertainties can then be quantified. However, for massive data sizes,\nlike those anticipated from the Square Kilometre Array (SKA), it will be\ndifficult if not impossible to apply any MCMC technique due to its inherent\ncomputational cost. We formulate Bayesian inference problems with\nsparsity-promoting priors (motivated by compressive sensing), for which we\nrecover maximum a posteriori (MAP) point estimators of radio interferometric\nimages by convex optimisation. Exploiting recent developments in the theory of\nprobability concentration, we quantify uncertainties by post-processing the\nrecovered MAP estimate. Three strategies to quantify uncertainties are\ndeveloped: (i) highest posterior density credible regions; (ii) local credible\nintervals (cf. error bars) for individual pixels and superpixels; and (iii)\nhypothesis testing of image structure. These forms of uncertainty\nquantification provide rich information for analysing radio interferometric\nobservations in a statistically robust manner. Our MAP-based methods are\napproximately $10^5$ times faster computationally than state-of-the-art MCMC\nmethods and, in addition, support highly distributed and parallelised\nalgorithmic structures. For the first time, our MAP-based techniques provide a\nmeans of quantifying uncertainties for radio interferometric imaging for\nrealistic data volumes and practical use, and scale to the emerging big-data\nera of radio astronomy.\n",
"title": "Uncertainty quantification for radio interferometric imaging: II. MAP estimation"
} | null | null | null | null | true | null | 20432 | null | Default | null | null |
null | {
"abstract": " The \\emph{Orbit Problem} consists of determining, given a linear\ntransformation $A$ on $\\mathbb{Q}^d$, together with vectors $x$ and $y$,\nwhether the orbit of $x$ under repeated applications of $A$ can ever reach $y$.\nThis problem was famously shown to be decidable by Kannan and Lipton in the\n1980s.\nIn this paper, we are concerned with the problem of synthesising suitable\n\\emph{invariants} $\\mathcal{P} \\subseteq \\mathbb{R}^d$, \\emph{i.e.}, sets that\nare stable under $A$ and contain $x$ and not $y$, thereby providing compact and\nversatile certificates of non-reachability. We show that whether a given\ninstance of the Orbit Problem admits a semialgebraic invariant is decidable,\nand moreover in positive instances we provide an algorithm to synthesise\nsuitable invariants of polynomial size.\nIt is worth noting that the existence of \\emph{semilinear} invariants, on the\nother hand, is (to the best of our knowledge) not known to be decidable.\n",
"title": "Semialgebraic Invariant Synthesis for the Kannan-Lipton Orbit Problem"
} | null | null | null | null | true | null | 20433 | null | Default | null | null |
null | {
"abstract": " The Large-Aperture Experiment to Detect the Dark Age (LEDA) was designed to\ndetect the predicted O(100)mK sky-averaged absorption of the Cosmic Microwave\nBackground by Hydrogen in the neutral pre- and intergalactic medium just after\nthe cosmological Dark Age. The spectral signature would be associated with\nemergence of a diffuse Ly$\\alpha$ background from starlight during 'Cosmic\nDawn'. Recently, Bowman et al. (2018) have reported detection of this predicted\nabsorption feature, with an unexpectedly large amplitude of 530 mK, centered at\n78 MHz. Verification of this result by an independent experiment, such as LEDA,\nis pressing. In this paper, we detail design and characterization of the LEDA\nradiometer systems, and a first-generation pipeline that instantiates a signal\npath model. Sited at the Owens Valley Radio Observatory Long Wavelength Array,\nLEDA systems include the station correlator, five well-separated redundant dual\npolarization radiometers and backend electronics. The radiometers deliver a\n30-85MHz band (16<z<34) and operate as part of the larger interferometric\narray, for purposes ultimately of in situ calibration. Here, we report on the\nLEDA system design, calibration approach, and progress in characterization as\nof January 2016. The LEDA systems are currently being modified to improve\nperformance near 78 MHz in order to verify the purported absorption feature.\n",
"title": "Design and characterization of the Large-Aperture Experiment to Detect the Dark Age (LEDA) radiometer systems"
} | null | null | [
"Physics"
]
| null | true | null | 20434 | null | Validated | null | null |
null | {
"abstract": " Making sense of a dataset in an automatic and unsupervised fashion is a\nchallenging problem in statistics and AI. Classical approaches for density\nestimation are usually not flexible enough to deal with the uncertainty\ninherent to real-world data: they are often restricted to fixed latent\ninteraction models and homogeneous likelihoods; they are sensitive to missing,\ncorrupt and anomalous data; moreover, their expressiveness generally comes at\nthe price of intractable inference. As a result, supervision from statisticians\nis usually needed to find the right model for the data. However, as domain\nexperts do not necessarily have to be experts in statistics, we propose\nAutomatic Bayesian Density Analysis (ABDA) to make density estimation\naccessible at large. ABDA automates the selection of adequate likelihood models\nfrom arbitrarily rich dictionaries while modeling their interactions via a deep\nlatent structure adaptively learned from data as a sum-product network. ABDA\ncasts uncertainty estimation at these local and global levels into a joint\nBayesian inference problem, providing robust and yet tractable inference.\nExtensive empirical evidence shows that ABDA is a suitable tool for automatic\nexploratory analysis of heterogeneous tabular data, allowing for missing value\nestimation, statistical data type and likelihood discovery, anomaly detection\nand dependency structure mining, on top of providing accurate density\nestimation.\n",
"title": "Automatic Bayesian Density Analysis"
} | null | null | null | null | true | null | 20435 | null | Default | null | null |
null | {
"abstract": " With the widespread use of machine learning (ML) techniques, ML as a service\nhas become increasingly popular. In this setting, an ML model resides on a\nserver and users can query the model with their data via an API. However, if\nthe user's input is sensitive, sending it to the server is not an option.\nEqually, the service provider does not want to share the model by sending it to\nthe client for protecting its intellectual property and pay-per-query business\nmodel. In this paper, we propose MLCapsule, a guarded offline deployment of\nmachine learning as a service. MLCapsule executes the machine learning model\nlocally on the user's client and therefore the data never leaves the client.\nMeanwhile, MLCapsule offers the service provider the same level of control and\nsecurity of its model as the commonly used server-side execution. In addition,\nMLCapsule is applicable to offline applications that require local execution.\nBeyond protecting against direct model access, we demonstrate that MLCapsule\nallows for implementing defenses against advanced attacks on machine learning\nmodels such as model stealing/reverse engineering and membership inference.\n",
"title": "MLCapsule: Guarded Offline Deployment of Machine Learning as a Service"
} | null | null | null | null | true | null | 20436 | null | Default | null | null |
null | {
"abstract": " The balance held by Brownian motion between temporal regularity and\nrandomness is embodied in a remarkable way by Levy's forgery of continuous\nfunctions. Here we describe how this property can be extended to forge\narbitrary dependences between two statistical systems, and then establish a new\nBrownian independence test based on fluctuating random paths. We also argue\nthat this result allows revisiting the theory of Brownian covariance from a\nphysical perspective and opens the possibility of engineering nonlinear\ncorrelation measures from more general functional integrals.\n",
"title": "Brownian forgery of statistical dependences"
} | null | null | null | null | true | null | 20437 | null | Default | null | null |
null | {
"abstract": " Working in the context of $\\mu$-abstract elementary classes ($\\mu$-AECs) -\nor, equivalently, accessible categories with all morphisms monomorphisms - we\nexamine the two natural notions of size that occur, namely cardinality of\nunderlying sets and internal size. The latter, purely category-theoretic,\nnotion generalizes e.g. density character in complete metric spaces and\ncardinality of orthogonal bases in Hilbert spaces. We consider the relationship\nbetween these notions under mild set-theoretic hypotheses, including weakenings\nof the singular cardinal hypothesis. We also establish preliminary results on\nthe existence and categoricity spectra of $\\mu$-AECs, including specific\nexamples showing dramatic failures of the eventual categoricity conjecture\n(with categoricity defined using cardinality) in $\\mu$-AECs.\n",
"title": "Internal sizes in $μ$-abstract elementary classes"
} | null | null | null | null | true | null | 20438 | null | Default | null | null |
null | {
"abstract": " The Yangtze River has been subject to heavy flooding throughout history, and\nin recent times severe floods such as those in 1998 have resulted in heavy loss\nof life and livelihoods. Dams along the river help to manage flood waters, and\nare important sources of electricity for the region. Being able to forecast\nhigh-impact events at long lead times therefore has enormous potential benefit.\nRecent improvements in seasonal forecasting mean that dynamical climate models\ncan start to be used directly for operational services. The teleconnection from\nEl Niño to Yangtze River basin rainfall meant that the strong El Niño in\nwinter 2015/2016 provided a valuable opportunity to test the application of a\ndynamical forecast system.\nThis paper therefore presents a case study of a real time seasonal forecast\nfor the Yangtze River basin, building on previous work demonstrating the\nretrospective skill of such a forecast. A simple forecasting methodology is\npresented, in which the forecast probabilities are derived from the historical\nrelationship between hindcast and observations. Its performance for 2016 is\ndiscussed. The heavy rainfall in the May-June-July period was correctly\nforecast well in advance. August saw anomalously low rainfall, and the\nforecasts for the June-July-August period correctly showed closer to average\nlevels. The forecasts contributed to the confidence of decision-makers across\nthe Yangtze River basin. Trials of climate services such as this help to\npromote appropriate use of seasonal forecasts, and highlight areas for future\nimprovements.\n",
"title": "Seasonal forecasts of the summer 2016 Yangtze River basin rainfall"
} | null | null | null | null | true | null | 20439 | null | Default | null | null |
null | {
"abstract": " Classical causal inference assumes a treatment meant for a given unit does\nnot have an effect on other units. When this \"no interference\" assumption is\nviolated, new types of spillover causal effects arise, and causal inference\nbecomes much more difficult. In addition, interference introduces a unique\ncomplication where outcomes may transmit treatment influences to each other,\nwhich is a relationship that has some features of a causal one, but is\nsymmetric. In settings where detailed temporal information on outcomes is not\navailable, addressing this complication using statistical inference methods\nbased on Directed Acyclic Graphs (DAGs) (Ogburn & VanderWeele, 2014) leads to\nconceptual difficulties.\nIn this paper, we develop a new approach to decomposing the spillover effect\ninto direct (also known as the contagion effect) and indirect (also known as\nthe infectiousness effect) components that extends the DAG based treatment\ndecomposition approach to mediation found in (Robins & Richardson, 2010) to\ncausal chain graph models (Lauritzen & Richardson, 2002). We show that when\nthese components of the spillover effect are identified in these models, they\nhave an identifying functional, which we call the symmetric mediation formula,\nthat generalizes the mediation formula in DAGs (Pearl, 2011). We further show\nthat, unlike assumptions in classical mediation analysis, an assumption\npermitting identification in our setting leads to restrictions on the observed\ndata law, making the assumption empirically falsifiable. Finally, we discuss\nstatistical inference for the components of the spillover effect in the special\ncase of two interacting outcomes, and discuss a maximum likelihood estimator,\nand a doubly robust estimator.\n",
"title": "Modeling Interference Via Symmetric Treatment Decomposition"
} | null | null | null | null | true | null | 20440 | null | Default | null | null |
null | {
"abstract": " A reliable, accurate, and affordable positioning service is highly required\nin wireless networks. In this paper, the novel Message Passing Hybrid\nLocalization (MPHL) algorithm is proposed to solve the problem of cooperative\ndistributed localization using distance and direction estimates. This hybrid\napproach combines two sensing modalities to reduce the uncertainty in\nlocalizing the network nodes. A statistical model is formulated for the\nproblem, and approximate minimum mean square error (MMSE) estimates of the node\nlocations are computed. The proposed MPHL is a distributed algorithm based on\nbelief propagation (BP) and Markov chain Monte Carlo (MCMC) sampling. It\nimproves the identifiability of the localization problem and reduces its\nsensitivity to the anchor node geometry, compared to distance-only or\ndirection-only localization techniques. For example, the unknown location of a\nnode can be found if it has only a single neighbor; and a whole network can be\nlocalized using only a single anchor node. Numerical results are presented\nshowing that the average localization error is significantly reduced in almost\nevery simulation scenario, about 50% in most cases, compared to the competing\nalgorithms.\n",
"title": "A Bayesian algorithm for distributed network localization using distance and direction data"
} | null | null | null | null | true | null | 20441 | null | Default | null | null |
null | {
"abstract": " We present several upper bounds for the height of global residues of rational\nforms on an affine variety. As a consequence, we deduce upper bounds for the\nheight of the coefficients in the Bergman-Weil trace formula.\nWe also present upper bounds for the degree and the height of the polynomials\nin the elimination theorem on an affine variety. This is an arithmetic analogue\nof Jelonek's effective elimination theorem, that plays a crucial role in the\nproof of our bounds for the height of global residues.\n",
"title": "Bounds for multivariate residues and for the polynomials in the elimination theorem"
} | null | null | null | null | true | null | 20442 | null | Default | null | null |
null | {
"abstract": " Shale gas plays an important role in reducing pollution and adjusting the\nstructure of world energy. Gas content estimation is particularly significant\nin shale gas resource evaluation. There exist various estimation methods, such\nas first principle methods and empirical models. However, resource evaluation\npresents many challenges, especially the insufficient accuracy of existing\nmodels and the high cost resulting from time-consuming adsorption experiments.\nIn this research, a low-cost and high-accuracy model based on geological\nparameters is constructed through statistical learning methods to estimate\nadsorbed shale gas content\n",
"title": "An adsorbed gas estimation model for shale gas reservoirs via statistical learning"
} | null | null | null | null | true | null | 20443 | null | Default | null | null |
null | {
"abstract": " This paper develops a novel methodology for using symbolic knowledge in deep\nlearning. From first principles, we derive a semantic loss function that\nbridges between neural output vectors and logical constraints. This loss\nfunction captures how close the neural network is to satisfying the constraints\non its output. An experimental evaluation shows that it effectively guides the\nlearner to achieve (near-)state-of-the-art results on semi-supervised\nmulti-class classification. Moreover, it significantly increases the ability of\nthe neural network to predict structured objects, such as rankings and paths.\nThese discrete concepts are tremendously difficult to learn, and benefit from a\ntight integration of deep learning and symbolic reasoning methods.\n",
"title": "A Semantic Loss Function for Deep Learning with Symbolic Knowledge"
} | null | null | null | null | true | null | 20444 | null | Default | null | null |
null | {
"abstract": " By tracking the divergence of two initially close trajectories in phase space\nin an Eulerian approach to forced turbulence, the relation between the maximal\nLyapunov exponent $\\lambda$, and the Reynolds number $Re$ is measured using\ndirect numerical simulations, performed on up to $2048^3$ collocation points.\nThe Lyapunov exponent is found to solely depend on the Reynolds number with\n$\\lambda \\propto Re^{0.53}$ and that after a transient period the divergence of\ntrajectories grows at the same rate at all scales. Finally a linear divergence\nis seen that is dependent on the energy forcing rate. Links are made with other\nchaotic systems.\n",
"title": "Chaotic properties of a turbulent isotropic fluid"
} | null | null | null | null | true | null | 20445 | null | Default | null | null |
null | {
"abstract": " As one of the most important semiconductors, silicon (Si) has been used to\nfabricate electronic devices, waveguides, detectors, and solar cells etc.\nHowever, its indirect bandgap hinders the use of Si for making good emitters1.\nFor integrated photonic circuits, Si-based emitters with sizes in the range of\n100-300 nm are highly desirable. Here, we show that efficient white light\nemission can be realized in spherical and cylindrical Si nanoparticles with\nfeature sizes of ~200 nm. The up-converted luminescence appears at the magnetic\nand electric multipole resonances when the nanoparticles are resonantly excited\nat their magnetic and electric dipole resonances by using femtosecond (fs)\nlaser pulses with ultralow low energy of ~40 pJ. The lifetime of the white\nlight is as short as ~52 ps, almost three orders of magnitude smaller than the\nstate-of-the-art results reported so far for Si (~10 ns). Our finding paves the\nway for realizing efficient Si-based emitters compatible with current\nsemiconductor fabrication technology, which can be integrated to photonic\ncircuits.\n",
"title": "White light emission from silicon nanoparticles"
} | null | null | [
"Physics"
]
| null | true | null | 20446 | null | Validated | null | null |
null | {
"abstract": " Humans are the ultimate ecosystem engineers who have profoundly transformed\nthe world's landscapes in order to enhance their survival. Somewhat\nparadoxically, however, sometimes the unforeseen effect of this ecosystem\nengineering is the very collapse of the population it intended to protect. Here\nwe use a spatial version of a standard population dynamics model of ecosystem\nengineers to study the colonization of unexplored virgin territories by a small\nsettlement of engineers. We find that during the expansion phase the population\ndensity reaches values much higher than those the environment can support in\nthe equilibrium situation. When the colonization front reaches the boundary of\nthe available space, the population density plunges sharply and attains its\nequilibrium value. The collapse takes place without warning and happens just\nafter the population reaches its peak number. We conclude that overpopulation\nand the consequent collapse of an expanding population of ecosystem engineers\nis a natural consequence of the nonlinear feedback between the population and\nenvironment variables.\n",
"title": "The collapse of ecosystem engineer populations"
} | null | null | null | null | true | null | 20447 | null | Default | null | null |
null | {
"abstract": " This paper describes a simple yet novel system for generating a highly\nviscous microjet. The jet is produced inside a wettable thin tube partially\nsubmerged in a liquid. The gas-liquid interface inside the tube, which is\ninitially concave, is kept much deeper than that outside the tube. An impulsive\nforce applied at the bottom of a liquid container leads to significant\nacceleration of the liquid inside the tube followed by flow-focusing due to the\nconcave interface. The jet generation process can be divided into two parts\nthat occur in different time scales, i.e. the Impact time (impact duration $\\le\nO(10^{-4})$ s) and Focusing time (focusing duration $\\gg O(10^{-4})$ s). In\nImpact time, the liquid accelerates suddenly due to the impact. In Focusing\ntime, the microjet emerges due to flow-focusing. In order to explain the sudden\nacceleration inside the tube in Impact time, we develop a physical model based\non a pressure impulse approach. Numerical simulations confirm the proposed\nmodel, indicating that the basic mechanism of the acceleration of the liquid\ndue to the impulsive force is elucidated. Remarkably, the viscous effect is\nnegligible in Impact time. In contrast, in Focusing time, the viscosity plays\nan important role in the microjet generation. We experimentally and numerically\ninvestigate the velocity of microjets with various viscosities. We find that\nhigher viscosities lead to reduction of the jet velocity, which can be\ndescribed by using Reynolds number (the ratio between the inertia force and the\nviscous force). This novel device may be a starting point for next-generation\ntechnologies, such as high-viscosity inkjet printers including bioprinters and\nneedle-free injection devices for minimally invasive medical treatments.\n",
"title": "Highly Viscous Microjet Generator"
} | null | null | null | null | true | null | 20448 | null | Default | null | null |
null | {
"abstract": " In this paper we give a close-to-sharp answer to the basic questions: When is\nthere a continuous way to add a point to a configuration of $n$ ordered points\non a surface $S$ of finite type so that all the points are still distinct? When\nthis is possible, what are all the ways to do it? More precisely, let\nPConf$_n(S)$ be the space of ordered $n$-tuples of distinct points in $S$. Let\n$f_n(S): \\text{PConf}_{n+1}(S) \\to \\text{PConf}_n(S)$ be the map given by\n$f_n(x_0,x_1,\\ldots,x_n):=(x_1,\\ldots,x_n)$. We will classify all continuous\nsections of $f_n$ by proving:\n1. If $S=\\mathbb{R}^2$ and $n>3$, any section of $f_{n}(S)$ is either \"adding\na point at infinity\" or \"adding a point near $x_k$\". (We define these two terms\nin Section 2.1, whether we can define \"adding a point near $x_k$\" or \"adding a\npoint at infinity\" depends in a delicate way on properties of $S$.)\n2. If $S=S^2$ a $2$-sphere and $n>4$, any section of $f_{n}(S)$ is \"adding a\npoint near $x_k$\", if $S=S^2$ and $n=2$, the bundle $f_n(S)$ does not have a\nsection. (We define this term in Section 3.2)\n3. If $S=S_g$ a surface of genus $g>1$ and for $n>1$, the bundle $f_{n}(S)$\ndoes not have a section.\n",
"title": "Section problems for configuration spaces of surfaces"
} | null | null | null | null | true | null | 20449 | null | Default | null | null |
null | {
"abstract": " We study the interplay of spin and charge coherence in a single-level quantum\ndot. A tunnel coupling to a superconducting lead induces superconducting\ncorrelations in the dot. With full spin symmetry retained, only even-singlet\nsuperconducting correlations are generated. An applied magnetic field or\nattached ferromagnetic leads partially or fully reduce the spin symmetry, and\nodd-triplet superconducting correlations are generated as well. For\nsingle-level quantum dots, no other superconducting correlations are possible.\nWe analyze, with the help of a diagrammatic real-time technique, the interplay\nof spin symmetry and superconductivity and its signatures in electronic\ntransport, in particular current and shot noise.\n",
"title": "Odd-triplet superconductivity in single-level quantum dots"
} | null | null | null | null | true | null | 20450 | null | Default | null | null |
null | {
"abstract": " This paper is an introduction to the membrane potential equation for neurons.\nIts properties are described, as well as sample applications. Networks of these\nequations can be used for modeling neuronal systems, which also process images\nand video sequences, respectively. Specifically, (i) a dynamic retina is\nproposed (based on a reaction-diffusion system), which predicts afterimages and\nsimple visual illusions, (ii) a system for texture segregation (texture\nelements are understood as even-symmetric contrast features), and (iii) a\nnetwork for detecting object approaches (inspired by the locust visual system).\n",
"title": "From Neuronal Models to Neuronal Dynamics and Image Processing"
} | null | null | null | null | true | null | 20451 | null | Default | null | null |
null | {
"abstract": " We present a compilation of LEGO Technic parts to provide easy-to-build\nconstructions of basic planar linkages. Some technical issues and their\npossible solutions are discussed. To solve questions on fine details---like\ndeciding whether the motion is an exactly straight line or not---we refer to\nthe dynamic mathematics software tool GeoGebra.\n",
"title": "A compilation of LEGO Technic parts to support learning experiments on linkages"
} | null | null | null | null | true | null | 20452 | null | Default | null | null |
null | {
"abstract": " The conjecture is formulated in an affine structure and linked with\ndimension=1 of the defined CA sets. Then some known results are proved in this\ncontext. The short intended proof relies on a direct yet unclear statement\nabout homogeneous dependence of algebraic equations. This might not be a\ncomplete proof or even one on the right track, but it may provoke more thoughts\nin this respect as expected.\n",
"title": "On the Casas-Alvero conjecture"
} | null | null | [
"Mathematics"
]
| null | true | null | 20453 | null | Validated | null | null |
null | {
"abstract": " Recent experiments show no statistical impact of seal length on the\nperformance of long elastomeric seals in relatively smooth test fixtures.\nMotivated by these results, we analytically and computationally investigate the\ncombined effects of seal length and compressibility on the maximum differential\npressure a seal can support. We present a Saint-Venant type analytic shear lag\nsolution for slightly compressible seals with large aspect ratios, which\ncompares well with nonlinear finite element simulations in regions far from the\nends of the seal. However, at the high- and low-pressure ends, where fracture\nis observed experimentally, the analytic solution is in poor agreement with\ndetailed finite element calculations. Nevertheless, we show that the analytic\nsolution provides far-field stress measures that correlate, over a range of\naspect ratios and bulk moduli, the calculated energy release rates for the\ngrowth of small cracks at the two ends of the seal. Thus a single finite\nelement simulation coupled with the analytic solution can be used to determine\ntendencies for fracture at the two ends of the seal over a wide range of\ngeometry and compressibility. Finally, using a hypothetical critical energy\nrelease rate, predictions for whether a crack on the high-pressure end will\nbegin to grow before or after a crack on the low-pressure end begins to grow\nare made using the analytic solution and compared with finite element\nsimulations for finite deformation, hyperelastic seals.\n",
"title": "Effect of compressibility and aspect ratio on performance of long elastic seals"
} | null | null | null | null | true | null | 20454 | null | Default | null | null |
null | {
"abstract": " Let $G$ be a reductive algebraic group over a $p$-adic field or number field\n$K$, and let $V$ be a $K$-linear faithful representation of $G$. A lattice\n$\\Lambda$ in the vector space $V$ defines a model $\\hat{G}_{\\Lambda}$ of $G$\nover $\\mathscr{O}_K$. One may wonder to what extent $\\Lambda$ is determined by\nthe group scheme $\\hat{G}_{\\Lambda}$. In this paper we prove that up to a\nnatural equivalence relation on the set of lattices there are only finitely\nmany $\\Lambda$ corresponding to one model $\\hat{G}_{\\Lambda}$. Furthermore, we\nrelate this fact to moduli spaces of abelian varieties as follows: let\n$\\mathscr{A}_{g,n}$ be the moduli space of principally polarised abelian\nvarieties of dimension $g$ with level $n$ structure. We prove that there are at\nmost finitely many special subvarieties of $\\mathscr{A}_{g,n}$ with a given\nintegral generic Mumford-Tate group.\n",
"title": "Integral models of reductive groups and integral Mumford-Tate groups"
} | null | null | null | null | true | null | 20455 | null | Default | null | null |
null | {
"abstract": " We give a general method of extending unital completely positive maps to\namalgamated free products of C*-algebras. As an application we give a dilation\ntheoretic proof of Boca's Theorem.\n",
"title": "A proof of Boca's Theorem"
} | null | null | [
"Mathematics"
]
| null | true | null | 20456 | null | Validated | null | null |
null | {
"abstract": " Convolution with Green's function of a differential operator appears in a lot\nof applications e.g. Lippmann-Schwinger integral equation. Algorithms for\ncomputing such are usually non-trivial and require non-uniform mesh. However,\nrecently Vico, Greengard and Ferrando developed method for computing\nconvolution with smooth functions with compact support with spectral accuracy,\nrequiring nothing more than Fast Fourier Transform (FFT). Their approach is\nvery suitable for the low-rank tensor implementation which we develop using\nQuantized Tensor Train (QTT) decomposition.\n",
"title": "Vico-Greengard-Ferrando quadratures in the tensor solver for integral equations"
} | null | null | null | null | true | null | 20457 | null | Default | null | null |
null | {
"abstract": " Many-body phenomena were always an integral part of physics comprising of\ncollective behaviors through self-organization, in systems consisting of many\ncomponents and degrees of freedom. We investigate the collective behaviors of\nstrongly interacting particles confined in one dimension. We show that\nmany-body orders with topological characteristics can be found at the Mott\ninsulator limit for hardcore bosons, at different fillings, without considering\nthe spin degree of freedom or long-range microscopic interactions. These orders\nhave unique properties like weak or strong quantum correlations (entanglement),\nquantified by the entanglement entropy, edge excitations/modes and gapped\nenergy spectrum with highly degenerate ground state, bearing resemblance to\ntopologically ordered phases of matter.\n",
"title": "Mott insulators of hardcore bosons in 1D: many-body orders, entanglement, edge modes"
} | null | null | [
"Physics"
]
| null | true | null | 20458 | null | Validated | null | null |
null | {
"abstract": " When performing Bayesian data analysis using a general linear mixed model,\nthe resulting posterior density is almost always analytically intractable.\nHowever, if proper conditionally conjugate priors are used, there is a simple\ntwo-block Gibbs sampler that is geometrically ergodic in nearly all practical\nsettings, including situations where $p > n$ (Abrahamsen and Hobert, 2017).\nUnfortunately, the (conditionally conjugate) multivariate normal prior on\n$\\beta$ does not perform well in the high-dimensional setting where $p \\gg n$.\nIn this paper, we consider an alternative model in which the multivariate\nnormal prior is replaced by the normal-gamma shrinkage prior developed by\nGriffin and Brown (2010). This change leads to a much more complex posterior\ndensity, and we develop a simple MCMC algorithm for exploring it. This\nalgorithm, which has both deterministic and random scan components, is easier\nto analyze than the more obvious three-step Gibbs sampler. Indeed, we prove\nthat the new algorithm is geometrically ergodic in most practical settings.\n",
"title": "Fast Monte Carlo Markov chains for Bayesian shrinkage models with random effects"
} | null | null | [
"Mathematics",
"Statistics"
]
| null | true | null | 20459 | null | Validated | null | null |
null | {
"abstract": " With the increasing of electric vehicle (EV) adoption in recent years, the\nimpact of EV charging activities to the power grid becomes more and more\nsignificant. In this article, an optimal scheduling algorithm which combines\nsmart EV charging and V2G gird service is developed to integrate EVs into power\ngrid as distributed energy resources, with improved system cost performance.\nSpecifically, an optimization problem is formulated and solved at each EV\ncharging station according to control signal from aggregated control center and\nuser charging behavior prediction by mean estimation and linear regression. The\ncontrol center collects distributed optimization results and updates the\ncontrol signal, periodically. The iteration continues until it converges to\noptimal scheduling. Experimental result shows this algorithm helps fill the\nvalley and shave the peak in electric load profiles within a microgrid, while\nthe energy demand of individual driver can be satisfied.\n",
"title": "Distributed Optimal Vehicle Grid Integration Strategy with User Behavior Prediction"
} | null | null | null | null | true | null | 20460 | null | Default | null | null |
null | {
"abstract": " Vulnerabilities in password managers are unremitting because current designs\nprovide large attack surfaces, both at the client and server. We describe and\nevaluate Horcrux, a password manager that is designed holistically to minimize\nand decentralize trust, while retaining the usability of a traditional password\nmanager. The prototype Horcrux client, implemented as a Firefox add-on, is\nsplit into two components, with code that has access to the user's master's\npassword and any key material isolated into a small auditable component,\nseparate from the complexity of managing the user interface. Instead of\nexposing actual credentials to the DOM, a dummy username and password are\nautofilled by the untrusted component. The trusted component intercepts and\nmodifies POST requests before they are encrypted and sent over the network. To\navoid trusting a centralized store, stored credentials are secret-shared over\nmultiple servers. To provide domain and username privacy, while maintaining\nresilience to off-line attacks on a compromised password store, we incorporate\ncuckoo hashing in a way that ensures an attacker cannot determine if a guessed\nmaster password is correct. Our approach only works for websites that do not\nmanipulate entered credentials in the browser client, so we conducted a\nlarge-scale experiment that found the technique appears to be compatible with\nover 98% of tested login forms.\n",
"title": "Horcrux: A Password Manager for Paranoids"
} | null | null | [
"Computer Science"
]
| null | true | null | 20461 | null | Validated | null | null |
null | {
"abstract": " This text is a survey on cross-validation. We define all classical\ncross-validation procedures, and we study their properties for two different\ngoals: estimating the risk of a given estimator, and selecting the best\nestimator among a given family. For the risk estimation problem, we compute the\nbias (which can also be corrected) and the variance of cross-validation\nmethods. For estimator selection, we first provide a first-order analysis\n(based on expectations). Then, we explain how to take into account second-order\nterms (from variance computations, and by taking into account the usefulness of\noverpenalization). This allows, in the end, to provide some guidelines for\nchoosing the best cross-validation method for a given learning problem.\n",
"title": "Cross-validation"
} | null | null | [
"Mathematics",
"Statistics"
]
| null | true | null | 20462 | null | Validated | null | null |
null | {
"abstract": " Objective: Using traditional approaches, a Brain-Computer Interface (BCI)\nrequires the collection of calibration data for new subjects prior to online\nuse. Calibration time can be reduced or eliminated e.g.~by transfer of a\npre-trained classifier or unsupervised adaptive classification methods which\nlearn from scratch and adapt over time. While such heuristics work well in\npractice, none of them can provide theoretical guarantees. Our objective is to\nmodify an event-related potential (ERP) paradigm to work in unison with the\nmachine learning decoder to achieve a reliable calibration-less decoding with a\nguarantee to recover the true class means.\nMethod: We introduce learning from label proportions (LLP) to the BCI\ncommunity as a new unsupervised, and easy-to-implement classification approach\nfor ERP-based BCIs. The LLP estimates the mean target and non-target responses\nbased on known proportions of these two classes in different groups of the\ndata. We modified a visual ERP speller to meet the requirements of the LLP. For\nevaluation, we ran simulations on artificially created data sets and conducted\nan online BCI study with N=13 subjects performing a copy-spelling task.\nResults: Theoretical considerations show that LLP is guaranteed to minimize\nthe loss function similarly to a corresponding supervised classifier. It\nperformed well in simulations and in the online application, where 84.5% of\ncharacters were spelled correctly on average without prior calibration.\nSignificance: The continuously adapting LLP classifier is the first\nunsupervised decoder for ERP BCIs guaranteed to find the true class means. This\nmakes it an ideal solution to avoid a tedious calibration and to tackle\nnon-stationarities in the data. Additionally, LLP works on complementary\nprinciples compared to existing unsupervised methods, allowing for their\nfurther enhancement when combined with LLP.\n",
"title": "Learning from Label Proportions in Brain-Computer Interfaces: Online Unsupervised Learning with Guarantees"
} | null | null | null | null | true | null | 20463 | null | Default | null | null |
null | {
"abstract": " The index coding problem has been generalized recently to accommodate\nreceivers which demand functions of messages and which possess functions of\nmessages. The connections between index coding and matroid theory have been\nwell studied in the recent past. Index coding solutions were first connected to\nmulti linear representation of matroids. For vector linear index codes discrete\npolymatroids which can be viewed as a generalization of the matroids was used.\nIt was shown that a vector linear solution to an index coding problem exists if\nand only if there exists a representable discrete polymatroid satisfying\ncertain conditions. In this work we explore the connections between generalized\nindex coding and discrete polymatroids. The conditions that needs to be\nsatisfied by a representable discrete polymatroid for a generalized index\ncoding problem to have a vector linear solution is established. From a discrete\npolymatroid we construct an index coding problem with coded side information\nand shows that if the index coding problem has a certain optimal length\nsolution then the discrete polymatroid satisfies certain properties. From a\nmatroid we construct a similar generalized index coding problem and shows that\nthe index coding problem has a binary scalar linear solution of optimal length\nif and only if the matroid is binary representable.\n",
"title": "Generalized Index Coding Problem and Discrete Polymatroids"
} | null | null | [
"Computer Science",
"Mathematics"
]
| null | true | null | 20464 | null | Validated | null | null |
null | {
"abstract": " We are concerned with the inverse scattering problem of extracting the\ngeometric structures of an unknown/inaccessible inhomogeneous medium by using\nthe corresponding acoustic far-field measurement. Using the intrinsic geometric\nproperties of the so-called interior transmission eigenfunctions, we develop a\nnovel inverse scattering scheme. The proposed method can efficiently capture\nthe cusp singularities of the support of the inhomogeneous medium. If further a\npriori information is available on the support of the medium, say, it is a\nconvex polyhedron, then one can actually recover its shape. Both theoretical\nanalysis and numerical experiments are provided. Our reconstruction method is\nnew to the literature and opens up a new direction in the study of inverse\nscattering problems.\n",
"title": "Reconstruction via the intrinsic geometric structures of interior transmission eigenfunctions"
} | null | null | [
"Mathematics"
]
| null | true | null | 20465 | null | Validated | null | null |
null | {
"abstract": " Thermoelectric energy conversion - the exploitation of the Seebeck effect to\nconvert waste heat into electricity - has attracted an increasing amount of\nresearch attention for energy harvesting technology. Niobium-doped strontium\ntitanate (SrTi1-xNbxO3) is one of the most promising thermoelectric material\ncandidates, particularly as it poses a much lesser environmental risk in\ncomparison to materials based on heavy metal elements. Two-dimensional electron\nconfinement, e.g. through the formation of superlattices or two-dimensional\nelectron gases, is recognized as an effective strategy to improve the\nthermoelectric performance of SrTi1-xNbxO3. Although electron confinement is\nclosely related to the electronic structure, the fundamental electronic phase\nbehavior of the SrTi1-xNbxO3 solid solution system has yet to be\ncomprehensively investigated. Here, we present a thermoelectric phase diagram\nfor the SrTi1-xNbxO3 (0.05 =< x =< 1) solid solution system, which we derived\nfrom the characterization of epitaxial films. We observed two thermoelectric\nphase boundaries in the system, which originate from the step-like decrease in\ncarrier effective mass at x ~ 0.3, and from a local minimum in carrier\nrelaxation time at x ~ 0.5. The origins of these phase boundaries are\nconsidered to be related to isovalent/heterovalent B-site substitution:\nparabolic Ti 3d orbitals dominate electron conduction for compositions with x <\n0.3, whereas the Nb 4d orbital dominates when x > 0.3. At x ~ 0.5, a tetragonal\ndistortion of the lattice, in which the B-site is composed of Ti4+ and Nb4+\nions, leads to the formation of tail-like impurity bands, which maximizes the\nelectron scattering. These results provide a foundation for further research\ninto improving the thermoelectric performance of SrTi1-xNbxO3.\n",
"title": "Thermoelectric phase diagram of the SrTiO3-SrNbO3 solid solution system"
} | null | null | null | null | true | null | 20466 | null | Default | null | null |
null | {
"abstract": " In this work we study the excitatory AMPA, and NMDA, and inhibitory GABAA\nreceptor mediated dynamical changes in neuronal networks of neonatal rat cortex\nin vitro. Extracellular network-wide activity was recorded with 59 planar\nelectrodes simultaneously under different pharmacological conditions. We\nanalyzed the changes of overall network activity and network-wide burst\nfrequency between baseline and AMPA receptor (AMPA-R) or NMDA receptor (NMDA-R)\ndriven activity, as well as between the latter states and disinhibited\nactivity. Additionally, spatiotemporal structures of pharmacologically modified\nbursts and recruitment of electrodes during the network bursts were studied.\nOur results show that AMPA-R and NMDA-R receptors have clearly distinct roles\nin network dynamics. AMPA-Rs are in greater charge to initiate network wide\nbursts. Therefore NMDA-Rs maintain the already initiated activity. GABAA\nreceptors (GABAA-Rs) inhibit AMPA-R driven network activity more strongly than\nNMDA-R driven activity during the bursts.\n",
"title": "AMPA, NMDA and GABAA receptor mediated network burst dynamics in cortical cultures in vitro"
} | null | null | null | null | true | null | 20467 | null | Default | null | null |
null | {
"abstract": " We use a coarse version of the fundamental group first introduced by Barcelo,\nKramer, Laubenbacher and Weaver to show that box spaces of finitely presented\ngroups detect the normal subgroups used to construct the box space, up to\nisomorphism. As a consequence we have that two finitely presented groups admit\ncoarsely equivalent box spaces if and only if they are commensurable via normal\nsubgroups. We also provide an example of two filtrations $(N_i)$ and $(M_i)$ of\na free group $F$ such that $M_i>N_i$ for all $i$ with $[M_i:N_i]$ uniformly\nbounded, but with $\\Box_{(N_i)}F$ not coarsely equivalent to $\\Box_{(M_i)}F$.\nFinally, we give some applications of the main theorem for rank gradient and\nthe first $\\ell^2$ Betti number, and show that the main theorem can be used to\nconstruct infinitely many coarse equivalence classes of box spaces with various\nproperties.\n",
"title": "Coarse fundamental groups and box spaces"
} | null | null | [
"Mathematics"
]
| null | true | null | 20468 | null | Validated | null | null |
null | {
"abstract": " We study the growth of degrees in many autonomous and non-autonomous lattice\nequations defined by quad rules with corner boundary values, some of which are\nknown to be integrable by other characterisations. Subject to an enabling\nconjecture, we prove polynomial growth for a large class of equations which\nincludes the Adler-Bobenko-Suris equations and Viallet's $Q_V$ and its\nnon-autonomous generalization. Our technique is to determine the ambient degree\ngrowth of the projective version of the lattice equations and to conjecture the\ngrowth of their common factors at each lattice vertex, allowing the true degree\ngrowth to be found. The resulting degrees satisfy a linear partial difference\nequation which is universal, i.e. the same for all the integrable lattice\nequations considered. When we take periodic reductions of these equations,\nwhich includes staircase initial conditions, we obtain from this linear partial\ndifference equation an ordinary difference equation for degrees that implies\nquadratic or linear degree growth. We also study growth of degree of several\nnon-integrable lattice equations. Exponential growth of degrees of these\nequations, and their mapping reductions, is also proved subject to a\nconjecture.\n",
"title": "Algebraic entropy of (integrable) lattice equations and their reductions"
} | null | null | null | null | true | null | 20469 | null | Default | null | null |
null | {
"abstract": " A laser heterodyne polarimeter (LHP) designed for the measurement of the\nbirefringence of dielectric super-mirrors is described and initial results are\nreported. The LHP does not require an optical resonator and so promises\nunprecedented accuracy in the measurement of the birefringence of individual\nmirrors. The working principle of the LHP can be applied to the measurement of\nvacuum birefringence and potentially ALPS (Any Light Particle Search).\n",
"title": "Measurement of mirror birefringence with laser heterodyne polarimetry"
} | null | null | [
"Physics"
]
| null | true | null | 20470 | null | Validated | null | null |
null | {
"abstract": " The availability of big data recorded from massively multiplayer online\nrole-playing games (MMORPGs) allows us to gain a deeper understanding of the\npotential connection between individuals' network positions and their economic\noutputs. We use a statistical filtering method to construct dependence networks\nfrom weighted friendship networks of individuals. We investigate the 30\ndistinct motif positions in the 13 directed triadic motifs which represent\nmicroscopic dependences among individuals. Based on the structural similarity\nof motif positions, we further classify individuals into different groups. The\nnode position diversity of individuals is found to be positively correlated\nwith their economic outputs. We also find that the economic outputs of leaf\nnodes are significantly lower than that of the other nodes in the same motif.\nOur findings shed light on understanding the influence of network structure on\neconomic activities and outputs in socioeconomic system.\n",
"title": "Individual position diversity in dependence socioeconomic networks increases economic output"
} | null | null | null | null | true | null | 20471 | null | Default | null | null |
null | {
"abstract": " The aim of this paper is to give an explicit formula for the nonsymmetric\nHeckman-Opdam's hypergeometric function of type $A_2$. This is obtained by\ndifferentiating the corresponding symmetric hypergeometric function.\n",
"title": "A formula for the nonsymmetric Opdam's hypergeometric function of type $A_2$"
} | null | null | null | null | true | null | 20472 | null | Default | null | null |
null | {
"abstract": " An algorithm for irreducible decomposition of representations of finite\ngroups over fields of characteristic zero is described. The algorithm uses the\nfact that the decomposition induces a partition of the invariant inner product\ninto a complete set of mutually orthogonal projectors. By expressing the\nprojectors through the basis elements of the centralizer ring of the\nrepresentation, the problem is reduced to solving systems of quadratic\nequations. The current implementation of the algorithm is able to split\nrepresentations of dimensions up to hundreds of thousands. Examples of\ncalculations are given.\n",
"title": "A new algorithm for irreducible decomposition of representations of finite groups"
} | null | null | null | null | true | null | 20473 | null | Default | null | null |
null | {
"abstract": " This note is a commentary on the model-theoretic interpretation of\nGrothendieck's double limit characterization of weak relative compactness.\n",
"title": "Stability and Grothendieck"
} | null | null | null | null | true | null | 20474 | null | Default | null | null |
null | {
"abstract": " In this work, we ask the following question: Can visual analogies, learned in\nan unsupervised way, be used in order to transfer knowledge between pairs of\ngames and even play one game using an agent trained for another game? We\nattempt to answer this research question by creating visual analogies between a\npair of games: a source game and a target game. For example, given a video\nframe in the target game, we map it to an analogous state in the source game\nand then attempt to play using a trained policy learned for the source game. We\ndemonstrate convincing visual mapping between four pairs of games (eight\nmappings), which are used to evaluate three transfer learning approaches.\n",
"title": "Visual Analogies between Atari Games for Studying Transfer Learning in RL"
} | null | null | null | null | true | null | 20475 | null | Default | null | null |
null | {
"abstract": " In the classical secretary problem, one attempts to find the maximum of an\nunknown and unlearnable distribution through sequential search. In many\nreal-world searches, however, distributions are not entirely unknown and can be\nlearned through experience. To investigate learning in such a repeated\nsecretary problem we conduct a large-scale behavioral experiment in which\npeople search repeatedly from fixed distributions. In contrast to prior\ninvestigations that find no evidence for learning in the classical scenario, in\nthe repeated setting we observe substantial learning resulting in near-optimal\nstopping behavior. We conduct a Bayesian comparison of multiple behavioral\nmodels which shows that participants' behavior is best described by a class of\nthreshold-based models that contains the theoretically optimal strategy.\nFitting such a threshold-based model to data reveals players' estimated\nthresholds to be surprisingly close to the optimal thresholds after only a\nsmall number of games.\n",
"title": "Learning in the Repeated Secretary Problem"
} | null | null | null | null | true | null | 20476 | null | Default | null | null |
null | {
"abstract": " This paper addresses challenges in flexibly modeling multimodal data that lie\non constrained spaces. Applications include climate or crime measurements in a\ngeographical area, or flow-cytometry experiments, where unsuitable recordings\nare discarded. A simple approach to modeling such data is through the use of\nmixture models, with each component following an appropriate truncated\ndistribution. Problems arise when the truncation involves complicated\nconstraints, leading to difficulties in specifying the component distributions,\nand in evaluating their normalization constants. Bayesian inference over the\nparameters of these models results in posterior distributions that are\ndoubly-intractable. We address this problem via an algorithm based on rejection\nsampling and data augmentation. We view samples from a truncated distribution\nas outcomes of a rejection sampling scheme, where proposals are made from a\nsimple mixture model, and are rejected if they violate the constraints. Our\nscheme proceeds by imputing the rejected samples given mixture parameters, and\nthen resampling parameters given all samples. We study two modeling approaches:\nmixtures of truncated components and truncated mixtures of components. In both\nsituations, we describe exact Markov chain Monte Carlo sampling algorithms, as\nwell as approximations that bound the number of rejected samples, achieving\ncomputational efficiency and lower variance at the cost of asymptotic bias.\nOverall, our methodology only requires practitioners to provide an indicator\nfunction for the set of interest. We present results on simulated data and\napply our algorithm to two problems, one involving flow-cytometry data, and the\nother, crime recorded in the city of Chicago.\n",
"title": "Flexible Mixture Modeling on Constrained Spaces"
} | null | null | null | null | true | null | 20477 | null | Default | null | null |
null | {
"abstract": " By analyzing spin-spin correlation functions at relatively short distances,\nwe show that equilibrium near-critical properties can be extracted at short\ntimes after quenches into the vicinity of a quantum critical point. The time\nscales after which equilibrium properties can be extracted are sufficiently\nshort so that the proposed scheme should be viable for quantum simulators of\nspin models based on ultracold atoms or trapped ions. Our results, analytic as\nwell as numeric, are for one-dimensional spin models, either integrable or\nnonintegrable, but we expect our conclusions to be valid in higher dimensions\nas well.\n",
"title": "Universal equilibrium scaling functions at short times after a quench"
} | null | null | null | null | true | null | 20478 | null | Default | null | null |
null | {
"abstract": " The CUORE experiment, a ton-scale cryogenic bolometer array, recently began\noperation at the Laboratori Nazionali del Gran Sasso in Italy. The array\nrepresents a significant advancement in this technology, and in this work we\napply it for the first time to a high-sensitivity search for a\nlepton-number--violating process: $^{130}$Te neutrinoless double-beta decay.\nExamining a total TeO$_2$ exposure of 86.3 kg$\\cdot$yr, characterized by an\neffective energy resolution of (7.7 $\\pm$ 0.5) keV FWHM and a background in the\nregion of interest of (0.014 $\\pm$ 0.002) counts/(keV$\\cdot$kg$\\cdot$yr), we\nfind no evidence for neutrinoless double-beta decay. The median statistical\nsensitivity of this search is $7.0\\times10^{24}$ yr. Including systematic\nuncertainties, we place a lower limit on the decay half-life of\n$T^{0\\nu}_{1/2}$($^{130}$Te) > $1.3\\times 10^{25}$ yr (90% C.L.). Combining\nthis result with those of two earlier experiments, Cuoricino and CUORE-0, we\nfind $T^{0\\nu}_{1/2}$($^{130}$Te) > $1.5\\times 10^{25}$ yr (90% C.L.), which is\nthe most stringent limit to date on this decay. Interpreting this result as a\nlimit on the effective Majorana neutrino mass, we find $m_{\\beta\\beta}<(110 -\n520)$ meV, where the range reflects the nuclear matrix element estimates\nemployed.\n",
"title": "First Results from CUORE: A Search for Lepton Number Violation via $0νββ$ Decay of $^{130}$Te"
} | null | null | null | null | true | null | 20479 | null | Default | null | null |
null | {
"abstract": " Recently, the Cauchy-Carlitz number was defined as the counterpart of the\nBernoulli-Carlitz number. Both numbers can be expressed explicitly in terms of\nso-called Stirling-Carlitz numbers. In this paper, we study the second analogue\nof Stirling-Carlitz numbers and give some general formulae, including Bernoulli\nand Cauchy numbers in formal power series with complex coefficients, and\nBernoulli-Carlitz and Cauchy-Carlitz numbers in function fields. We also give\nsome applications of Hasse-Teichmüller derivative to hypergeometric Bernoulli\nand Cauchy numbers in terms of associated Stirling numbers.\n",
"title": "Bernoulli-Carlitz and Cauchy-Carlitz numbers with Stirling-Carlitz numbers"
} | null | null | [
"Mathematics"
]
| null | true | null | 20480 | null | Validated | null | null |
null | {
"abstract": " Networks provide an informative, yet non-redundant description of complex\nsystems only if links represent truly dyadic relationships that cannot be\ndirectly traced back to node-specific properties such as size, importance, or\ncoordinates in some embedding space. In any real-world network, some links may\nbe reducible, and others irreducible, to such local properties. This dichotomy\npersists despite the steady increase in data availability and resolution, which\nactually determines an even stronger need for filtering techniques aimed at\ndiscerning essential links from non-essential ones. Here we introduce a\nrigorous method that, for any desired level of statistical significance,\noutputs the network backbone that is irreducible to the local properties of\nnodes, i.e. their degrees and strengths. Unlike previous approaches, our method\nemploys an exact maximum-entropy formulation guaranteeing that the filtered\nnetwork encodes only the links that cannot be inferred from local information.\nExtensive empirical analysis confirms that this approach uncovers essential\nbackbones that are otherwise hidden amidst many redundant relationships and\ninaccessible to other methods. For instance, we retrieve the hub-and-spoke\nskeleton of the US airport network and many specialised patterns of\ninternational trade. Being irreducible to local transportation and economic\nconstraints of supply and demand, these backbones single out genuinely\nhigher-order wiring principles.\n",
"title": "Irreducible network backbones: unbiased graph filtering via maximum entropy"
} | null | null | null | null | true | null | 20481 | null | Default | null | null |
null | {
"abstract": " We present Synkhronos, an extension to Theano for multi-GPU computations\nleveraging data parallelism. Our framework provides automated execution and\nsynchronization across devices, allowing users to continue to write serial\nprograms without risk of race conditions. The NVIDIA Collective Communication\nLibrary is used for high-bandwidth inter-GPU communication. Further\nenhancements to the Theano function interface include input slicing (with\naggregation) and input indexing, which perform common data-parallel computation\npatterns efficiently. One example use case is synchronous SGD, which has\nrecently been shown to scale well for a growing set of deep learning problems.\nWhen training ResNet-50, we achieve a near-linear speedup of 7.5x on an NVIDIA\nDGX-1 using 8 GPUs, relative to Theano-only code running a single GPU in\nisolation. Yet Synkhronos remains general to any data-parallel computation\nprogrammable in Theano. By implementing parallelism at the level of individual\nTheano functions, our framework uniquely addresses a niche between manual\nmulti-device programming and prescribed multi-GPU training routines.\n",
"title": "Synkhronos: a Multi-GPU Theano Extension for Data Parallelism"
} | null | null | [
"Computer Science"
]
| null | true | null | 20482 | null | Validated | null | null |
null | {
"abstract": " Cell division site positioning is precisely regulated to generate correctly\nsized and shaped daughters. We uncover a novel strategy to position the FtsZ\ncytokinetic ring at midcell in the social bacterium Myxococcus xanthus. PomX,\nPomY and the nucleoid-binding ParA/MinD ATPase PomZ self-assemble forming a\nlarge nucleoid-associated complex that localizes at the division site before\nFtsZ to directly guide and stimulate division. PomXYZ localization is generated\nthrough self-organized biased random motion on the nucleoid towards midcell and\nconstrained motion at midcell. Experiments and theory show that PomXYZ motion\nis produced by diffusive PomZ fluxes on the nucleoid into the complex. Flux\ndifferences scale with the intracellular asymmetry of the complex and are\nconverted into a local PomZ concentration gradient across the complex with\ntranslocation towards the higher PomZ concentration. At midcell, fluxes\nequalize resulting in constrained motion. Flux-based mechanisms may represent a\ngeneral paradigm for positioning of macromolecular structures in bacteria.\n",
"title": "The PomXYZ Proteins Self-Organize on the Bacterial Nucleoid to Stimulate Cell Division"
} | null | null | [
"Quantitative Biology"
]
| null | true | null | 20483 | null | Validated | null | null |
null | {
"abstract": " We consider a system of polynomials $f_1,\\ldots, f_R\\in\n\\mathbb{Z}[x_1,\\ldots, x_n]$ of the same degree with non-singular local zeros\nand in many variables. Generalising the work of Birch (1962) we find\nquantitative asymptotics (in terms of the maximum of the absolute value of the\ncoefficients of these polynomials) for the number of integer zeros of this\nsystem within a growing box. Using a quantitative version of the\nNullstellensatz, we obtain a quantitative strong approximation result, i.e. an\nupper bound on the smallest integer zero provided the system of polynomials is\nnon-singular.\n",
"title": "Quantitative Results on Diophantine Equations in Many Variables"
} | null | null | null | null | true | null | 20484 | null | Default | null | null |
null | {
"abstract": " Let $G = GL_N$ over an algebraically closed field of odd characteristic, and\n$\\theta$ an involutive automorphism on $G$ such that $H = (G^{\\theta})^0$ is\nisomorphic to $SO_N$. Then $G^{\\iota\\theta} = \\{ g \\in G \\mid \\theta(g) =\ng^{-1} \\}$ is regarded as a symmetric space $G/G^{\\theta}$. Let\n$G^{\\iota\\theta}_{uni}$ be the set of unipotent elements in $G^{\\iota\\theta}$.\n$H$ acts on $G^{\\iota\\theta}_{uni}$ by the conjugation. As an analogue of the\ngeneralized Springer correspondence in the case of reductive groups, we\nestablish in this paper the generalized Springer correspondence between\n$H$-orbits in $G^{\\iota\\theta}_{uni}$ and irreducible representations of\nvarious symmetric groups.\n",
"title": "Generalized Springer correspondence for symmetric spaces associated to orthogonal groups"
} | null | null | null | null | true | null | 20485 | null | Default | null | null |
null | {
"abstract": " Our work focuses on the problem of predicting the transfer of pediatric\npatients from the general ward of a hospital to the pediatric intensive care\nunit. Using data collected over 5.5 years from the electronic health records of\ntwo medical facilities, we develop classifiers based on adaptive boosting and\ngradient tree boosting. We further combine these learned classifiers into an\nensemble model and compare its performance to a modified pediatric early\nwarning score (PEWS) baseline that relies on expert defined guidelines. To\ngauge model generalizability, we perform an inter-facility evaluation where we\ntrain our algorithm on data from one facility and perform evaluation on a\nhidden test dataset from a separate facility. We show that improvements are\nwitnessed over the PEWS baseline in accuracy (0.77 vs. 0.69), sensitivity (0.80\nvs. 0.68), specificity (0.74 vs. 0.70) and AUROC (0.85 vs. 0.73).\n",
"title": "An Ensemble Boosting Model for Predicting Transfer to the Pediatric Intensive Care Unit"
} | null | null | null | null | true | null | 20486 | null | Default | null | null |
null | {
"abstract": " We propose a minority route choice game to investigate the effect of the\nnetwork structure on traffic network performance under the assumption of\ndrivers' bounded rationality. We investigate ring-and-hub topologies to capture\nthe nature of traffic networks in cities, and employ a minority game-based\ninductive learning process to model the characteristic behavior under the route\nchoice scenario. Through numerical experiments, we find that topological\nchanges in traffic networks induce a phase transition from an uncongested phase\nto a congested phase. Understanding this phase transition is helpful in\nplanning new traffic networks.\n",
"title": "Effects of Network Structure on the Performance of a Modeled Traffic Network under Drivers' Bounded Rationality"
} | null | null | null | null | true | null | 20487 | null | Default | null | null |
null | {
"abstract": " Massive content about user's social, personal and professional life stored on\nOnline Social Networks (OSNs) has attracted not only the attention of\nresearchers and social analysts but also the cyber criminals. These cyber\ncriminals penetrate illegally into an OSN by establishing fake profiles or by\ndesigning bots and exploit the vulnerabilities of an OSN to carry out illegal\nactivities. With the growth of technology cyber crimes have been increasing\nmanifold. Daily reports of the security and privacy threats in the OSNs demand\nnot only the intelligent automated detection systems that can identify and\nalleviate fake profiles in real time but also the reinforcement of the security\nand privacy laws to curtail the cyber crime. In this paper, we have studied\nvarious categories of fake profiles like compromised profiles, cloned profiles\nand online bots (spam-bots, social-bots, like-bots and influential-bots) on\ndifferent OSN sites along with existing cyber laws to mitigate their threats.\nIn order to design fake profile detection systems, we have highlighted\ndifferent category of fake profile features which are capable to distinguish\ndifferent kinds of fake entities from real ones. Another major challenges faced\nby researchers while building the fake profile detection systems is the\nunavailability of data specific to fake users. The paper addresses this\nchallenge by providing extremely obliging data collection techniques along with\nsome existing data sources. Furthermore, an attempt is made to present several\nmachine learning techniques employed to design different fake profile detection\nsystems.\n",
"title": "Sneak into Devil's Colony- A study of Fake Profiles in Online Social Networks and the Cyber Law"
} | null | null | null | null | true | null | 20488 | null | Default | null | null |
null | {
"abstract": " Networked system often relies on distributed algorithms to achieve a global\ncomputation goal with iterative local information exchanges between neighbor\nnodes. To preserve data privacy, a node may add a random noise to its original\ndata for information exchange at each iteration. Nevertheless, a neighbor node\ncan estimate other's original data based on the information it received. The\nestimation accuracy and data privacy can be measured in terms of $(\\epsilon,\n\\delta)$-data-privacy, defined as the probability of $\\epsilon$-accurate\nestimation (the difference of an estimation and the original data is within\n$\\epsilon$) is no larger than $\\delta$ (the disclosure probability). How to\noptimize the estimation and analyze data privacy is a critical and open issue.\nIn this paper, a theoretical framework is developed to investigate how to\noptimize the estimation of neighbor's original data using the local information\nreceived, named optimal distributed estimation. Then, we study the disclosure\nprobability under the optimal estimation for data privacy analysis. We further\napply the developed framework to analyze the data privacy of the\nprivacy-preserving average consensus algorithm and identify the optimal noises\nfor the algorithm.\n",
"title": "Preserving Data-Privacy with Added Noises: Optimal Estimation and Privacy Analysis"
} | null | null | null | null | true | null | 20489 | null | Default | null | null |
null | {
"abstract": " Catalytic swimmers have attracted much attention as alternatives to\nbiological systems for examining collective microscopic dynamics and the\nresponse to physico-chemical signals. Yet, understanding and predicting even\nthe most fundamental characteristics of their individual propulsion still\nraises important challenges. While chemical asymmetry is widely recognized as\nthe cornerstone of catalytic propulsion, different experimental studies have\nreported that particles with identical chemical properties may propel in\nopposite directions. Here, we show that, beyond its chemical properties, the\ndetailed shape of a catalytic swimmer plays an essential role in determining\nits direction of motion, demonstrating the compatibility of the classical\ntheoretical framework with experimental observations.\n",
"title": "Geometric tuning of self-propulsion for Janus catalytic particles"
} | null | null | null | null | true | null | 20490 | null | Default | null | null |
null | {
"abstract": " We consider the Bogolubov-de Gennes equations giving an equivalent\nformulation of the BCS theory of superconductivity. We are interested in the\ncase when the magnetic field is present. We (a) discuss their general features,\n(b) isolate key physical classes of solutions (normal, vortex and vortex\nlattice states) and (c) prove existence of the normal, vortex and vortex\nlattice states and stability/instability of the normal states for large/small\ntemperature or/and magnetic fields.\n",
"title": "On the Bogolubov-de Gennes Equations"
} | null | null | null | null | true | null | 20491 | null | Default | null | null |
null | {
"abstract": " Penalized likelihood methods are widely used for high-dimensional regression.\nAlthough many methods have been proposed and the associated theory is now\nwell-developed, the relative efficacy of different methods in finite-sample\nsettings, as encountered in practice, remains incompletely understood. There is\ntherefore a need for empirical investigations in this area that can offer\npractical insight and guidance to users of these methods. In this paper we\npresent a large-scale comparison of penalized regression methods. We\ndistinguish between three related goals: prediction, variable selection and\nvariable ranking. Our results span more than 1,800 data-generating scenarios,\nallowing us to systematically consider the influence of various factors (sample\nsize, dimensionality, sparsity, signal strength and multicollinearity). We\nconsider several widely-used methods (Lasso, Elastic Net, Ridge Regression,\nSCAD, the Dantzig Selector as well as Stability Selection). We find\nconsiderable variation in performance between methods, with results dependent\non details of the data-generating scenario and the specific goal. Our results\nsupport a `no panacea' view, with no unambiguous winner across all scenarios,\neven in this restricted setting where all data align well with the assumptions\nunderlying the methods. Lasso is well-behaved, performing competitively in many\nscenarios, while SCAD is highly variable. Substantial benefits from a\nRidge-penalty are only seen in the most challenging scenarios with strong\nmulti-collinearity. The results are supported by semi-synthetic analyzes using\ngene expression data from cancer samples. Our empirical results complement\nexisting theory and provide a resource to compare methods across a range of\nscenarios and metrics.\n",
"title": "High-dimensional regression in practice: an empirical study of finite-sample prediction, variable selection and ranking"
} | null | null | null | null | true | null | 20492 | null | Default | null | null |
null | {
"abstract": " Machine learning is a field of computer science that builds algorithms that\nlearn. In many cases, machine learning algorithms are used to recreate a human\nability like adding a caption to a photo, driving a car, or playing a game.\nWhile the human brain has long served as a source of inspiration for machine\nlearning, little effort has been made to directly use data collected from\nworking brains as a guide for machine learning algorithms. Here we demonstrate\na new paradigm of \"neurally-weighted\" machine learning, which takes fMRI\nmeasurements of human brain activity from subjects viewing images, and infuses\nthese data into the training process of an object recognition learning\nalgorithm to make it more consistent with the human brain. After training,\nthese neurally-weighted classifiers are able to classify images without\nrequiring any additional neural data. We show that our neural-weighting\napproach can lead to large performance gains when used with traditional machine\nvision features, as well as to significant improvements with already\nhigh-performing convolutional neural network features. The effectiveness of\nthis approach points to a path forward for a new class of hybrid machine\nlearning algorithms which take both inspiration and direct constraints from\nneuronal data.\n",
"title": "Using Human Brain Activity to Guide Machine Learning"
} | null | null | [
"Computer Science"
]
| null | true | null | 20493 | null | Validated | null | null |
null | {
"abstract": " We use a deep learning model trained only on a patient's blood oxygenation\ndata (measurable with an inexpensive fingertip sensor) to predict impending\nhypoxemia (low blood oxygen) more accurately than trained anesthesiologists\nwith access to all the data recorded in a modern operating room. We also\nprovide a simple way to visualize the reason why a patient's risk is low or\nhigh by assigning weight to the patient's past blood oxygen values. This work\nhas the potential to provide cutting-edge clinical decision support in\nlow-resource settings, where rates of surgical complication and death are\nsubstantially greater than in high-resource areas.\n",
"title": "Anesthesiologist-level forecasting of hypoxemia with only SpO2 data using deep learning"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20494 | null | Validated | null | null |
null | {
"abstract": " We investigate the possibility of extending the non-functionally complete\nlogic of a collection of Boolean connectives by the addition of further Boolean\nconnectives that make the resulting set of connectives functionally complete.\nMore precisely, we will be interested in checking whether an axiomatization for\nClassical Propositional Logic may be produced by merging Hilbert-style calculi\nfor two disjoint incomplete fragments of it. We will prove that the answer to\nthat problem is a negative one, unless one of the components includes only\ntop-like connectives.\n",
"title": "Merging fragments of classical logic"
} | null | null | null | null | true | null | 20495 | null | Default | null | null |
null | {
"abstract": " We propose a new variational Bayes estimator for high-dimensional copulas\nwith discrete, or a combination of discrete and continuous, margins. The method\nis based on a variational approximation to a tractable augmented posterior, and\nis faster than previous likelihood-based approaches. We use it to estimate\ndrawable vine copulas for univariate and multivariate Markov ordinal and mixed\ntime series. These have dimension $rT$, where $T$ is the number of observations\nand $r$ is the number of series, and are difficult to estimate using previous\nmethods. The vine pair-copulas are carefully selected to allow for\nheteroskedasticity, which is a feature of most ordinal time series data. When\ncombined with flexible margins, the resulting time series models also allow for\nother common features of ordinal data, such as zero inflation, multiple modes\nand under- or over-dispersion. Using six example series, we illustrate both the\nflexibility of the time series copula models, and the efficacy of the\nvariational Bayes estimator for copulas of up to 792 dimensions and 60\nparameters. This far exceeds the size and complexity of copula models for\ndiscrete data that can be estimated using previous methods.\n",
"title": "Variational Bayes Estimation of Discrete-Margined Copula Models with Application to Time Series"
} | null | null | null | null | true | null | 20496 | null | Default | null | null |
null | {
"abstract": " Scene modeling is very crucial for robots that need to perceive, reason about\nand manipulate the objects in their environments. In this paper, we adapt and\nextend Boltzmann Machines (BMs) for contextualized scene modeling. Although\nthere are many models on the subject, ours is the first to bring together\nobjects, relations, and affordances in a highly-capable generative model. For\nthis end, we introduce a hybrid version of BMs where relations and affordances\nare introduced with shared, tri-way connections into the model. Moreover, we\ncontribute a dataset for relation estimation and modeling studies. We evaluate\nour method in comparison with several baselines on object estimation,\nout-of-context object detection, relation estimation, and affordance estimation\ntasks. Moreover, to illustrate the generative capability of the model, we show\nseveral example scenes that the model is able to generate.\n",
"title": "COSMO: Contextualized Scene Modeling with Boltzmann Machines"
} | null | null | null | null | true | null | 20497 | null | Default | null | null |
null | {
"abstract": " In order to perform complex actions in human environments, an autonomous\nrobot needs the ability to understand the environment, that is, to gather and\nmaintain spatial knowledge. Topological map is commonly used for representing\nlarge scale, global maps such as floor plans. Although much work has been done\nin topological map extraction, we have found little previous work on the\nproblem of learning the topological map using a probabilistic model. Learning a\ntopological map means learning the structure of the large-scale space and\ndependency between places, for example, how the evidence of a group of places\ninfluence the attributes of other places. This is an important step towards\nplanning complex actions in the environment. In this thesis, we consider the\nproblem of using probabilistic deep learning model to learn the topological\nmap, which is essentially a sparse undirected graph where nodes represent\nplaces annotated with their semantic attributes (e.g. place category). We\npropose to use a novel probabilistic deep model, Sum-Product Networks (SPNs),\ndue to their unique properties. We present two methods for learning topological\nmaps using SPNs: the place grid method and the template-based method. We\ncontribute an algorithm that builds SPNs for graphs using template models. Our\nexperiments evaluate the ability of our models to enable robots to infer\nsemantic attributes and detect maps with novel semantic attribute arrangements.\nOur results demonstrate their understanding of the topological map structure\nand spatial relations between places.\n",
"title": "Learning Large-Scale Topological Maps Using Sum-Product Networks"
} | null | null | null | null | true | null | 20498 | null | Default | null | null |
null | {
"abstract": " We study numerically the entanglement entropy and spatial correlations of the\none dimensional transverse field Ising model with three different\nperturbations. First, we focus on the out of equilibrium, steady state with an\nenergy current passing through the system. By employing a variety of\nmatrix-product state based methods, we confirm the phase diagram and compute\nthe entanglement entropy. Second, we consider a small perturbation that takes\nthe system away from integrability and calculate the correlations, the central\ncharge and the entanglement entropy. Third, we consider periodically weakened\nbonds, exploring the phase diagram and entanglement properties first in the\nsituation when the weak and strong bonds alternate (period two-bonds) and then\nthe general situation of a period of n bonds. In the latter case we find a\ncritical weak bond that scales with the transverse field as $J'_c/J$ =\n$(h/J)^n$, where $J$ is the strength of the strong bond, $J'$ of the weak bond\nand $h$ the transverse field. We explicitly show that the energy current is not\na conserved quantity in this case.\n",
"title": "Entanglement scaling and spatial correlations of the transverse field Ising model with perturbations"
} | null | null | null | null | true | null | 20499 | null | Default | null | null |
null | {
"abstract": " Strongly interacting many-body systems consisting of fermions or bosons can\nhost exotic quasiparticles with anyonic statistics. Here, we demonstrate that\nmany-body systems of anyons can also form anyonic quasi-particles. The charge\nand statistics of the emergent anyons can be different from those of the\noriginal anyons.\n",
"title": "Anyonic excitations of hardcore anyons"
} | null | null | [
"Physics"
]
| null | true | null | 20500 | null | Validated | null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.