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": " For high-dimensional sparse linear models, how to construct confidence\nintervals for coefficients remains a difficult question. The main reason is the\ncomplicated limiting distributions of common estimators such as the Lasso.\nSeveral confidence interval construction methods have been developed, and\nBootstrap Lasso+OLS is notable for its simple technicality, good\ninterpretability, and comparable performance with other more complicated\nmethods. However, Bootstrap Lasso+OLS depends on the beta-min assumption, a\ntheoretic criterion that is often violated in practice. In this paper, we\nintroduce a new method called Bootstrap Lasso+Partial Ridge (LPR) to relax this\nassumption. LPR is a two-stage estimator: first using Lasso to select features\nand subsequently using Partial Ridge to refit the coefficients. Simulation\nresults show that Bootstrap LPR outperforms Bootstrap Lasso+OLS when there\nexist small but non-zero coefficients, a common situation violating the\nbeta-min assumption. For such coefficients, compared to Bootstrap Lasso+OLS,\nconfidence intervals constructed by Bootstrap LPR have on average 50% larger\ncoverage probabilities. Bootstrap LPR also has on average 35% shorter\nconfidence interval lengths than the de-sparsified Lasso methods, regardless of\nwhether linear models are misspecified. Additionally, we provide theoretical\nguarantees of Bootstrap LPR under appropriate conditions and implement it in\nthe R package \"HDCI.\"\n",
"title": "A Bootstrap Lasso + Partial Ridge Method to Construct Confidence Intervals for Parameters in High-dimensional Sparse Linear Models"
}
| null | null | null | null | true | null |
3201
| null |
Default
| null | null |
null |
{
"abstract": " At the core of many important machine learning problems faced by online\nstreaming services is a need to model how users interact with the content.\nThese problems can often be reduced to a combination of 1) sequentially\nrecommending items to the user, and 2) exploiting the user's interactions with\nthe items as feedback for the machine learning model. Unfortunately, there are\nno public datasets currently available that enable researchers to explore this\ntopic. In order to spur that research, we release the Music Streaming Sessions\nDataset (MSSD), which consists of approximately 150 million listening sessions\nand associated user actions. Furthermore, we provide audio features and\nmetadata for the approximately 3.7 million unique tracks referred to in the\nlogs. This is the largest collection of such track metadata currently available\nto the public. This dataset enables research on important problems including\nhow to model user listening and interaction behaviour in streaming, as well as\nMusic Information Retrieval (MIR), and session-based sequential\nrecommendations.\n",
"title": "The Music Streaming Sessions Dataset"
}
| null | null | null | null | true | null |
3202
| null |
Default
| null | null |
null |
{
"abstract": " Class imbalance classification is a challenging research problem in data\nmining and machine learning, as most of the real-life datasets are often\nimbalanced in nature. Existing learning algorithms maximise the classification\naccuracy by correctly classifying the majority class, but misclassify the\nminority class. However, the minority class instances are representing the\nconcept with greater interest than the majority class instances in real-life\napplications. Recently, several techniques based on sampling methods\n(under-sampling of the majority class and over-sampling the minority class),\ncost-sensitive learning methods, and ensemble learning have been used in the\nliterature for classifying imbalanced datasets. In this paper, we introduce a\nnew clustering-based under-sampling approach with boosting (AdaBoost)\nalgorithm, called CUSBoost, for effective imbalanced classification. The\nproposed algorithm provides an alternative to RUSBoost (random under-sampling\nwith AdaBoost) and SMOTEBoost (synthetic minority over-sampling with AdaBoost)\nalgorithms. We evaluated the performance of CUSBoost algorithm with the\nstate-of-the-art methods based on ensemble learning like AdaBoost, RUSBoost,\nSMOTEBoost on 13 imbalance binary and multi-class datasets with various\nimbalance ratios. The experimental results show that the CUSBoost is a\npromising and effective approach for dealing with highly imbalanced datasets.\n",
"title": "CUSBoost: Cluster-based Under-sampling with Boosting for Imbalanced Classification"
}
| null | null | null | null | true | null |
3203
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we study the problem of discovering a timeline of events in a\ntemporal network. We model events as dense subgraphs that occur within\nintervals of network activity. We formulate the event-discovery task as an\noptimization problem, where we search for a partition of the network timeline\ninto k non-overlapping intervals, such that the intervals span subgraphs with\nmaximum total density. The output is a sequence of dense subgraphs along with\ncorresponding time intervals, capturing the most interesting events during the\nnetwork lifetime.\nA naive solution to our optimization problem has polynomial but prohibitively\nhigh running time complexity. We adapt existing recent work on dynamic\ndensest-subgraph discovery and approximate dynamic programming to design a fast\napproximation algorithm. Next, to ensure richer structure, we adjust the\nproblem formulation to encourage coverage of a larger set of nodes. This\nproblem is NP-hard even for static graphs. However, on static graphs a simple\ngreedy algorithm leads to approximate solution due to submodularity. We\nextended this greedy approach for the case of temporal networks. However, the\napproximation guarantee does not hold. Nevertheless, according to the\nexperiments, the algorithm finds good quality solutions.\n",
"title": "Finding events in temporal networks: Segmentation meets densest-subgraph discovery"
}
| null | null |
[
"Computer Science"
] | null | true | null |
3204
| null |
Validated
| null | null |
null |
{
"abstract": " We study configuration spaces of linkages whose underlying graph are polygons\nwith diagonal constrains, or more general, partial two-trees. We show that\n(with an appropriate definition) the oriented area is a Bott-Morse function on\nthe configuration space. Its critical points are described and Bott-Morse\nindices are computed. This paper is a generalization of analogous results for\npolygonal linkages (obtained earlier by G. Khimshiashvili, G. Panina, and A.\nZhukova).\n",
"title": "On the area of constrained polygonal linkages"
}
| null | null | null | null | true | null |
3205
| null |
Default
| null | null |
null |
{
"abstract": " The small-cluster exact-diagonalization calculations and the projector\nquantum Monte Carlo method are used to examine the competing effects of\ngeometrical frustration and interaction on ferromagnetism in the Hubbard model\non the Shastry-Sutherland lattice. It is shown that the geometrical frustration\nstabilizes the ferromagnetic state at high electron concentrations ($n \\gtrsim\n7/4$), where strong correlations between ferromagnetism and the shape of the\nnoninteracting density of states are observed. In particular, it is found that\nferromagnetism is stabilized only for these values of frustration parameters,\nwhich lead to the single peaked noninterating density of states at the band\nedge. Once, two or more peaks appear in the noninteracting density of states at\nthe band egde the ferromagnetic state is suppressed. This opens a new route\ntowards the understanding of ferromagnetism in strongly correlated systems.\n",
"title": "Effects of geometrical frustration on ferromagnetism in the Hubbard model on the Shastry-Sutherland lattice"
}
| null | null | null | null | true | null |
3206
| null |
Default
| null | null |
null |
{
"abstract": " Process discovery techniques return process models that are either formal\n(precisely describing the possible behaviors) or informal (merely a \"picture\"\nnot allowing for any form of formal reasoning). Formal models are able to\nclassify traces (i.e., sequences of events) as fitting or non-fitting. Most\nprocess mining approaches described in the literature produce such models. This\nis in stark contrast with the over 25 available commercial process mining tools\nthat only discover informal process models that remain deliberately vague on\nthe precise set of possible traces. There are two main reasons why vendors\nresort to such models: scalability and simplicity. In this paper, we propose to\ncombine the best of both worlds: discovering hybrid process models that have\nformal and informal elements. As a proof of concept we present a discovery\ntechnique based on hybrid Petri nets. These models allow for formal reasoning,\nbut also reveal information that cannot be captured in mainstream formal\nmodels. A novel discovery algorithm returning hybrid Petri nets has been\nimplemented in ProM and has been applied to several real-life event logs. The\nresults clearly demonstrate the advantages of remaining \"vague\" when there is\nnot enough \"evidence\" in the data or standard modeling constructs do not \"fit\".\nMoreover, the approach is scalable enough to be incorporated in\nindustrial-strength process mining tools.\n",
"title": "Learning Hybrid Process Models From Events: Process Discovery Without Faking Confidence"
}
| null | null |
[
"Computer Science"
] | null | true | null |
3207
| null |
Validated
| null | null |
null |
{
"abstract": " Formal languages theory is useful for the study of natural language. In\nparticular, it is of interest to study the adequacy of the grammatical\nformalisms to express syntactic phenomena present in natural language. First,\nit helps to draw hypothesis about the nature and complexity of the\nspeaker-hearer linguistic competence, a fundamental question in linguistics and\nother cognitive sciences. Moreover, from an engineering point of view, it\nallows the knowledge of practical limitations of applications based on those\nformalisms. In this article I introduce the adequacy problem of grammatical\nformalisms for natural language, also introducing some formal language theory\nconcepts required for this discussion. Then, I review the formalisms that have\nbeen proposed in history, and the arguments that have been given to support or\nreject their adequacy.\n-----\nLa teoría de lenguajes formales es útil para el estudio de los lenguajes\nnaturales. En particular, resulta de interés estudiar la adecuación de los\nformalismos gramaticales para expresar los fenómenos sintácticos presentes\nen el lenguaje natural. Primero, ayuda a trazar hipótesis acerca de la\nnaturaleza y complejidad de las competencias lingüísticas de los\nhablantes-oyentes del lenguaje, un interrogante fundamental de la\nlingüística y otras ciencias cognitivas. Además, desde el punto de vista\nde la ingeniería, permite conocer limitaciones prácticas de las\naplicaciones basadas en dichos formalismos. En este artículo hago una\nintroducción al problema de la adecuación de los formalismos gramaticales\npara el lenguaje natural, introduciendo también algunos conceptos de la\nteoría de lenguajes formales necesarios para esta discusión. Luego, hago un\nrepaso de los formalismos que han sido propuestos a lo largo de la historia, y\nde los argumentos que se han dado para sostener o refutar su adecuación.\n",
"title": "El Lenguaje Natural como Lenguaje Formal"
}
| null | null |
[
"Computer Science"
] | null | true | null |
3208
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper, we find an upper bound for the CP-rank of a matrix over a\ntropical semiring, according to the vertex clique cover of the graph prescribed\nby the pattern of the matrix. We study the graphs that beget the patterns of\nmatrices with the lowest possible CP-ranks and prove that any such graph must\nhave its diameter equal to 2.\n",
"title": "Bounds for the completely positive rank of a symmetric matrix over a tropical semiring"
}
| null | null | null | null | true | null |
3209
| null |
Default
| null | null |
null |
{
"abstract": " This paper proposes a new algorithm for Gaussian process classification based\non posterior linearisation (PL). In PL, a Gaussian approximation to the\nposterior density is obtained iteratively using the best possible linearisation\nof the conditional mean of the labels and accounting for the linearisation\nerror. Considering three widely-used likelihood functions, in general, PL\nprovides lower classification errors in real data sets than expectation\npropagation and Laplace algorithms.\n",
"title": "Gaussian process classification using posterior linearisation"
}
| null | null |
[
"Statistics"
] | null | true | null |
3210
| null |
Validated
| null | null |
null |
{
"abstract": " Pairwise comparisons are an important tool of modern (multiple criteria)\ndecision making. Since human judgments are often inconsistent, many studies\nfocused on the ways how to express and measure this inconsistency, and several\ninconsistency indices were proposed as an alternative to Saaty inconsistency\nindex and inconsistency ratio for reciprocal pairwise comparisons matrices.\nThis paper aims to: firstly, introduce a new measure of inconsistency of\npairwise comparisons and to prove its basic properties; secondly, to postulate\nan additional axiom, an upper boundary axiom, to an existing set of axioms; and\nthe last, but not least, the paper provides proofs of satisfaction of this\nadditional axiom by selected inconsistency indices as well as it provides their\nnumerical comparison.\n",
"title": "On Inconsistency Indices and Inconsistency Axioms in Pairwise Comparisons"
}
| null | null |
[
"Computer Science"
] | null | true | null |
3211
| null |
Validated
| null | null |
null |
{
"abstract": " We study the existence of monotone heteroclinic traveling waves for a general\nFisher-Burgers equation with nonlinear and possibly density-dependent\ndiffusion. Such a model arises, for instance, in physical phenomena where a\nsaturation effect appears for large values of the gradient. We give an estimate\nfor the critical speed (namely, the first speed for which a monotone\nheteroclinic traveling wave exists) for some different shapes of the reaction\nterm, and we analyze its dependence on a small real parameter when this brakes\nthe diffusion, complementing our study with some numerical simulations.\n",
"title": "Heteroclinic traveling fronts for a generalized Fisher-Burgers equation with saturating diffusion"
}
| null | null | null | null | true | null |
3212
| null |
Default
| null | null |
null |
{
"abstract": " I describe a method for computer algebra that helps with laborious\ncalculations typically encountered in theoretical microhydrodynamics. The\nprogram mimics how humans calculate by matching patterns and making\nreplacements according to the rules of algebra and calculus. This note gives an\noverview and walks through an example, while the accompanying code repository\ncontains the implementation details, a tutorial, and more examples. The code\nrepository is attached as supplementary material to this note, and maintained\nat this https URL\n",
"title": "Computer Algebra for Microhydrodynamics"
}
| null | null |
[
"Computer Science",
"Physics"
] | null | true | null |
3213
| null |
Validated
| null | null |
null |
{
"abstract": " This paper describes a micro fluorescence in situ hybridization\n({\\mu}FISH)-based rapid detection of cytogenetic biomarkers on formalin-fixed\nparaffin embedded (FFPE) tissue sections. We demonstrated this method in the\ncontext of detecting human epidermal growth factor 2 (HER2) in breast tissue\nsections. This method uses a non-contact microfluidic scanning probe (MFP),\nwhich localizes FISH probes at the micrometer length-scale to selected cells of\nthe tissue section. The scanning ability of the MFP allows for a versatile\nimplementation of FISH on tissue sections. We demonstrated the use of\noligonucleotide FISH probes in ethylene carbonate-based buffer enabling rapid\nhybridization within < 1 min for chromosome enumeration and 10-15 min for\nassessment of the HER2 status in FFPE sections. We further demonstrated\nrecycling of FISH probes for multiple sequential tests using a defined volume\nof probes by forming hierarchical hydrodynamic flow confinements. This\nmicroscale method is compatible with the standard FISH protocols and with the\nInstant Quality (IQ) FISH assay, reduces the FISH probe consumption ~100-fold\nand the hybridization time 4-fold, resulting in an assay turnaround time of < 3\nh. We believe rapid {\\mu}FISH has the potential of being used in pathology\nworkflows as a standalone method or in combination with other molecular methods\nfor diagnostic and prognostic analysis of FFPE sections.\n",
"title": "Rapid micro fluorescence in situ hybridization in tissue sections"
}
| null | null | null | null | true | null |
3214
| null |
Default
| null | null |
null |
{
"abstract": " Neural time-series data contain a wide variety of prototypical signal\nwaveforms (atoms) that are of significant importance in clinical and cognitive\nresearch. One of the goals for analyzing such data is hence to extract such\n'shift-invariant' atoms. Even though some success has been reported with\nexisting algorithms, they are limited in applicability due to their heuristic\nnature. Moreover, they are often vulnerable to artifacts and impulsive noise,\nwhich are typically present in raw neural recordings. In this study, we address\nthese issues and propose a novel probabilistic convolutional sparse coding\n(CSC) model for learning shift-invariant atoms from raw neural signals\ncontaining potentially severe artifacts. In the core of our model, which we\ncall $\\alpha$CSC, lies a family of heavy-tailed distributions called\n$\\alpha$-stable distributions. We develop a novel, computationally efficient\nMonte Carlo expectation-maximization algorithm for inference. The maximization\nstep boils down to a weighted CSC problem, for which we develop a\ncomputationally efficient optimization algorithm. Our results show that the\nproposed algorithm achieves state-of-the-art convergence speeds. Besides,\n$\\alpha$CSC is significantly more robust to artifacts when compared to three\ncompeting algorithms: it can extract spike bursts, oscillations, and even\nreveal more subtle phenomena such as cross-frequency coupling when applied to\nnoisy neural time series.\n",
"title": "Learning the Morphology of Brain Signals Using Alpha-Stable Convolutional Sparse Coding"
}
| null | null | null | null | true | null |
3215
| null |
Default
| null | null |
null |
{
"abstract": " We describe two recently proposed machine learning approaches for discovering\nemerging trends in fatal accidental drug overdoses. The Gaussian Process Subset\nScan enables early detection of emerging patterns in spatio-temporal data,\naccounting for both the non-iid nature of the data and the fact that detecting\nsubtle patterns requires integration of information across multiple spatial\nareas and multiple time steps. We apply this approach to 17 years of\ncounty-aggregated data for monthly opioid overdose deaths in the New York City\nmetropolitan area, showing clear advantages in the utility of discovered\npatterns as compared to typical anomaly detection approaches.\nTo detect and characterize emerging overdose patterns that differentially\naffect a subpopulation of the data, including geographic, demographic, and\nbehavioral patterns (e.g., which combinations of drugs are involved), we apply\nthe Multidimensional Tensor Scan to 8 years of case-level overdose data from\nAllegheny County, PA. We discover previously unidentified overdose patterns\nwhich reveal unusual demographic clusters, show impacts of drug legislation,\nand demonstrate potential for early detection and targeted intervention. These\napproaches to early detection of overdose patterns can inform prevention and\nresponse efforts, as well as understanding the effects of policy changes.\n",
"title": "Machine Learning for Drug Overdose Surveillance"
}
| null | null | null | null | true | null |
3216
| null |
Default
| null | null |
null |
{
"abstract": " Polyphonic sound event detection (polyphonic SED) is an interesting but\nchallenging task due to the concurrence of multiple sound events. Recently, SED\nmethods based on convolutional neural networks (CNN) and recurrent neural\nnetworks (RNN) have shown promising performance. Generally, CNN are designed\nfor local feature extraction while RNN are used to model the temporal\ndependency among these local features. Despite their success, it is still\ninsufficient for existing deep learning techniques to separate individual sound\nevent from their mixture, largely due to the overlapping characteristic of\nfeatures. Motivated by the success of Capsule Networks (CapsNet), we propose a\nmore suitable capsule based approach for polyphonic SED. Specifically, several\ncapsule layers are designed to effectively select representative frequency\nbands for each individual sound event. The temporal dependency of capsule's\noutputs is then modeled by a RNN. And a dynamic threshold method is proposed\nfor making the final decision based on RNN outputs. Experiments on the TUT-SED\nSynthetic 2016 dataset show that the proposed approach obtains an F1-score of\n68.8% and an error rate of 0.45, outperforming the previous state-of-the-art\nmethod of 66.4% and 0.48, respectively.\n",
"title": "A Capsule based Approach for Polyphonic Sound Event Detection"
}
| null | null | null | null | true | null |
3217
| null |
Default
| null | null |
null |
{
"abstract": " The global gyrokinetic toroidal code (GTC) has been recently upgraded to do\nsimulations in non-axisymmetric equilibrium configuration, such as\nstellarators. Linear simulation of ion temperature gradient (ITG) driven\ninstabilities has been done in Wendelstein7-X (W7-X) and Large Helical Device\n(LHD) stellarators using GTC. Several results are discussed to study\ncharacteristics of ITG in stellarators, including toroidal grids convergence,\nnmodes number convergence, poloidal and parallel spectrums, and electrostatic\npotential mode structure on flux surface.\n",
"title": "Linear simulation of ion temperature gradient driven instabilities in W7-X and LHD stellarators using GTC"
}
| null | null | null | null | true | null |
3218
| null |
Default
| null | null |
null |
{
"abstract": " A nanofabrication process for realizing optical nanoantennas carved from a\nsingle-crystal gold plate is presented in this communication. The method relies\non synthesizing two-dimensional micron-size gold crystals followed by the dry\netching of a desired antenna layout. The fabrication of single-crystal optical\nnanoantennas with standard electron-beam lithography tool and dry etching\nreactor represents an alternative technological solution to focused ion beam\nmilling of the objects. The process is exemplified by engineering nanorod\nantennas. Dark-field spectroscopy indicates that optical antennas produced from\nsingle crystal flakes have reduced localized surface plasmon resonance losses\ncompared to amorphous designs of similar shape. The present process is easily\napplicable to other metals such as silver or copper and offers a design\nflexibility not found in crystalline particles synthesized by colloidal\nchemistry.\n",
"title": "Advanced engineering of single-crystal gold nanoantennas"
}
| null | null | null | null | true | null |
3219
| null |
Default
| null | null |
null |
{
"abstract": " We consider the connections among `clumped' residual allocation models\n(RAMs), a general class of stick-breaking processes including Dirichlet\nprocesses, and the occupation laws of certain discrete space time-inhomogeneous\nMarkov chains related to simulated annealing and other applications. An\nintermediate structure is introduced in a given RAM, where proportions between\nsuccessive indices in a list are added or clumped together to form another RAM.\nIn particular, when the initial RAM is a Griffiths-Engen-McCloskey (GEM)\nsequence and the indices are given by the random times that an auxiliary Markov\nchain jumps away from its current state, the joint law of the intermediate RAM\nand the locations visited in the sojourns is given in terms of a `disordered'\nGEM sequence, and an induced Markov chain. Through this joint law, we identify\na large class of `stick breaking' processes as the limits of empirical\noccupation measures for associated time-inhomogeneous Markov chains.\n",
"title": "Stick-breaking processes, clumping, and Markov chain occupation laws"
}
| null | null | null | null | true | null |
3220
| null |
Default
| null | null |
null |
{
"abstract": " Rydberg atoms have attracted considerable interest due to their huge\ninteraction among each other and with external fields. They demonstrate\ncharacteristic scaling laws in dependence on the principal quantum number $n$\nfor features such as the magnetic field for level crossing. While bearing\nstriking similarities to Rydberg atoms, fundamentally new insights may be\nobtained for Rydberg excitons, as the crystal environment gives easy optical\naccess to many states within an exciton multiplet. Here we study experimentally\nand theoretically the scaling of several characteristic parameters of Rydberg\nexcitons with $n$. From absorption spectra in magnetic field we find for the\nfirst crossing of levels with adjacent principal quantum numbers a $B_r \\propto\nn^{-4}$ dependence of the resonance field strength, $B_r$, due to the dominant\nparamagnetic term unlike in the atomic case where the diamagnetic contribution\nis decisive. By contrast, in electric field we find scaling laws just like for\nRydberg atoms. The resonance electric field strength scales as $E_r \\propto\nn^{-5}$. We observe anticrossings of the states belonging to multiplets with\ndifferent principal quantum numbers. The energy splittings at the avoided\ncrossings scale as $n^{-4}$ which we relate to the crystal specific deviation\nof the exciton Hamiltonian from the hydrogen model. We observe the exciton\npolarizability in the electric field to scale as $n^7$. In magnetic field the\ncrossover field strength from a hydrogen-like exciton to a magnetoexciton\ndominated by electron and hole Landau level quantization scales as $n^{-3}$.\nThe ionization voltages demonstrate a $n^{-4}$ scaling as for atoms. The width\nof the absorption lines remains constant before dissociation for high enough\n$n$, while for small $n \\lesssim 12$ an exponential increase with the field is\nfound. These results are in excellent agreement with theoretical calculations.\n",
"title": "Scaling laws of Rydberg excitons"
}
| null | null | null | null | true | null |
3221
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we propose a novel variable selection approach in the\nframework of multivariate linear models taking into account the dependence that\nmay exist between the responses. It consists in estimating beforehand the\ncovariance matrix of the responses and to plug this estimator in a Lasso\ncriterion, in order to obtain a sparse estimator of the coefficient matrix. The\nproperties of our approach are investigated both from a theoretical and a\nnumerical point of view. More precisely, we give general conditions that the\nestimators of the covariance matrix and its inverse have to satisfy in order to\nrecover the positions of the null and non null entries of the coefficient\nmatrix when the size of the covariance matrix is not fixed and can tend to\ninfinity. We prove that these conditions are satisfied in the particular case\nof some Toeplitz matrices. Our approach is implemented in the R package\nMultiVarSel available from the Comprehensive R Archive Network (CRAN) and is\nvery attractive since it benefits from a low computational load. We also assess\nthe performance of our methodology using synthetic data and compare it with\nalternative approaches. Our numerical experiments show that including the\nestimation of the covariance matrix in the Lasso criterion dramatically\nimproves the variable selection performance in many cases.\n",
"title": "Variable selection in multivariate linear models with high-dimensional covariance matrix estimation"
}
| null | null |
[
"Mathematics",
"Statistics"
] | null | true | null |
3222
| null |
Validated
| null | null |
null |
{
"abstract": " Background. Models of cancer-induced neuropathy are designed by injecting\ncancer cells near the peripheral nerves. The interference of tissue-resident\nimmune cells does not allow a direct contact with nerve fibres which affects\nthe tumor microenvironment and the invasion process. Methods. Anaplastic\ntumor-1 (AT-1) cells were inoculated within the sciatic nerves (SNs) of male\nCopenhagen rats. Lumbar dorsal root ganglia (DRGs) and the SNs were collected\non days 3, 7, 14, and 21. SN tissues were examined for morphological changes\nand DRG tissues for immunofluorescence, electrophoretic tendency, and mRNA\nquantification. Hypersensitivities to cold, mechanical, and thermal stimuli\nwere determined. HC-030031, a selective TRPA1 antagonist, was used to treat\ncold allodynia. Results. Nociception thresholds were identified on day 6.\nImmunofluorescent micrographs showed overexpression of TRPA1 on days 7 and 14\nand of CGRP on day 14 until day 21. Both TRPA1 and CGRP were coexpressed on the\nsame cells. Immunoblots exhibited an increase in TRPA1 expression on day 14.\nTRPA1 mRNA underwent an increase on day 7 (normalized to 18S). Injection of\nHC-030031 transiently reversed the cold allodynia. Conclusion. A novel and a\npromising model of cancer-induced neuropathy was established, and the role of\nTRPA1 and CGRP in pain transduction was examined.\n",
"title": "A Novel Model of Cancer-Induced Peripheral Neuropathy and the Role of TRPA1 in Pain Transduction"
}
| null | null | null | null | true | null |
3223
| null |
Default
| null | null |
null |
{
"abstract": " We present a novel method for frequentist statistical inference in\n$M$-estimation problems, based on stochastic gradient descent (SGD) with a\nfixed step size: we demonstrate that the average of such SGD sequences can be\nused for statistical inference, after proper scaling. An intuitive analysis\nusing the Ornstein-Uhlenbeck process suggests that such averages are\nasymptotically normal. From a practical perspective, our SGD-based inference\nprocedure is a first order method, and is well-suited for large scale problems.\nTo show its merits, we apply it to both synthetic and real datasets, and\ndemonstrate that its accuracy is comparable to classical statistical methods,\nwhile requiring potentially far less computation.\n",
"title": "Statistical inference using SGD"
}
| null | null | null | null | true | null |
3224
| null |
Default
| null | null |
null |
{
"abstract": " Massive network exploration is an important research direction with many\napplications. In such a setting, the network is, usually, modeled as a graph\n$G$, whereas any structural information of interest is extracted by inspecting\nthe way nodes are connected together. In the case where the adjacency matrix or\nthe adjacency list of $G$ is available, one can directly apply graph mining\nalgorithms to extract useful knowledge. However, there are cases where this is\nnot possible because the graph is \\textit{hidden} or \\textit{implicit}, meaning\nthat the edges are not recorded explicitly in the form of an adjacency\nrepresentation. In such a case, the only alternative is to pose a sequence of\n\\textit{edge probing queries} asking for the existence or not of a particular\ngraph edge. However, checking all possible node pairs is costly (quadratic on\nthe number of nodes). Thus, our objective is to pose as few edge probing\nqueries as possible, since each such query is expected to be costly. In this\nwork, we center our focus on the \\textit{core decomposition} of a hidden graph.\nIn particular, we provide an efficient algorithm to detect the maximal subgraph\nof $S_k$ of $G$ where the induced degree of every node $u \\in S_k$ is at least\n$k$. Performance evaluation results demonstrate that significant performance\nimprovements are achieved in comparison to baseline approaches.\n",
"title": "Core Discovery in Hidden Graphs"
}
| null | null | null | null | true | null |
3225
| null |
Default
| null | null |
null |
{
"abstract": " We consider the Bradlow equation for vortices which was recently found by\nManton and find a two-parameter class of analytic solutions in closed form on\nnontrivial geometries with non-constant curvature. The general solution to our\nclass of metrics is given by a hypergeometric function and the area of the\nvortex domain by the Gaussian hypergeometric function.\n",
"title": "Some exact Bradlow vortex solutions"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3226
| null |
Validated
| null | null |
null |
{
"abstract": " Let $\\Pi_q$ be an arbitrary finite projective plane of order $q$. A subset\n$S$ of its points is called saturating if any point outside $S$ is collinear\nwith a pair of points from $S$. Applying probabilistic tools we improve the\nupper bound on the smallest possible size of the saturating set to\n$\\lceil\\sqrt{3q\\ln{q}}\\rceil+ \\lceil(\\sqrt{q}+1)/2\\rceil$. The same result is\npresented using an algorithmic approach as well, which points out the\nconnection with the transversal number of uniform multiple intersecting\nhypergraphs.\n",
"title": "Saturating sets in projective planes and hypergraph covers"
}
| null | null | null | null | true | null |
3227
| null |
Default
| null | null |
null |
{
"abstract": " Following the seminal work of Nesterov, accelerated optimization methods have\nbeen used to powerfully boost the performance of first-order, gradient-based\nparameter estimation in scenarios where second-order optimization strategies\nare either inapplicable or impractical. Not only does accelerated gradient\ndescent converge considerably faster than traditional gradient descent, but it\nalso performs a more robust local search of the parameter space by initially\novershooting and then oscillating back as it settles into a final\nconfiguration, thereby selecting only local minimizers with a basis of\nattraction large enough to contain the initial overshoot. This behavior has\nmade accelerated and stochastic gradient search methods particularly popular\nwithin the machine learning community. In their recent PNAS 2016 paper,\nWibisono, Wilson, and Jordan demonstrate how a broad class of accelerated\nschemes can be cast in a variational framework formulated around the Bregman\ndivergence, leading to continuum limit ODE's. We show how their formulation may\nbe further extended to infinite dimension manifolds (starting here with the\ngeometric space of curves and surfaces) by substituting the Bregman divergence\nwith inner products on the tangent space and explicitly introducing a\ndistributed mass model which evolves in conjunction with the object of interest\nduring the optimization process. The co-evolving mass model, which is\nintroduced purely for the sake of endowing the optimization with helpful\ndynamics, also links the resulting class of accelerated PDE based optimization\nschemes to fluid dynamical formulations of optimal mass transport.\n",
"title": "Accelerated Optimization in the PDE Framework: Formulations for the Active Contour Case"
}
| null | null | null | null | true | null |
3228
| null |
Default
| null | null |
null |
{
"abstract": " We study the conditions under which one is able to efficiently apply\nvariance-reduction and acceleration schemes on finite sum optimization\nproblems. First, we show that, perhaps surprisingly, the finite sum structure\nby itself, is not sufficient for obtaining a complexity bound of\n$\\tilde{\\cO}((n+L/\\mu)\\ln(1/\\epsilon))$ for $L$-smooth and $\\mu$-strongly\nconvex individual functions - one must also know which individual function is\nbeing referred to by the oracle at each iteration. Next, we show that for a\nbroad class of first-order and coordinate-descent finite sum algorithms\n(including, e.g., SDCA, SVRG, SAG), it is not possible to get an `accelerated'\ncomplexity bound of $\\tilde{\\cO}((n+\\sqrt{n L/\\mu})\\ln(1/\\epsilon))$, unless\nthe strong convexity parameter is given explicitly. Lastly, we show that when\nthis class of algorithms is used for minimizing $L$-smooth and convex finite\nsums, the optimal complexity bound is $\\tilde{\\cO}(n+L/\\epsilon)$, assuming\nthat (on average) the same update rule is used in every iteration, and\n$\\tilde{\\cO}(n+\\sqrt{nL/\\epsilon})$, otherwise.\n",
"title": "Limitations on Variance-Reduction and Acceleration Schemes for Finite Sum Optimization"
}
| null | null | null | null | true | null |
3229
| null |
Default
| null | null |
null |
{
"abstract": " Stability is an important aspect of a classification procedure because\nunstable predictions can potentially reduce users' trust in a classification\nsystem and also harm the reproducibility of scientific conclusions. The major\ngoal of our work is to introduce a novel concept of classification instability,\ni.e., decision boundary instability (DBI), and incorporate it with the\ngeneralization error (GE) as a standard for selecting the most accurate and\nstable classifier. Specifically, we implement a two-stage algorithm: (i)\ninitially select a subset of classifiers whose estimated GEs are not\nsignificantly different from the minimal estimated GE among all the candidate\nclassifiers; (ii) the optimal classifier is chosen as the one achieving the\nminimal DBI among the subset selected in stage (i). This general selection\nprinciple applies to both linear and nonlinear classifiers. Large-margin\nclassifiers are used as a prototypical example to illustrate the above idea.\nOur selection method is shown to be consistent in the sense that the optimal\nclassifier simultaneously achieves the minimal GE and the minimal DBI. Various\nsimulations and real examples further demonstrate the advantage of our method\nover several alternative approaches.\n",
"title": "Stability Enhanced Large-Margin Classifier Selection"
}
| null | null |
[
"Statistics"
] | null | true | null |
3230
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper we analyze the capacitary potential due to a charged body in\norder to deduce sharp analytic and geometric inequalities, whose equality cases\nare saturated by domains with spherical symmetry. In particular, for a regular\nbounded domain $\\Omega \\subset \\mathbb{R}^n$, $n\\geq 3$, we prove that if the\nmean curvature $H$ of the boundary obeys the condition $$ - \\bigg[\n\\frac{1}{\\text{Cap}(\\Omega)} \\bigg]^{\\frac{1}{n-2}} \\leq \\frac{H}{n-1} \\leq\n\\bigg[ \\frac{1}{\\text{Cap}(\\Omega)} \\bigg]^{\\frac{1}{n-2}} , $$ then $\\Omega$\nis a round ball.\n",
"title": "Some Sphere Theorems in Linear Potential Theory"
}
| null | null | null | null | true | null |
3231
| null |
Default
| null | null |
null |
{
"abstract": " Non-Gaussian component analysis (NGCA) is a problem in multidimensional data\nanalysis which, since its formulation in 2006, has attracted considerable\nattention in statistics and machine learning. In this problem, we have a random\nvariable $X$ in $n$-dimensional Euclidean space. There is an unknown subspace\n$\\Gamma$ of the $n$-dimensional Euclidean space such that the orthogonal\nprojection of $X$ onto $\\Gamma$ is standard multidimensional Gaussian and the\northogonal projection of $X$ onto $\\Gamma^{\\perp}$, the orthogonal complement\nof $\\Gamma$, is non-Gaussian, in the sense that all its one-dimensional\nmarginals are different from the Gaussian in a certain metric defined in terms\nof moments. The NGCA problem is to approximate the non-Gaussian subspace\n$\\Gamma^{\\perp}$ given samples of $X$.\nVectors in $\\Gamma^{\\perp}$ correspond to `interesting' directions, whereas\nvectors in $\\Gamma$ correspond to the directions where data is very noisy. The\nmost interesting applications of the NGCA model is for the case when the\nmagnitude of the noise is comparable to that of the true signal, a setting in\nwhich traditional noise reduction techniques such as PCA don't apply directly.\nNGCA is also related to dimension reduction and to other data analysis problems\nsuch as ICA. NGCA-like problems have been studied in statistics for a long time\nusing techniques such as projection pursuit.\nWe give an algorithm that takes polynomial time in the dimension $n$ and has\nan inverse polynomial dependence on the error parameter measuring the angle\ndistance between the non-Gaussian subspace and the subspace output by the\nalgorithm. Our algorithm is based on relative entropy as the contrast function\nand fits under the projection pursuit framework. The techniques we develop for\nanalyzing our algorithm maybe of use for other related problems.\n",
"title": "Non-Gaussian Component Analysis using Entropy Methods"
}
| null | null | null | null | true | null |
3232
| null |
Default
| null | null |
null |
{
"abstract": " Recent years have seen a surprising connection between the physics of\nscattering amplitudes and a class of mathematical objects--the positive\nGrassmannian, positive loop Grassmannians, tree and loop Amplituhedra--which\nhave been loosely referred to as \"positive geometries\". The connection between\nthe geometry and physics is provided by a unique differential form canonically\ndetermined by the property of having logarithmic singularities (only) on all\nthe boundaries of the space, with residues on each boundary given by the\ncanonical form on that boundary. In this paper we initiate an exploration of\n\"positive geometries\" and \"canonical forms\" as objects of study in their own\nright in a more general mathematical setting. We give a precise definition of\npositive geometries and canonical forms, introduce general methods for finding\nforms for more complicated positive geometries from simpler ones, and present\nnumerous examples of positive geometries in projective spaces, Grassmannians,\nand toric, cluster and flag varieties. We also illustrate a number of\nstrategies for computing canonical forms which yield interesting\nrepresentations for the forms associated with wide classes of positive\ngeometries, ranging from the simplest Amplituhedra to new expressions for the\nvolume of arbitrary convex polytopes.\n",
"title": "Positive Geometries and Canonical Forms"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3233
| null |
Validated
| null | null |
null |
{
"abstract": " Computational Thinking (CT) has been described as an essential skill which\neveryone should learn and can therefore include in their skill set. Seymour\nPapert is credited as concretising Computational Thinking in 1980 but since\nWing popularised the term in 2006 and brought it to the international\ncommunity's attention, more and more research has been conducted on CT in\neducation. The aim of this systematic literary review is to give educators and\neducation researchers an overview of what work has been carried out in the\ndomain, as well as potential gaps and opportunities that still exist.\nOverall it was found in this review that, although there is a lot of work\ncurrently being done around the world in many different educational contexts,\nthe work relating to CT is still in its infancy. Along with the need to create\nan agreed-upon definition of CT lots of countries are still in the process of,\nor have not yet started, introducing CT into curriculums in all levels of\neducation. It was also found that Computer Science/Computing, which could be\nthe most obvious place to teach CT, has yet to become a mainstream subject in\nsome countries, although this is improving. Of encouragement to educators is\nthe wealth of tools and resources being developed to help teach CT as well as\nmore and more work relating to curriculum development. For those teachers\nlooking to incorporate CT into their schools or classes then there are\nbountiful options which include programming, hands-on exercises and more. The\nneed for more detailed lesson plans and curriculum structure however, is\nsomething that could be of benefit to teachers.\n",
"title": "Computational Thinking in Education: Where does it Fit? A systematic literary review"
}
| null | null | null | null | true | null |
3234
| null |
Default
| null | null |
null |
{
"abstract": " Spincaloritronic signal generation due to thermal spin injection and spin\ntransport is demonstrated in a non-degenerate Si spin valve. The spin-dependent\nSeebeck effect is used for the spincaloritronic signal generation, and the\nthermal gradient of about 200 mK at an interface of Fe and Si enables\ngenerating a spin voltage of 8 {\\mu}V at room temperature. A simple expansion\nof a conventional spin drift-diffusion model with taking into account the\nspin-dependent Seebeck effect shows semiconductor materials are quite potential\nfor the spincaloritronic signal generation comparing with metallic materials,\nwhich can allow efficient heat recycling in semiconductor spin devices.\n",
"title": "Spincaloritronic signal generation in non-degenerate Si"
}
| null | null | null | null | true | null |
3235
| null |
Default
| null | null |
null |
{
"abstract": " We quantify uncertainties in the location and magnitude of extreme pressure\nspots revealed from large scale multi-phase flow simulations of cloud\ncavitation collapse. We examine clouds containing 500 cavities and quantify\nuncertainties related to their initial spatial arrangement. The resulting\n2000-dimensional space is sampled using a non-intrusive and computationally\nefficient Multi-Level Monte Carlo (MLMC) methodology. We introduce novel\noptimal control variate coefficients to enhance the variance reduction in MLMC.\nThe proposed optimal fidelity MLMC leads to more than two orders of magnitude\nspeedup when compared to standard Monte Carlo methods. We identify large\nuncertainties in the location and magnitude of the peak pressure pulse and\npresent its statistical correlations and joint probability density functions\nwith the geometrical characteristics of the cloud. Characteristic properties of\nspatial cloud structure are identified as potential causes of significant\nuncertainties in exerted collapse pressures.\n",
"title": "Optimal fidelity multi-level Monte Carlo for quantification of uncertainty in simulations of cloud cavitation collapse"
}
| null | null | null | null | true | null |
3236
| null |
Default
| null | null |
null |
{
"abstract": " Pollution in urban centres is becoming a major societal problem. While\npollution is a concern for all urban dwellers, cyclists are one of the most\nexposed groups due to their proximity to vehicle tailpipes. Consequently, new\nsolutions are required to help protect citizens, especially cyclists, from the\nharmful effects of exhaust-gas emissions. In this context, hybrid vehicles\n(HVs) offer new actuation possibilities that can be exploited in this\ndirection. More specifically, such vehicles when working together as a group,\nhave the ability to dynamically lower the emissions in a given area, thus\nbenefiting citizens, whilst still giving the vehicle owner the flexibility of\nusing an Internal Combustion Engine (ICE). This paper aims to develop an\nalgorithm, that can be deployed in such vehicles, whereby geofences (virtual\ngeographic boundaries) are used to specify areas of low pollution around\ncyclists. The emissions level inside the geofence is controlled via a coin\ntossing algorithm to switch the HV motor into, and out of, electric mode, in a\nmanner that is in some sense optimal. The optimality criterion is based on how\npolluting vehicles inside the geofence are, and the expected density of\ncyclists near each vehicle. The algorithm is triggered once a vehicle detects a\ncyclist. Implementations are presented, both in simulation, and in a real\nvehicle, and the system is tested using a Hardware-In-the-Loop (HIL) platform\n(video provided).\n",
"title": "A New Take on Protecting Cyclists in Smart Cities"
}
| null | null | null | null | true | null |
3237
| null |
Default
| null | null |
null |
{
"abstract": " During the accretion phase of a core-collapse supernovae, large amplitude\nturbulence is generated by the combination of the standing accretion shock\ninstability and convection driven by neutrino heating. The turbulence directly\naffects the dynamics of the explosion, but there is also the possibility of an\nadditional, indirect, feedback mechanism due to the effect turbulence can have\nupon neutrino flavor evolution and thus the neutrino heating. In this paper we\nconsider the effect of turbulence during the accretion phase upon neutrino\nevolution, both numerically and analytically. Adopting representative supernova\nprofiles taken from the accretion phase of a supernova simulation, we find the\nnumerical calculations exhibit no effect from turbulence. We explain this\nabsence using two analytic descriptions: the Stimulated Transition model and\nthe Distorted Phase Effect model. In the Stimulated Transition model turbulence\neffects depend upon six different lengthscales, and three criteria must be\nsatisfied between them if one is to observe a change in the flavor evolution\ndue to Stimulated Transition. We further demonstrate that the Distorted Phase\nEffect depends upon the presence of multiple semi-adiabatic MSW resonances or\ndiscontinuities that also can be expressed as a relationship between three of\nthe same lengthscales. When we examine the supernova profiles used in the\nnumerical calculations we find the three Stimulated Transition criteria cannot\nbe satisfied, independent of the form of the turbulence power spectrum, and\nthat the same supernova profiles lack the multiple semi-adiabatic MSW\nresonances or discontinuities necessary to produce a Distorted Phase Effect.\nThus we conclude that even though large amplitude turbulence is present in\nsupernova during the accretion phase, it has no effect upon neutrino flavor\nevolution.\n",
"title": "The effect upon neutrinos of core-collapse supernova accretion phase turbulence"
}
| null | null | null | null | true | null |
3238
| null |
Default
| null | null |
null |
{
"abstract": " From only positive (P) and unlabeled (U) data, a binary classifier could be\ntrained with PU learning, in which the state of the art is unbiased PU\nlearning. However, if its model is very flexible, empirical risks on training\ndata will go negative, and we will suffer from serious overfitting. In this\npaper, we propose a non-negative risk estimator for PU learning: when getting\nminimized, it is more robust against overfitting, and thus we are able to use\nvery flexible models (such as deep neural networks) given limited P data.\nMoreover, we analyze the bias, consistency, and mean-squared-error reduction of\nthe proposed risk estimator, and bound the estimation error of the resulting\nempirical risk minimizer. Experiments demonstrate that our risk estimator fixes\nthe overfitting problem of its unbiased counterparts.\n",
"title": "Positive-Unlabeled Learning with Non-Negative Risk Estimator"
}
| null | null | null | null | true | null |
3239
| null |
Default
| null | null |
null |
{
"abstract": " The possibly unbiased selection process in surveys of the Sunyaev Zel'dovich\neffect can unveil new populations of galaxy clusters. We performed a weak\nlensing analysis of the PSZ2LenS sample, i.e. the PSZ2 galaxy clusters detected\nby the Planck mission in the sky portion covered by the lensing surveys\nCFHTLenS and RCSLenS. PSZ2LenS consists of 35 clusters and it is a\nstatistically complete and homogeneous subsample of the PSZ2 catalogue. The\nPlanck selected clusters appear to be unbiased tracers of the massive end of\nthe cosmological haloes. The mass concentration relation of the sample is in\nexcellent agreement with predictions from the Lambda cold dark matter model.\nThe stacked lensing signal is detected at 14 sigma significance over the radial\nrange 0.1<R<3.2 Mpc/h, and is well described by the cuspy dark halo models\npredicted by numerical simulations. We confirmed that Planck estimated masses\nare biased low by b_SZ= 27+-11(stat)+-8(sys) per cent with respect to weak\nlensing masses. The bias is higher for the cosmological subsample, b_SZ=\n40+-14+-(stat)+-8(sys) per cent.\n",
"title": "PSZ2LenS. Weak lensing analysis of the Planck clusters in the CFHTLenS and in the RCSLenS"
}
| null | null | null | null | true | null |
3240
| null |
Default
| null | null |
null |
{
"abstract": " A fundamental goal in network neuroscience is to understand how activity in\none region drives activity elsewhere, a process referred to as effective\nconnectivity. Here we propose to model this causal interaction using\nintegro-differential equations and causal kernels that allow for a rich\nanalysis of effective connectivity. The approach combines the tractability and\nflexibility of autoregressive modeling with the biophysical interpretability of\ndynamic causal modeling. The causal kernels are learned nonparametrically using\nGaussian process regression, yielding an efficient framework for causal\ninference. We construct a novel class of causal covariance functions that\nenforce the desired properties of the causal kernels, an approach which we call\nGP CaKe. By construction, the model and its hyperparameters have biophysical\nmeaning and are therefore easily interpretable. We demonstrate the efficacy of\nGP CaKe on a number of simulations and give an example of a realistic\napplication on magnetoencephalography (MEG) data.\n",
"title": "GP CaKe: Effective brain connectivity with causal kernels"
}
| null | null | null | null | true | null |
3241
| null |
Default
| null | null |
null |
{
"abstract": " A boundary behavior of ring mappings on Riemannian manifolds, which are\ngeneralization of quasiconformal mappings by Gehring, is investigated. In terms\nof prime ends, there are obtained theorems about continuous extension to a\nboundary of classes mentioned above. In the terms mentioned above, there are\nobtained results about equicontinuity of these classes in the closure of the\ndomain.\n",
"title": "On boundary behavior of mappings on Riemannian manifolds in terms of prime ends"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3242
| null |
Validated
| null | null |
null |
{
"abstract": " Recently, cloud storage and processing have been widely adopted. Mobile users\nin one family or one team may automatically backup their photos to the same\nshared cloud storage space. The powerful face detector trained and provided by\na 3rd party may be used to retrieve the photo collection which contains a\nspecific group of persons from the cloud storage server. However, the privacy\nof the mobile users may be leaked to the cloud server providers. In the\nmeanwhile, the copyright of the face detector should be protected. Thus, in\nthis paper, we propose a protocol of privacy preserving face retrieval in the\ncloud for mobile users, which protects the user photos and the face detector\nsimultaneously. The cloud server only provides the resources of storage and\ncomputing and can not learn anything of the user photos and the face detector.\nWe test our protocol inside several families and classes. The experimental\nresults reveal that our protocol can successfully retrieve the proper photos\nfrom the cloud server and protect the user photos and the face detector.\n",
"title": "Privacy Preserving Face Retrieval in the Cloud for Mobile Users"
}
| null | null | null | null | true | null |
3243
| null |
Default
| null | null |
null |
{
"abstract": " This work provides a simplified proof of the statistical minimax optimality\nof (iterate averaged) stochastic gradient descent (SGD), for the special case\nof least squares. This result is obtained by analyzing SGD as a stochastic\nprocess and by sharply characterizing the stationary covariance matrix of this\nprocess. The finite rate optimality characterization captures the constant\nfactors and addresses model mis-specification.\n",
"title": "A Markov Chain Theory Approach to Characterizing the Minimax Optimality of Stochastic Gradient Descent (for Least Squares)"
}
| null | null | null | null | true | null |
3244
| null |
Default
| null | null |
null |
{
"abstract": " It is becoming increasingly common to see large collections of network data\nobjects -- that is, data sets in which a network is viewed as a fundamental\nunit of observation. As a result, there is a pressing need to develop\nnetwork-based analogues of even many of the most basic tools already standard\nfor scalar and vector data. In this paper, our focus is on averages of\nunlabeled, undirected networks with edge weights. Specifically, we (i)\ncharacterize a certain notion of the space of all such networks, (ii) describe\nkey topological and geometric properties of this space relevant to doing\nprobability and statistics thereupon, and (iii) use these properties to\nestablish the asymptotic behavior of a generalized notion of an empirical mean\nunder sampling from a distribution supported on this space. Our results rely on\na combination of tools from geometry, probability theory, and statistical shape\nanalysis. In particular, the lack of vertex labeling necessitates working with\na quotient space modding out permutations of labels. This results in a\nnontrivial geometry for the space of unlabeled networks, which in turn is found\nto have important implications on the types of probabilistic and statistical\nresults that may be obtained and the techniques needed to obtain them.\n",
"title": "Averages of Unlabeled Networks: Geometric Characterization and Asymptotic Behavior"
}
| null | null | null | null | true | null |
3245
| null |
Default
| null | null |
null |
{
"abstract": " We define the Abelian distribution and study its basic properties. Abelian\ndistributions arise in the context of neural modeling and describe the size of\nneural avalanches in fully-connected integrate-and-fire models of\nself-organized criticality in neural systems.\n",
"title": "The Abelian distribution"
}
| null | null |
[
"Physics"
] | null | true | null |
3246
| null |
Validated
| null | null |
null |
{
"abstract": " We introduce PULSEDYN, a particle dynamics program in $C++$, to solve\nmany-body nonlinear systems in one dimension. PULSEDYN is designed to make\ncomputing accessible to non-specialists in the field of nonlinear dynamics of\nmany-body systems and to ensure transparency and easy benchmarking of numerical\nresults for an integrable model (Toda chain) and three non-integrable models\n(Fermi-Pasta-Ulam-Tsingou, Morse and Lennard-Jones). To achieve the latter, we\nhave made our code open source and free to distribute. We examine (i) soliton\npropagation and two-soliton collision in the Toda system, (ii) the recurrence\nphenomenon in the Fermi-Pasta-Ulam-Tsingou system and the decay of a single\nlocalized nonlinear excitation in the same system through quasi-equilibrium to\nan equipartitioned state, and SW propagation in chains with (iii) Morse and\n(iv) Lennard-Jones potentials. We recover well known results from theory and\nother numerical results in the literature. We have obtained these results by\nsetting up a parameter file interface which allows the code to be used as a\nblack box. Therefore, we anticipate that the code would prove useful to\nstudents and non-specialists. At the same time, PULSEDYN provides\nscientifically accurate simulations thus making the study of rich dynamical\nprocesses broadly accessible.\n",
"title": "PULSEDYN - A dynamical simulation tool for studying strongly nonlinear chains"
}
| null | null | null | null | true | null |
3247
| null |
Default
| null | null |
null |
{
"abstract": " This paper presents the design of a control model to navigate the\ndifferential mobile robot to reach the desired destination from an arbitrary\ninitial pose. The designed model is divided into two stages: the state\nestimation and the stabilization control. In the state estimation, an extended\nKalman filter is employed to optimally combine the information from the system\ndynamics and measurements. Two Lyapunov functions are constructed that allow a\nhybrid feedback control law to execute the robot movements. The asymptotical\nstability and robustness of the closed loop system are assured. Simulations and\nexperiments are carried out to validate the effectiveness and applicability of\nthe proposed approach.\n",
"title": "Stabilization Control of the Differential Mobile Robot Using Lyapunov Function and Extended Kalman Filter"
}
| null | null |
[
"Computer Science"
] | null | true | null |
3248
| null |
Validated
| null | null |
null |
{
"abstract": " Big data sets must be carefully partitioned into statistically similar data\nsubsets that can be used as representative samples for big data analysis tasks.\nIn this paper, we propose the random sample partition (RSP) data model to\nrepresent a big data set as a set of non-overlapping data subsets, called RSP\ndata blocks, where each RSP data block has a probability distribution similar\nto the whole big data set. Under this data model, efficient block level\nsampling is used to randomly select RSP data blocks, replacing expensive record\nlevel sampling to select sample data from a big distributed data set on a\ncomputing cluster. We show how RSP data blocks can be employed to estimate\nstatistics of a big data set and build models which are equivalent to those\nbuilt from the whole big data set. In this approach, analysis of a big data set\nbecomes analysis of few RSP data blocks which have been generated in advance on\nthe computing cluster. Therefore, the new method for data analysis based on RSP\ndata blocks is scalable to big data.\n",
"title": "A Random Sample Partition Data Model for Big Data Analysis"
}
| null | null | null | null | true | null |
3249
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we present a family of bivariate copulas by transforming a\ngiven copula function with two increasing functions, named as transformed\ncopula. One distinctive characteristic of the transformed copula is its\nsingular component along the main diagonal. Conditions guaranteeing the\ntransformed function to be a copula function are provided, and several classes\nof the transformed copulas are given. The singular component along the main\ndiagonal of the transformed copula is verified, and the tail dependence\ncoefficients of the transformed copulas are obtained. Finally, some properties\nof the transformed copula are discussed, such as the totally positive of order\n2 and the concordance order.\n",
"title": "A family of transformed copulas with singular component"
}
| null | null | null | null | true | null |
3250
| null |
Default
| null | null |
null |
{
"abstract": " Machine learning can extract information from neural recordings, e.g.,\nsurface EEG, ECoG and {\\mu}ECoG, and therefore plays an important role in many\nresearch and clinical applications. Deep learning with artificial neural\nnetworks has recently seen increasing attention as a new approach in brain\nsignal decoding. Here, we apply a deep learning approach using convolutional\nneural networks to {\\mu}ECoG data obtained with a wireless, chronically\nimplanted system in an ovine animal model. Regularized linear discriminant\nanalysis (rLDA), a filter bank component spatial pattern (FBCSP) algorithm and\nconvolutional neural networks (ConvNets) were applied to auditory evoked\nresponses captured by {\\mu}ECoG. We show that compared with rLDA and FBCSP,\nsignificantly higher decoding accuracy can be obtained by ConvNets trained in\nan end-to-end manner, i.e., without any predefined signal features. Deep\nlearning thus proves a promising technique for {\\mu}ECoG-based brain-machine\ninterfacing applications.\n",
"title": "Deep Learning for micro-Electrocorticographic (μECoG) Data"
}
| null | null | null | null | true | null |
3251
| null |
Default
| null | null |
null |
{
"abstract": " We consider the problem of recovering a low-rank matrix from its clipped\nobservations. Clipping is conceivable in many scientific areas that obstructs\nstatistical analyses. On the other hand, matrix completion (MC) methods can\nrecover a low-rank matrix from various information deficits by using the\nprinciple of low-rank completion. However, the current theoretical guarantees\nfor low-rank MC do not apply to clipped matrices, as the deficit depends on the\nunderlying values. Therefore, the feasibility of clipped matrix completion\n(CMC) is not trivial. In this paper, we first provide a theoretical guarantee\nfor the exact recovery of CMC by using a trace-norm minimization algorithm.\nFurthermore, we propose practical CMC algorithms by extending ordinary MC\nmethods. Our extension is to use the squared hinge loss in place of the squared\nloss for reducing the penalty of over-estimation on clipped entries. We also\npropose a novel regularization term tailored for CMC. It is a combination of\ntwo trace-norm terms, and we theoretically bound the recovery error under the\nregularization. We demonstrate the effectiveness of the proposed methods\nthrough experiments using both synthetic and benchmark data for recommendation\nsystems.\n",
"title": "Clipped Matrix Completion: A Remedy for Ceiling Effects"
}
| null | null | null | null | true | null |
3252
| null |
Default
| null | null |
null |
{
"abstract": " Supervised learning based on a deep neural network recently has achieved\nsubstantial improvement on speech enhancement. Denoising networks learn mapping\nfrom noisy speech to clean one directly, or to a spectra mask which is the\nratio between clean and noisy spectrum. In either case, the network is\noptimized by minimizing mean square error (MSE) between predefined labels and\nnetwork output of spectra or time-domain signal. However, existing schemes have\neither of two critical issues: spectra and metric mismatches. The spectra\nmismatch is a well known issue that any spectra modification after short-time\nFourier transform (STFT), in general, cannot be fully recovered after inverse\nSTFT. The metric mismatch is that a conventional MSE metric is sub-optimal to\nmaximize our target metrics, signal-to-distortion ratio (SDR) and perceptual\nevaluation of speech quality (PESQ). This paper presents a new end-to-end\ndenoising framework with the goal of joint SDR and PESQ optimization. First,\nthe network optimization is performed on the time-domain signals after ISTFT to\navoid spectra mismatch. Second, two loss functions which have improved\ncorrelations with SDR and PESQ metrics are proposed to minimize metric\nmismatch. The experimental result showed that the proposed denoising scheme\nsignificantly improved both SDR and PESQ performance over the existing methods.\n",
"title": "End-to-End Multi-Task Denoising for joint SDR and PESQ Optimization"
}
| null | null | null | null | true | null |
3253
| null |
Default
| null | null |
null |
{
"abstract": " We propose a Monte Carlo algorithm to sample from high-dimensional\nprobability distributions that combines Markov chain Monte Carlo (MCMC) and\nimportance sampling. We provide a careful theoretical analysis, including\nguarantees on robustness to high-dimensionality, explicit comparison with\nstandard MCMC and illustrations of the potential improvements in efficiency.\nSimple and concrete intuition is provided for when the novel scheme is expected\nto outperform standard schemes. When applied to Bayesian Variable Selection\nproblems, the novel algorithm is orders of magnitude more efficient than\navailable alternative sampling schemes and allows to perform fast and reliable\nfully Bayesian inferences with tens of thousands regressors.\n",
"title": "Scalable Importance Tempering and Bayesian Variable Selection"
}
| null | null | null | null | true | null |
3254
| null |
Default
| null | null |
null |
{
"abstract": " Convolutional dictionary learning (CDL) estimates shift invariant basis\nadapted to multidimensional data. CDL has proven useful for image denoising or\ninpainting, as well as for pattern discovery on multivariate signals. As\nestimated patterns can be positioned anywhere in signals or images,\noptimization techniques face the difficulty of working in extremely high\ndimensions with millions of pixels or time samples, contrarily to standard\npatch-based dictionary learning. To address this optimization problem, this\nwork proposes a distributed and asynchronous algorithm, employing locally\ngreedy coordinate descent and an asynchronous locking mechanism that does not\nrequire a central server. This algorithm can be used to distribute the\ncomputation on a number of workers which scales linearly with the encoded\nsignal's size. Experiments confirm the scaling properties which allows us to\nlearn patterns on large scales images from the Hubble Space Telescope.\n",
"title": "Distributed Convolutional Dictionary Learning (DiCoDiLe): Pattern Discovery in Large Images and Signals"
}
| null | null | null | null | true | null |
3255
| null |
Default
| null | null |
null |
{
"abstract": " This paper derives new formulations for designing dominant pole placement\nbased proportional-integral-derivative (PID) controllers to handle second order\nprocesses with time delays (SOPTD). Previously, similar attempts have been made\nfor pole placement in delay-free systems. The presence of the time delay term\nmanifests itself as a higher order system with variable number of interlaced\npoles and zeros upon Pade approximation, which makes it difficult to achieve\nprecise pole placement control. We here report the analytical expressions to\nconstrain the closed loop dominant and non-dominant poles at the desired\nlocations in the complex s-plane, using a third order Pade approximation for\nthe delay term. However, invariance of the closed loop performance with\ndifferent time delay approximation has also been verified using increasing\norder of Pade, representing a closed to reality higher order delay dynamics.\nThe choice of the nature of non-dominant poles e.g. all being complex, real or\na combination of them modifies the characteristic equation and influences the\nachievable stability regions. The effect of different types of non-dominant\npoles and the corresponding stability regions are obtained for nine test-bench\nprocesses indicating different levels of open-loop damping and lag to delay\nratio. Next, we investigate which expression yields a wider stability region in\nthe design parameter space by using Monte Carlo simulations while uniformly\nsampling a chosen design parameter space. Various time and frequency domain\ncontrol performance parameters are investigated next, as well as their\ndeviations with uncertain process parameters, using thousands of Monte Carlo\nsimulations, around the robust stable solution for each of the nine test-bench\nprocesses.\n",
"title": "Performance Analysis of Robust Stable PID Controllers Using Dominant Pole Placement for SOPTD Process Models"
}
| null | null |
[
"Statistics"
] | null | true | null |
3256
| null |
Validated
| null | null |
null |
{
"abstract": " In this note we present an $\\infty$-categorical framework for descent along\nadjunctions and a general formula for counting conjugates up to equivalence\nwhich unifies several known formulae from different fields.\n",
"title": "On Conjugates and Adjoint Descent"
}
| null | null | null | null | true | null |
3257
| null |
Default
| null | null |
null |
{
"abstract": " There has been relatively little attention to incorporating linguistic prior\nto neural machine translation. Much of the previous work was further\nconstrained to considering linguistic prior on the source side. In this paper,\nwe propose a hybrid model, called NMT+RNNG, that learns to parse and translate\nby combining the recurrent neural network grammar into the attention-based\nneural machine translation. Our approach encourages the neural machine\ntranslation model to incorporate linguistic prior during training, and lets it\ntranslate on its own afterward. Extensive experiments with four language pairs\nshow the effectiveness of the proposed NMT+RNNG.\n",
"title": "Learning to Parse and Translate Improves Neural Machine Translation"
}
| null | null | null | null | true | null |
3258
| null |
Default
| null | null |
null |
{
"abstract": " This paper addresses the boundary stabilization of a flexible wing model,\nboth in bending and twisting displacements, under unsteady aerodynamic loads,\nand in presence of a store. The wing dynamics is captured by a distributed\nparameter system as a coupled Euler-Bernoulli and Timoshenko beam model. The\nproblem is tackled in the framework of semigroup theory, and a Lyapunov-based\nstability analysis is carried out to assess that the system energy, as well as\nthe bending and twisting displacements, decay exponentially to zero. The\neffectiveness of the proposed boundary control scheme is evaluated based on\nsimulations.\n",
"title": "Boundary feedback stabilization of a flexible wing model under unsteady aerodynamic loads"
}
| null | null | null | null | true | null |
3259
| null |
Default
| null | null |
null |
{
"abstract": " We study vertex colourings of digraphs so that no out-neighbourhood is\nmonochromatic and call such a colouring an {\\bf out-colouring}. The problem of\ndeciding whether a given digraph has an out-colouring with only two colours\n(called a 2-out-colouring) is ${\\cal\nNP}$-complete. We show that for every choice of positive integers $r,k$ there\nexists a $k$-strong bipartite tournament which needs at least $r$ colours in\nevery out-colouring. Our main results are on tournaments and semicomplete\ndigraphs. We prove that, except for the Paley tournament $P_7$, every strong\nsemicomplete digraph of minimum out-degree at least 3 has a 2-out-colouring.\nFurthermore, we show that every semicomplete digraph on at least 7 vertices has\na 2-out-colouring if and only if it has a {\\bf balanced} such colouring, that\nis, the difference between the number of vertices that receive colour 1 and\ncolour 2 is at most one. In the second half of the paper we consider the\ngeneralization of 2-out-colourings to vertex partitions $(V_1,V_2)$ of a\ndigraph $D$ so that each of the three digraphs induced by respectively, the\nvertices of $V_1$, the vertices of $V_2$ and all arcs between $V_1$ and $V_2$\nhave minimum out-degree $k$ for a prescribed integer $k\\geq 1$. Using\nprobabilistic arguments we prove that there exists an absolute positive\nconstant $c$ so that every semicomplete digraph of minimum out-degree at least\n$2k+c\\sqrt{k}$ has such a partition. This is tight up to the value of $c$.\n",
"title": "Out-colourings of Digraphs"
}
| null | null | null | null | true | null |
3260
| null |
Default
| null | null |
null |
{
"abstract": " Locality Sensitive Hashing (LSH) based algorithms have already shown their\npromise in finding approximate nearest neighbors in high dimen- sional data\nspace. However, there are certain scenarios, as in sequential data, where the\nproximity of a pair of points cannot be captured without considering their\nsurroundings or context. In videos, as for example, a particular frame is\nmeaningful only when it is seen in the context of its preceding and following\nframes. LSH has no mechanism to handle the con- texts of the data points. In\nthis article, a novel scheme of Context based Locality Sensitive Hashing\n(conLSH) has been introduced, in which points are hashed together not only\nbased on their closeness, but also because of similar context. The contribution\nmade in this article is three fold. First, conLSH is integrated with a recently\nproposed fast optimal sequence alignment algorithm (FOGSAA) using a layered\napproach. The resultant method is applied to video retrieval for extracting\nsimilar sequences. The pro- posed algorithm yields more than 80% accuracy on an\naverage in different datasets. It has been found to save 36.3% of the total\ntime, consumed by the exhaustive search. conLSH reduces the search space to\napproximately 42% of the entire dataset, when compared with an exhaustive\nsearch by the aforementioned FOGSAA, Bag of Words method and the standard LSH\nimplementations. Secondly, the effectiveness of conLSH is demon- strated in\naction recognition of the video clips, which yields an average gain of 12.83%\nin terms of classification accuracy over the state of the art methods using\nSTIP descriptors. The last but of great significance is that this article\nprovides a way of automatically annotating long and composite real life videos.\nThe source code of conLSH is made available at\nthis http URL\n",
"title": "An Improved Video Analysis using Context based Extension of LSH"
}
| null | null | null | null | true | null |
3261
| null |
Default
| null | null |
null |
{
"abstract": " For a group $H$ and a non empty subset $\\Gamma\\subseteq H$, the commuting\ngraph $G=\\mathcal{C}(H,\\Gamma)$ is the graph with $\\Gamma$ as the node set and\nwhere any $x,y \\in \\Gamma$ are joined by an edge if $x$ and $y$ commute in $H$.\nWe prove that any simple graph can be obtained as a commuting graph of a\nCoxeter group, solving the realizability problem in this setup. In particular\nwe can recover every Dynkin diagram of ADE type as a commuting graph. Thanks to\nthe relation between the ADE classification and finite subgroups of\n$\\SL(2,\\C)$, we are able to rephrase results from the {\\em McKay\ncorrespondence} in terms of generators of the corresponding Coxeter groups. We\nfinish the paper studying commuting graphs $\\mathcal{C}(H,\\Gamma)$ for every\nfinite subgroup $H\\subset\\SL(2,\\C)$ for different subsets $\\Gamma\\subseteq H$,\nand investigating metric properties of them when $\\Gamma=H$.\n",
"title": "Commuting graphs on Coxeter groups, Dynkin diagrams and finite subgroups of $SL(2,\\mathbb{C})$"
}
| null | null | null | null | true | null |
3262
| null |
Default
| null | null |
null |
{
"abstract": " Using the energy method we investigate the stability of pure conduction in\nPearson's model for Bénard-Marangoni convection in a layer of fluid at\ninfinite Prandtl number. Upon extending the space of admissible perturbations\nto the conductive state, we find an exact solution to the energy stability\nvariational problem for a range of thermal boundary conditions describing\nperfectly conducting, imperfectly conducting, and insulating boundaries. Our\nanalysis extends and improves previous results, and shows that with the energy\nmethod global stability can be proven up to the linear instability threshold\nonly when the top and bottom boundaries of the fluid layer are insulating.\nContrary to the well-known Rayleigh-Bénard convection setup, therefore,\nenergy stability theory does not exclude the possibility of subcritical\ninstabilities against finite-amplitude perturbations.\n",
"title": "Exact energy stability of Bénard-Marangoni convection at infinite Prandtl number"
}
| null | null |
[
"Physics"
] | null | true | null |
3263
| null |
Validated
| null | null |
null |
{
"abstract": " We present an elementary proof of a conjecture proposed by I. Rasa in 2017\nwhich is an inequality involving Bernstein basis polynomials and convex\nfunctions. It was affirmed in positive by A. Komisarski and T. Rajba very\nrecently by the use of stochastic convex orderings.\n",
"title": "A sharpening of a problem on Bernstein polynomials and convex functions"
}
| null | null | null | null | true | null |
3264
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we prove the following pointwise and curvature-free estimates\non convexity radius, injectivity radius and local behavior of geodesics in a\ncomplete Riemannian manifold $M$: 1) the convexity radius of $p$,\n$\\operatorname{conv}(p)\\ge\n\\min\\{\\frac{1}{2}\\operatorname{inj}(p),\\operatorname{foc}(B_{\\operatorname{inj}(p)}(p))\\}$,\nwhere $\\operatorname{inj}(p)$ is the injectivity radius of $p$ and\n$\\operatorname{foc}(B_r(p))$ is the focal radius of open ball centered at $p$\nwith radius $r$; 2) for any two points $p,q$ in $M$, $\\operatorname{inj}(q)\\ge\n\\min\\{\\operatorname{inj}(p), \\operatorname{conj}(q)\\}-d(p,q),$ where\n$\\operatorname{conj}(q)$ is the conjugate radius of $q$; 3) for any\n$0<r<\\min\\{\\operatorname{inj}(p),\\frac{1}{2}\\operatorname{conj}(B_{\\operatorname{inj}(p)}(p))\\}$,\nany (not necessarily minimizing) geodesic in $B_r(p)$ has length $\\le 2r$. We\nalso clarify two different concepts on convexity radius and give examples to\nillustrate that the one more frequently used in literature is not continuous.\n",
"title": "Local Estimate on Convexity Radius and decay of injectivity radius in a Riemannian manifold"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3265
| null |
Validated
| null | null |
null |
{
"abstract": " We report experimental results of the static magnetization, ESR and NMR\nspectroscopic measurements of the Ni-hybrid compound\nNiCl$_3$C$_6$H$_5$CH$_2$CH$_2$NH$_3$. In this material NiCl$_3$ octahedra are\nstructurally arranged in chains along the crystallographic $a$-axis. According\nto the static susceptibility and ESR data Ni$^{2+}$ spins $S = 1$ are isotropic\nand are coupled antiferromagnetically (AFM) along the chain with the exchange\nconstant $J = 25.5$ K. These are important prerequisites for the realization of\nthe so-called Haldane spin-1 chain with the spin-singlet ground state and a\nquantum spin gap. However, experimental results evidence AFM order at $T_{\\rm\nN} \\approx 10$ K presumably due to small interchain couplings. Interestingly,\nfrequency-, magnetic field-, and temperature-dependent ESR measurements, as\nwell as the NMR data, reveal signatures which could presumably indicate an\ninhomogeneous ground state of co-existent mesoscopically spatially separated\nAFM ordered and spin-singlet state regions similar to the situation observed\nbefore in some spin-diluted Haldane magnets.\n",
"title": "Magnetic properties of the spin-1 chain compound NiCl$_3$C$_6$H$_5$CH$_2$CH$_2$NH$_3$"
}
| null | null | null | null | true | null |
3266
| null |
Default
| null | null |
null |
{
"abstract": " We investigate the formation of multiple-core-hole states of molecular\nnitrogen interacting with a free-electron laser pulse. We obtain bound and\ncontinuum molecular orbitals in the single-center expansion scheme and use\nthese orbitals to calculate photo-ionization and Auger decay rates. Using these\nrates, we compute the atomic ion yields generated in this interaction. We track\nthe population of all states throughout this interaction and compute the\nproportion of the population which accesses different core-hole states. We also\ninvestigate the pulse parameters that favor the formation of these core-hole\nstates for 525 eV and 1100 eV photons.\n",
"title": "Multiple core hole formation by free-electron laser radiation in molecular nitrogen"
}
| null | null | null | null | true | null |
3267
| null |
Default
| null | null |
null |
{
"abstract": " Few ideas have enjoyed as large an impact on deep learning as convolution.\nFor any problem involving pixels or spatial representations, common intuition\nholds that convolutional neural networks may be appropriate. In this paper we\nshow a striking counterexample to this intuition via the seemingly trivial\ncoordinate transform problem, which simply requires learning a mapping between\ncoordinates in (x,y) Cartesian space and one-hot pixel space. Although\nconvolutional networks would seem appropriate for this task, we show that they\nfail spectacularly. We demonstrate and carefully analyze the failure first on a\ntoy problem, at which point a simple fix becomes obvious. We call this solution\nCoordConv, which works by giving convolution access to its own input\ncoordinates through the use of extra coordinate channels. Without sacrificing\nthe computational and parametric efficiency of ordinary convolution, CoordConv\nallows networks to learn either complete translation invariance or varying\ndegrees of translation dependence, as required by the end task. CoordConv\nsolves the coordinate transform problem with perfect generalization and 150\ntimes faster with 10--100 times fewer parameters than convolution. This stark\ncontrast raises the question: to what extent has this inability of convolution\npersisted insidiously inside other tasks, subtly hampering performance from\nwithin? A complete answer to this question will require further investigation,\nbut we show preliminary evidence that swapping convolution for CoordConv can\nimprove models on a diverse set of tasks. Using CoordConv in a GAN produced\nless mode collapse as the transform between high-level spatial latents and\npixels becomes easier to learn. A Faster R-CNN detection model trained on MNIST\nshowed 24% better IOU when using CoordConv, and in the RL domain agents playing\nAtari games benefit significantly from the use of CoordConv layers.\n",
"title": "An Intriguing Failing of Convolutional Neural Networks and the CoordConv Solution"
}
| null | null |
[
"Statistics"
] | null | true | null |
3268
| null |
Validated
| null | null |
null |
{
"abstract": " We study the ideal structure of reduced crossed product of topological\ndynamical systems of a countable discrete group. More concretely, for a compact\nHausdorff space $X$ with an action of a countable discrete group $\\Gamma$, we\nconsider the absence of a non-zero ideals in the reduced crossed product $C(X)\n\\rtimes_r \\Gamma$ which has a zero intersection with $C(X)$. We characterize\nthis condition by a property for amenable subgroups of the stabilizer subgroups\nof $X$ in terms of the Chabauty space of $\\Gamma$. This generalizes Kennedy's\nalgebraic characterization of the simplicity for a reduced group\n$\\mathrm{C}^{*}$-algebra of a countable discrete group.\n",
"title": "Uniformly recurrent subgroups and the ideal structure of reduced crossed products"
}
| null | null | null | null | true | null |
3269
| null |
Default
| null | null |
null |
{
"abstract": " Understanding and discovering knowledge from GPS (Global Positioning System)\ntraces of human activities is an essential topic in mobility-based urban\ncomputing. We propose TrajectoryNet-a neural network architecture for\npoint-based trajectory classification to infer real world human transportation\nmodes from GPS traces. To overcome the challenge of capturing the underlying\nlatent factors in the low-dimensional and heterogeneous feature space imposed\nby GPS data, we develop a novel representation that embeds the original feature\nspace into another space that can be understood as a form of basis expansion.\nWe also enrich the feature space via segment-based information and use Maxout\nactivations to improve the predictive power of Recurrent Neural Networks\n(RNNs). We achieve over 98% classification accuracy when detecting four types\nof transportation modes, outperforming existing models without additional\nsensory data or location-based prior knowledge.\n",
"title": "TrajectoryNet: An Embedded GPS Trajectory Representation for Point-based Classification Using Recurrent Neural Networks"
}
| null | null | null | null | true | null |
3270
| null |
Default
| null | null |
null |
{
"abstract": " We focus on two supervised visual reasoning tasks whose labels encode a\nsemantic relational rule between two or more objects in an image: the MNIST\nParity task and the colorized Pentomino task. The objects in the images undergo\nrandom translation, scaling, rotation and coloring transformations. Thus these\ntasks involve invariant relational reasoning. We report uneven performance of\nvarious deep CNN models on these two tasks. For the MNIST Parity task, we\nreport that the VGG19 model soundly outperforms a family of ResNet models.\nMoreover, the family of ResNet models exhibits a general sensitivity to random\ninitialization for the MNIST Parity task. For the colorized Pentomino task, now\nboth the VGG19 and ResNet models exhibit sluggish optimization and very poor\ntest generalization, hovering around 30% test error. The CNN we tested all\nlearn hierarchies of fully distributed features and thus encode the distributed\nrepresentation prior. We are motivated by a hypothesis from cognitive\nneuroscience which posits that the human visual cortex is modularized, and this\nallows the visual cortex to learn higher order invariances. To this end, we\nconsider a modularized variant of the ResNet model, referred to as a Residual\nMixture Network (ResMixNet) which employs a mixture-of-experts architecture to\ninterleave distributed representations with more specialized, modular\nrepresentations. We show that very shallow ResMixNets are capable of learning\neach of the two tasks well, attaining less than 2% and 1% test error on the\nMNIST Parity and the colorized Pentomino tasks respectively. Most importantly,\nthe ResMixNet models are extremely parameter efficient: generalizing better\nthan various non-modular CNNs that have over 10x the number of parameters.\nThese experimental results support the hypothesis that modularity is a robust\nprior for learning invariant relational reasoning.\n",
"title": "Modularity Matters: Learning Invariant Relational Reasoning Tasks"
}
| null | null | null | null | true | null |
3271
| null |
Default
| null | null |
null |
{
"abstract": " We offer the proofs that complete our article introducing the propositional\ncalculus called semi-intuitionistic logic with strong negation.\n",
"title": "Proofs of some Propositions of the semi-Intuitionistic Logic with Strong Negation"
}
| null | null | null | null | true | null |
3272
| null |
Default
| null | null |
null |
{
"abstract": " These lecture notes on entanglement in topological systems are part of the\n48th IFF Spring School 2017 on Topological Matter: Topological Insulators,\nSkyrmions and Majoranas at the Forschungszentrum Juelich, Germany. They cover a\nshort discussion on topologically ordered phases and review the two main tools\navailable for detecting topological order - the entanglement entropy and the\nentanglement spectrum.\n",
"title": "Entanglement in topological systems"
}
| null | null | null | null | true | null |
3273
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we study the existence and the stability in the sense of\nLyapunov of solutions for\\ differential inclusions governed by the normal cone\nto a prox-regular set and subject to a Lipschitzian perturbation. We prove that\nsuch, apparently, more general nonsmooth dynamics can be indeed remodelled into\nthe classical theory of differential inclusions involving maximal monotone\noperators. This result is new in the literature and permits us to make use of\nthe rich and abundant achievements in this class of monotone operators to\nderive the desired existence result and stability analysis, as well as the\ncontinuity and differentiability properties of the solutions. This going back\nand forth between these two models of differential inclusions is made possible\nthanks to a viability result for maximal monotone operators. As an application,\nwe study a Luenberger-like observer, which is shown to converge exponentially\nto the actual state when the initial value of the state's estimation remains in\na neighborhood of the initial value of the original system.\n",
"title": "Equivalence between Differential Inclusions Involving Prox-regular sets and maximal monotone operators"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3274
| null |
Validated
| null | null |
null |
{
"abstract": " Synthetic biological macromolecule of magnetoferritin containing an iron\noxide core inside a protein shell (apoferritin) is prepared with different\ncontent of iron. Its structure in aqueous solution is analyzed by small-angle\nsynchrotron X-ray (SAXS) and neutron (SANS) scattering. The loading factor (LF)\ndefined as the average number of iron atoms per protein is varied up to LF=800.\nWith an increase of the LF, the scattering curves exhibit a relative increase\nin the total scattered intensity, a partial smearing and a shift of the match\npoint in the SANS contrast variation data. The analysis shows an increase in\nthe polydispersity of the proteins and a corresponding effective increase in\nthe relative content of magnetic material against the protein moiety of the\nshell with the LF growth. At LFs above ~150, the apoferritin shell undergoes\nstructural changes, which is strongly indicative of the fact that the shell\nstability is affected by iron oxide presence.\n",
"title": "Effect of iron oxide loading on magnetoferritin structure in solution as revealed by SAXS and SANS"
}
| null | null | null | null | true | null |
3275
| null |
Default
| null | null |
null |
{
"abstract": " Most multispectral remote sensors (e.g. QuickBird, IKONOS, and Landsat 7\nETM+) provide low-spatial high-spectral resolution multispectral (MS) or\nhigh-spatial low-spectral resolution panchromatic (PAN) images, separately. In\norder to reconstruct a high-spatial/high-spectral resolution multispectral\nimage volume, either the information in MS and PAN images are fused (i.e.\npansharpening) or super-resolution reconstruction (SRR) is used with only MS\nimages captured on different dates. Existing methods do not utilize temporal\ninformation of MS and high spatial resolution of PAN images together to improve\nthe resolution. In this paper, we propose a multiframe SRR algorithm using\npansharpened MS images, taking advantage of both temporal and spatial\ninformation available in multispectral imagery, in order to exceed spatial\nresolution of given PAN images. We first apply pansharpening to a set of\nmultispectral images and their corresponding PAN images captured on different\ndates. Then, we use the pansharpened multispectral images as input to the\nproposed wavelet-based multiframe SRR method to yield full volumetric SRR. The\nproposed SRR method is obtained by deriving the subband relations between\nmultitemporal MS volumes. We demonstrate the results on Landsat 7 ETM+ images\ncomparing our method to conventional techniques.\n",
"title": "Volumetric Super-Resolution of Multispectral Data"
}
| null | null | null | null | true | null |
3276
| null |
Default
| null | null |
null |
{
"abstract": " Empirical scoring functions based on either molecular force fields or\ncheminformatics descriptors are widely used, in conjunction with molecular\ndocking, during the early stages of drug discovery to predict potency and\nbinding affinity of a drug-like molecule to a given target. These models\nrequire expert-level knowledge of physical chemistry and biology to be encoded\nas hand-tuned parameters or features rather than allowing the underlying model\nto select features in a data-driven procedure. Here, we develop a general\n3-dimensional spatial convolution operation for learning atomic-level chemical\ninteractions directly from atomic coordinates and demonstrate its application\nto structure-based bioactivity prediction. The atomic convolutional neural\nnetwork is trained to predict the experimentally determined binding affinity of\na protein-ligand complex by direct calculation of the energy associated with\nthe complex, protein, and ligand given the crystal structure of the binding\npose. Non-covalent interactions present in the complex that are absent in the\nprotein-ligand sub-structures are identified and the model learns the\ninteraction strength associated with these features. We test our model by\npredicting the binding free energy of a subset of protein-ligand complexes\nfound in the PDBBind dataset and compare with state-of-the-art cheminformatics\nand machine learning-based approaches. We find that all methods achieve\nexperimental accuracy and that atomic convolutional networks either outperform\nor perform competitively with the cheminformatics based methods. Unlike all\nprevious protein-ligand prediction systems, atomic convolutional networks are\nend-to-end and fully-differentiable. They represent a new data-driven,\nphysics-based deep learning model paradigm that offers a strong foundation for\nfuture improvements in structure-based bioactivity prediction.\n",
"title": "Atomic Convolutional Networks for Predicting Protein-Ligand Binding Affinity"
}
| null | null |
[
"Computer Science",
"Physics",
"Statistics"
] | null | true | null |
3277
| null |
Validated
| null | null |
null |
{
"abstract": " We define the $L$-measure on the set of Dirichlet characters as an analogue\nof the Plancherel measure, once considered as a measure on the irreducible\ncharacters of the symmetric group.\nWe compare the two measures and study the limit in distribution of characters\nevaluations when the size of the underlying group grows. These evaluations are\nproven to converge in law to imaginary exponentials of a Cauchy distribution in\nthe same way as the rescaled windings of the complex Brownian motion. This\ncontrasts with the case of the symmetric group where the renormalised\ncharacters converge in law to Gaussians after rescaling (Kerov Central Limit\nTheorem).\n",
"title": "Random characters under the $L$-measure, I : Dirichlet characters"
}
| null | null | null | null | true | null |
3278
| null |
Default
| null | null |
null |
{
"abstract": " In the present study, superheating treatment has been applied on A357\nreinforced with 0.5 wt. % (Composite 1) and 1.0 wt.% (Composite 2) continuous\nstainless steel composite. In Composite 1 the microstructure displayed poor\nbonding between matrix and reinforcement interface. Poor bonding associated\nwith large voids also can be seen in Composite 1. The results also showed that\ncoarser eutectic silicon (Si) particles were less intensified around the\nmatrix-reinforcement interface. From energy dispersive spectrometry (EDS)\nelemental mapping, it was clearly shown that the distribution of eutectic Si\nparticles were less concentrated at poor bonding regions associated with large\nvoids. Meanwhile in Composite 2, the microstructure displayed good bonding\ncombined with more concentrated finer eutectic Si particles around the\nmatrix-reinforcement interface. From EDS elemental mapping, it was clearly\nshowed more concentrated of eutectic Si particles were distributed at the good\nbonding area. The superheating prior to casting has influenced the\nmicrostructure and tends to produce finer, rounded and preferred oriented\n{\\alpha}-Al dendritic structures.\n",
"title": "The Effects of Superheating Treatment on Distribution of Eutectic Silicon Particles in A357-Continuous Stainless Steel Composite"
}
| null | null | null | null | true | null |
3279
| null |
Default
| null | null |
null |
{
"abstract": " We discuss different types of human-robot interaction paradigms in the\ncontext of training end-to-end reinforcement learning algorithms. We provide a\ntaxonomy to categorize the types of human interaction and present our\nCycle-of-Learning framework for autonomous systems that combines different\nhuman-interaction modalities with reinforcement learning. Two key concepts\nprovided by our Cycle-of-Learning framework are how it handles the integration\nof the different human-interaction modalities (demonstration, intervention, and\nevaluation) and how to define the switching criteria between them.\n",
"title": "Cycle-of-Learning for Autonomous Systems from Human Interaction"
}
| null | null | null | null | true | null |
3280
| null |
Default
| null | null |
null |
{
"abstract": " Background: Choosing the most performing method in terms of outcome\nprediction or variables selection is a recurring problem in prognosis studies,\nleading to many publications on methods comparison. But some aspects have\nreceived little attention. First, most comparison studies treat prediction\nperformance and variable selection aspects separately. Second, methods are\neither compared within a binary outcome setting (based on an arbitrarily chosen\ndelay) or within a survival setting, but not both. In this paper, we propose a\ncomparison methodology to weight up those different settings both in terms of\nprediction and variables selection, while incorporating advanced machine\nlearning strategies. Methods: Using a high-dimensional case study on a\nsickle-cell disease (SCD) cohort, we compare 8 statistical methods. In the\nbinary outcome setting, we consider logistic regression (LR), support vector\nmachine (SVM), random forest (RF), gradient boosting (GB) and neural network\n(NN); while on the survival analysis setting, we consider the Cox Proportional\nHazards (PH), the CURE and the C-mix models. We then compare performances of\nall methods both in terms of risk prediction and variable selection, with a\nfocus on the use of Elastic-Net regularization technique. Results: Among all\nassessed statistical methods assessed, the C-mix model yields the better\nperformances in both the two considered settings, as well as interesting\ninterpretation aspects. There is some consistency in selected covariates across\nmethods within a setting, but not much across the two settings. Conclusions: It\nappears that learning withing the survival setting first, and then going back\nto a binary prediction using the survival estimates significantly enhance\nbinary predictions.\n",
"title": "Comparison of methods for early-readmission prediction in a high-dimensional heterogeneous covariates and time-to-event outcome framework"
}
| null | null | null | null | true | null |
3281
| null |
Default
| null | null |
null |
{
"abstract": " This contribution will show how Access play a strong role in the creation and\nstructuring of DARIAH, a European Digital Research Infrastructure in Arts and\nHumanities.To achieve this goal, this contribution will develop the concept of\nAccess from five examples: Interdisciplinarity point of view, Manage\ncontradiction between national and international perspectives, Involve\ndifferent communities (not only researchers stakeholders), Manage tools and\nservices, Develop and use new collaboration tools. We would like to demonstrate\nthat speaking about Access always implies a selection, a choice, even in the\nperspective of \"Open Access\".\n",
"title": "How the notion of ACCESS guides the organization of a European research infrastructure: the example of DARIAH"
}
| null | null |
[
"Computer Science"
] | null | true | null |
3282
| null |
Validated
| null | null |
null |
{
"abstract": " In this work we mainly consider the dynamics and scattering of a narrow\nsoliton of NLS equation with a potential in $\\mathbb{R}^3$, where the\nasymptotic state of the system can be far from the initial state in parameter\nspace. Specifically, if we let a narrow soliton state with initial velocity\n$\\upsilon_{0}$ to interact with an extra potential $V(x)$, then the velocity\n$\\upsilon_{+}$ of outgoing solitary wave in infinite time will in general be\nvery different from $\\upsilon_{0}$. In contrast to our present work, previous\nworks proved that the soliton is asymptotically stable under the assumption\nthat $\\upsilon_{+}$ stays close to $\\upsilon_{0}$ in a certain manner.\n",
"title": "Soliton-potential interactions for nonlinear Schrödinger equation in $\\mathbb{R}^3$"
}
| null | null | null | null | true | null |
3283
| null |
Default
| null | null |
null |
{
"abstract": " Incommensurately modulated twin structure of nyerereite Na1.64K0.36Ca(CO3)2\nhas been first determined in the (3+1)D symmetry group Cmcm({\\alpha}00)00s with\nmodulation vector q = 0.383a*. Unit-cell values are a = 5.062(1), b = 8.790(1),\nc = 12.744(1) {\\AA}. Three orthorhombic components are related by threefold\nrotation about [001]. Discontinuous crenel functions are used to describe\noccupation modulation of Ca and some CO3 groups. Strong displacive modulation\nof the oxygen atoms in vertexes of such CO3 groups is described using\nx-harmonics in crenel intervals. The Na, K atoms occupy mixed sites whose\noccupation modulation is described by two ways using either complementary\nharmonic functions or crenels. The nyerereite structure has been compared both\nwith commensurately modulated structure of K-free Na2Ca(CO3)2 and with widely\nknown incommensurately modulated structure of {\\gamma}-Na2CO3.\n",
"title": "Incommensurately modulated twin structure of nyerereite Na1.64K0.36Ca(CO3)2"
}
| null | null | null | null | true | null |
3284
| null |
Default
| null | null |
null |
{
"abstract": " The cortex exhibits self-sustained highly-irregular activity even under\nresting conditions, whose origin and function need to be fully understood. It\nis believed that this can be described as an \"asynchronous state\" stemming from\nthe balance between excitation and inhibition, with important consequences for\ninformation-processing, though a competing hypothesis claims it stems from\ncritical dynamics. By analyzing a parsimonious neural-network model with\nexcitatory and inhibitory interactions, we elucidate a noise-induced mechanism\ncalled \"Jensen's force\" responsible for the emergence of a novel phase of\narbitrarily-low but self-sustained activity, which reproduces all the\nexperimental features of asynchronous states. The simplicity of our framework\nallows for a deep understanding of asynchronous states from a broad\nstatistical-mechanics perspective and of the phase transitions to other\nstandard phases it exhibits, opening the door to reconcile, asynchronous-state\nand critical-state hypotheses. We argue that Jensen's forces are measurable\nexperimentally and might be relevant in contexts beyond neuroscience.\n",
"title": "Jensen's force and the statistical mechanics of cortical asynchronous states"
}
| null | null | null | null | true | null |
3285
| null |
Default
| null | null |
null |
{
"abstract": " This paper deals with relative normalizations of skew ruled surfaces in the\nEuclidean space $\\mathbb{E}^{3}$. In section 2 we investigate some new formulae\nconcerning the Pick invariant, the relative curvature, the relative mean\ncurvature and the curvature of the relative metric of a relatively normalized\nruled surface $\\varPhi$ and in section 3 we introduce some special\nnormalizations of it. All ruled surfaces and their corresponding normalizations\nthat make $\\varPhi$ an improper or a proper relative sphere are determined in\nsection 4. In the last section we study ruled surfaces, which are\n\\emph{centrally} normalized, i.e., their relative normals at each point lie on\nthe corresponding central plane. Especially we study various properties of the\nTchebychev vector field. We conclude the paper by the study of the central\nimage of $\\varPhi$.\n",
"title": "Notes on relative normalizations of ruled surfaces in the three-dimensional Euclidean space"
}
| null | null |
[
"Mathematics"
] | null | true | null |
3286
| null |
Validated
| null | null |
null |
{
"abstract": " GeTe wins the renewed research interest due to its giant bulk Rashba spin\norbit coupling (SOC), and becomes the father of a new multifunctional material,\ni.e., ferroelectric Rashba semiconductor. In the present work, we investigate\nRashba SOC at the interface of the ferroelectric semiconductor superlattice\nGeTe(111)/InP(111) by using the first principles calculation. Contribution of\nthe interface electric field and the ferroelectric field to Rashba SOC is\nrevealed. A large modulation to Rashba SOC and a reversal of the spin\npolarization is obtained by switching the ferroelectric polarization. Our\ninvestigation about GeTe(111)/InP(111) superlattice is of great importance in\nthe application of ferroelectric Rashba semiconductor in the spin field effect\ntransistor.\n",
"title": "Ferroelectric control of the giant Rashba spin orbit coupling in GeTe(111)/InP(111) superlattice"
}
| null | null |
[
"Physics"
] | null | true | null |
3287
| null |
Validated
| null | null |
null |
{
"abstract": " We show, theoretically, that the phase of the interlayer exchange coupling\n(IEC) undergoes a topological change of approximately $2\\pi$ as the chemical\npotential of the ferromagnetic (FM) lead moves across a hybridization gap (HG).\nThe effect is largely independent of the detailed parameters of the system, in\nparticular the width of the gap. The implication is that for a narrow gap, a\nsmall perturbation in the chemical potential of the lead can give a sign\nreversal of the exchange coupling. This offers the possibility of controlling\nmagnetization switching in spintronic devices such as MRAM, with little power\nconsumption. Furthermore we believe that this effect has already been\nindirectly observed, in existing measurements of the IEC as a function of\ntemperature and of doping of the leads.\n",
"title": "Topological phase of the interlayer exchange coupling with application to magnetic switching"
}
| null | null | null | null | true | null |
3288
| null |
Default
| null | null |
null |
{
"abstract": " Observations of the CMB today allow us to answer detailed questions about the\nproperties of our Universe, targeting both standard and non-standard physics.\nIn this paper, we study the effects of varying fundamental constants (i.e., the\nfine-structure constant, $\\alpha_{\\rm EM}$, and electron rest mass, $m_{\\rm\ne}$) around last scattering using the recombination codes CosmoRec and\nRecfast++. We approach the problem in a pedagogical manner, illustrating the\nimportance of various effects on the free electron fraction, Thomson visibility\nfunction and CMB power spectra, highlighting various degeneracies. We\ndemonstrate that the simpler Recfast++ treatment (based on a three-level atom\napproach) can be used to accurately represent the full computation of CosmoRec.\nWe also include explicit time-dependent variations using a phenomenological\npower-law description. We reproduce previous Planck 2013 results in our\nanalysis. Assuming constant variations relative to the standard values, we find\nthe improved constraints $\\alpha_{\\rm EM}/\\alpha_{\\rm EM,0}=0.9993\\pm 0.0025$\n(CMB only) and $m_{\\rm e}/m_{\\rm e,0}= 1.0039 \\pm 0.0074$ (including BAO) using\nPlanck 2015 data. For a redshift-dependent variation, $\\alpha_{\\rm\nEM}(z)=\\alpha_{\\rm EM}(z_0)\\,[(1+z)/1100]^p$ with $\\alpha_{\\rm\nEM}(z_0)\\equiv\\alpha_{\\rm EM,0}$ at $z_0=1100$, we obtain $p=0.0008\\pm 0.0025$.\nAllowing simultaneous variations of $\\alpha_{\\rm EM}(z_0)$ and $p$ yields\n$\\alpha_{\\rm EM}(z_0)/\\alpha_{\\rm EM,0} = 0.9998\\pm 0.0036$ and $p = 0.0006\\pm\n0.0036$. We also discuss combined limits on $\\alpha_{\\rm EM}$ and $m_{\\rm e}$.\nOur analysis shows that existing data is not only sensitive to the value of the\nfundamental constants around recombination but also its first time derivative.\nThis suggests that a wider class of varying fundamental constant models can be\nprobed using the CMB.\n",
"title": "New constraints on time-dependent variations of fundamental constants using Planck data"
}
| null | null | null | null | true | null |
3289
| null |
Default
| null | null |
null |
{
"abstract": " We describe a parallel, adaptive, multi-block algorithm for explicit\nintegration of time dependent partial differential equations on two-dimensional\nCartesian grids. The grid layout we consider consists of a nested hierarchy of\nfixed size, non-overlapping, logically Cartesian grids stored as leaves in a\nquadtree. Dynamic grid refinement and parallel partitioning of the grids is\ndone through the use of the highly scalable quadtree/octree library p4est.\nBecause our concept is multi-block, we are able to easily solve on a variety of\ngeometries including the cubed sphere. In this paper, we pay special attention\nto providing details of the parallel ghost-filling algorithm needed to ensure\nthat both corner and edge ghost regions around each grid hold valid values.\nWe have implemented this algorithm in the ForestClaw code using single-grid\nsolvers from ClawPack, a software package for solving hyperbolic PDEs using\nfinite volumes methods. We show weak and strong scalability results for scalar\nadvection problems on two-dimensional manifold domains on 1 to 64Ki MPI\nprocesses, demonstrating neglible regridding overhead.\n",
"title": "ForestClaw: A parallel algorithm for patch-based adaptive mesh refinement on a forest of quadtrees"
}
| null | null | null | null | true | null |
3290
| null |
Default
| null | null |
null |
{
"abstract": " Triangle counting is a key algorithm for large graph analysis. The Graphulo\nlibrary provides a framework for implementing graph algorithms on the Apache\nAccumulo distributed database. In this work we adapt two algorithms for\ncounting triangles, one that uses the adjacency matrix and another that also\nuses the incidence matrix, to the Graphulo library for server-side processing\ninside Accumulo. Cloud-based experiments show a similar performance profile for\nthese different approaches on the family of power law Graph500 graphs, for\nwhich data skew increasingly bottlenecks. These results motivate the design of\nskew-aware hybrid algorithms that we propose for future work.\n",
"title": "Distributed Triangle Counting in the Graphulo Matrix Math Library"
}
| null | null | null | null | true | null |
3291
| null |
Default
| null | null |
null |
{
"abstract": " The growing importance and utilization of measuring brain waves (e.g. EEG\nsignals of eye state) in brain-computer interface (BCI) applications\nhighlighted the need for suitable classification methods. In this paper, a\ncomparison between three of well-known classification methods (i.e. support\nvector machine (SVM), hidden Markov map (HMM), and radial basis function (RBF))\nfor EEG based eye state classification was achieved. Furthermore, a suggested\nmethod that is based on ensemble model was tested. The suggested (ensemble\nsystem) method based on a voting algorithm with two kernels: random forest (RF)\nand Kstar classification methods. The performance was tested using three\nmeasurement parameters: accuracy, mean absolute error (MAE), and confusion\nmatrix. Results showed that the proposed method outperforms the other tested\nmethods. For instance, the suggested method's performance was 97.27% accuracy\nand 0.13 MAE.\n",
"title": "Ensemble Classifier for Eye State Classification using EEG Signals"
}
| null | null | null | null | true | null |
3292
| null |
Default
| null | null |
null |
{
"abstract": " Diet design for vegetarian health is challenging due to the limited food\nrepertoire of vegetarians. This challenge can be partially overcome by\nquantitative, data-driven approaches that utilise massive nutritional\ninformation collected for many different foods. Based on large-scale data of\nfoods' nutrient compositions, the recent concept of nutritional fitness helps\nquantify a nutrient balance within each food with regard to satisfying daily\nnutritional requirements. Nutritional fitness offers prioritisation of\nrecommended foods using the foods' occurrence in nutritionally adequate food\ncombinations. Here, we systematically identify nutritionally recommendable\nfoods for semi- to strict vegetarian diets through the computation of\nnutritional fitness. Along with commonly recommendable foods across different\ndiets, our analysis reveals favourable foods specific to each diet, such as\nimmature lima beans for a vegan diet as an amino acid and choline source, and\nmushrooms for ovo-lacto vegetarian and vegan diets as a vitamin D source.\nFurthermore, we find that selenium and other essential micronutrients can be\nsubject to deficiency in plant-based diets, and suggest nutritionally-desirable\ndietary patterns. We extend our analysis to two hypothetical scenarios of\nhighly personalised, plant-based methionine-restricted diets. Our\nnutrient-profiling approach may provide a useful guide for designing different\ntypes of personalised vegetarian diets.\n",
"title": "Nutritionally recommended food for semi- to strict vegetarian diets based on large-scale nutrient composition data"
}
| null | null | null | null | true | null |
3293
| null |
Default
| null | null |
null |
{
"abstract": " We show that learning methods interpolating the training data can achieve\noptimal rates for the problems of nonparametric regression and prediction with\nsquare loss.\n",
"title": "Does data interpolation contradict statistical optimality?"
}
| null | null | null | null | true | null |
3294
| null |
Default
| null | null |
null |
{
"abstract": " Search of new frustrated magnetic systems is of a significant importance for\nphysics studying the condensed matter. The platform for geometric frustration\nof magnetic systems can be provided by copper oxocentric tetrahedra (OCu4)\nforming the base of crystalline structures of copper minerals from Tolbachik\nvolcanos in Kamchatka. The present work was devoted to a new frustrated\nantiferromagnetic - kamchatkite (KCu3OCl(SO4)2). The calculation of the sign\nand strength of magnetic couplings in KCu3OCl(SO4)2 has been performed on the\nbasis of structural data by the phenomenological crystal chemistry method with\ntaking into account corrections on the Jahn-Teller orbital degeneracy of Cu2.\nIt has been established that kamchatkite (KCu3OCl(SO4)2) contains AFM\nspin-frustrated chains of the pyrochlore type composed of cone-sharing Cu4\ntetrahedra. Strong AFM intrachain and interchain couplings compete with each\nother. Frustration of magnetic couplings in tetrahedral chains is combined with\nthe presence of electric polarization.\n",
"title": "Spin-Frustrated Pyrochlore Chains in the Volcanic Mineral Kamchatkite (KCu3OCl(SO4)2)"
}
| null | null |
[
"Physics"
] | null | true | null |
3295
| null |
Validated
| null | null |
null |
{
"abstract": " Previous CNN-based video super-resolution approaches need to align multiple\nframes to the reference. In this paper, we show that proper frame alignment and\nmotion compensation is crucial for achieving high quality results. We\naccordingly propose a `sub-pixel motion compensation' (SPMC) layer in a CNN\nframework. Analysis and experiments show the suitability of this layer in video\nSR. The final end-to-end, scalable CNN framework effectively incorporates the\nSPMC layer and fuses multiple frames to reveal image details. Our\nimplementation can generate visually and quantitatively high-quality results,\nsuperior to current state-of-the-arts, without the need of parameter tuning.\n",
"title": "Detail-revealing Deep Video Super-resolution"
}
| null | null | null | null | true | null |
3296
| null |
Default
| null | null |
null |
{
"abstract": " Large organizations often have users in multiple sites which are connected\nover the Internet. Since resources are limited, communication between these\nsites needs to be carefully orchestrated for the most benefit to the\norganization. We present a Mission-optimized Overlay Network (MON), a hybrid\noverlay network architecture for maximizing utility to the organization. We\ncombine an offline and an online system to solve non-concave utility\nmaximization problems. The offline tier, the Predictive Flow Optimizer (PFO),\ncreates plans for routing traffic using a model of network conditions. The\nonline tier, MONtra, is aware of the precise local network conditions and is\nable to react quickly to problems within the network. Either tier alone is\ninsufficient. The PFO may take too long to react to network changes. MONtra\nonly has local information and cannot optimize non-concave mission utilities.\nHowever, by combining the two systems, MON is robust and achieves near-optimal\nutility under a wide range of network conditions. While best-effort overlay\nnetworks are well studied, our work is the first to design overlays that are\noptimized for mission utility.\n",
"title": "MON: Mission-optimized Overlay Networks"
}
| null | null | null | null | true | null |
3297
| null |
Default
| null | null |
null |
{
"abstract": " We show that some results from the theory of group automata and monoid\nautomata still hold for more general classes of monoids and models. Extending\nprevious work for finite automata over commutative groups, we demonstrate a\ncontext-free language that can not be recognized by any rational monoid\nautomaton over a finitely generated permutable monoid. We show that the class\nof languages recognized by rational monoid automata over finitely generated\ncompletely simple or completely 0-simple permutable monoids is a semi-linear\nfull trio. Furthermore, we investigate valence pushdown automata, and prove\nthat they are only as powerful as (finite) valence automata. We observe that\ncertain results proven for monoid automata can be easily lifted to the case of\ncontext-free valence grammars.\n",
"title": "Generalized Results on Monoids as Memory"
}
| null | null | null | null | true | null |
3298
| null |
Default
| null | null |
null |
{
"abstract": " Multicarrier-low density spreading multiple access (MC-LDSMA) is a promising\nmultiple access technique that enables near optimum multiuser detection. In\nMC-LDSMA, each user's symbol spread on a small set of subcarriers, and each\nsubcarrier is shared by multiple users. The unique structure of MC-LDSMA makes\nthe radio resource allocation more challenging comparing to some well-known\nmultiple access techniques. In this paper, we study the radio resource\nallocation for single-cell MC-LDSMA system. Firstly, we consider the\nsingle-user case, and derive the optimal power allocation and subcarriers\npartitioning schemes. Then, by capitalizing on the optimal power allocation of\nthe Gaussian multiple access channel, we provide an optimal solution for\nMC-LDSMA that maximizes the users' weighted sum-rate under relaxed constraints.\nDue to the prohibitive complexity of the optimal solution, suboptimal\nalgorithms are proposed based on the guidelines inferred by the optimal\nsolution. The performance of the proposed algorithms and the effect of\nsubcarrier loading and spreading are evaluated through Monte Carlo simulations.\nNumerical results show that the proposed algorithms significantly outperform\nconventional static resource allocation, and MC-LDSMA can improve the system\nperformance in terms of spectral efficiency and fairness in comparison with\nOFDMA.\n",
"title": "Radio Resource Allocation for Multicarrier-Low Density Spreading Multiple Access"
}
| null | null | null | null | true | null |
3299
| null |
Default
| null | null |
null |
{
"abstract": " Many important problems are characterized by the eigenvalues of a large\nmatrix. For example, the difficulty of many optimization problems, such as\nthose arising from the fitting of large models in statistics and machine\nlearning, can be investigated via the spectrum of the Hessian of the empirical\nloss function. Network data can be understood via the eigenstructure of a graph\nLaplacian matrix using spectral graph theory. Quantum simulations and other\nmany-body problems are often characterized via the eigenvalues of the solution\nspace, as are various dynamic systems. However, naive eigenvalue estimation is\ncomputationally expensive even when the matrix can be represented; in many of\nthese situations the matrix is so large as to only be available implicitly via\nproducts with vectors. Even worse, one may only have noisy estimates of such\nmatrix vector products. In this work, we combine several different techniques\nfor randomized estimation and show that it is possible to construct unbiased\nestimators to answer a broad class of questions about the spectra of such\nimplicit matrices, even in the presence of noise. We validate these methods on\nlarge-scale problems in which graph theory and random matrix theory provide\nground truth.\n",
"title": "Estimating the Spectral Density of Large Implicit Matrices"
}
| null | null | null | null | true | null |
3300
| null |
Default
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.