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": " Casimir forces between material surfaces at close proximity of less than 200\nnm can lead to increased chaotic behavior of actuating devices depending on the\nstrength of the Casimir interaction. We investigate these phenomena for phase\nchange materials in torsional oscillators, where the amorphous to crystalline\nphase transitions lead to transitions between high and low Casimir force and\ntorque states respectively, without material compositions. For a conservative\nsystem bifurcation curve and Poincare maps analysis show the absence of chaotic\nbehavior but with the crystalline phase (high force/torque state) favoring more\nunstable behavior and stiction. However, for a non-conservative system chaotic\nbehavior can occur introducing significant risk for stiction, which is again\nmore pronounced for the crystalline phase. The latter illustrates the more\ngeneral scenario that stronger Casimir forces and torques increase the\npossibility for chaotic behavior. The latter is making impossible to predict\nwhether stiction or stable actuation will occur on a long term basis, and it is\nsetting limitations in the design of micro/nano devices operating at short\nrange nanoscale separations.\n",
"title": "Chaotic behavior in Casimir oscillators: A case study for phase change materials"
} | null | null | null | null | true | null | 20501 | null | Default | null | null |
null | {
"abstract": " Driven by the interest of reasoning about probabilistic programming\nlanguages, we set out to study a notion of unicity of normal forms for them. To\nprovide a tractable proof method for it, we define a property of distribution\nconfluence which is shown to imply the desired uniqueness (even for infinite\nsequences of reduction) and further properties. We then carry over several\ncriteria from the classical case, such as Newman's lemma, to simplify proving\nconfluence in concrete languages. Using these criteria, we obtain simple proofs\nof confluence for $\\lambda_1$, an affine probabilistic $\\lambda$-calculus, and\nfor Q$^*$, a quantum programming language for which a related property has\nalready been proven in the literature.\n",
"title": "Confluence in Probabilistic Rewriting"
} | null | null | null | null | true | null | 20502 | null | Default | null | null |
null | {
"abstract": " In this paper we consider the development of numerical schemes for mean-field\nequations describing the collective behavior of a large group of interacting\nagents. The schemes are based on a generalization of the classical Chang-Cooper\napproach and are capable to preserve the main structural properties of the\nsystems, namely nonnegativity of the solution, physical conservation laws,\nentropy dissipation and stationary solutions. In particular, the methods here\nderived are second order accurate in transient regimes whereas they can reach\narbitrary accuracy asymptotically for large times. Several examples are\nreported to show the generality of the approach.\n",
"title": "Structure preserving schemes for mean-field equations of collective behavior"
} | null | null | null | null | true | null | 20503 | null | Default | null | null |
null | {
"abstract": " Localization-based imaging has revolutionized fluorescence optical microscopy\nand has also enabled unprecedented ultrasound images of microvascular\nstructures in deep tissues. Herein, we introduce a new concept of localization\noptoacoustic tomography (LOAT) that employs rapid sequential acquisition of\nthree-dimensional optoacoustic images from flowing absorbing particles. We show\nthat the new method enables breaking through the spatial resolution barrier of\nacoustic diffraction while further enhancing the visibility of structures under\nlimited-view tomographic conditions. Given the intrinsic sensitivity of\noptoacoustics to multiple hemodynamic and oxygenation parameters, LOAT may\nenable new level of performance in studying functional and anatomical\nalterations of microcirculation.\n",
"title": "Localization optoacoustic tomography"
} | null | null | null | null | true | null | 20504 | null | Default | null | null |
null | {
"abstract": " Within the standard model of many-body localization, i.e., the disordered\nchain of spinless fermions, we investigate how the interaction affects the\nmany-body states in the basis of noninteracting localized Anderson states. From\nthis starting point we follow the approach that uses a reduced basis of\nmany-body states. Together with an extrapolation to the full basis, it proves\nto be efficient for the evaluation of the stiffnesses of local observables,\nwhich remain finite within the non-ergodic regime and represent the hallmark of\nthe many-body localization (MBL). The method enables a larger span of system\nsizes and, within the MBL regime, allows for a more careful analysis of the\nsize scaling of calculated quantities. On the other hand, the survival\nstiffness as the representative of non--local quantities, reveals limitations\nof the reduced-basis approach.\n",
"title": "Reduced-basis approach to many-body localization"
} | null | null | null | null | true | null | 20505 | null | Default | null | null |
null | {
"abstract": " Spectroscopic properties, useful for plasma diagnostics and astrophysics, of\na few rubidium-like ions are studied here. We choose one of the simplest, but\ncorrelationally challenging series where $d-$ and $f-$ orbitals are present in\nthe core and/or valence shells with $4d$ $^2D_{3/2}$ as the ground state. We\nstudy different correlation characteristics of this series and make precise\ncalculations of electronic structure and rates of electromagnetic transitions.\nOur calculated lifetimes and transition rates are compared with other available\nexperimental and theoretical values. Radiative rates of vacuum ultra-violet\nelectromagnetic transitions of the long lived Tc$^{6+}$ ion, useful in several\nareas of physics and chemistry, are estimated. To the best of our knowledge,\nthere is no literature for most of these transitions.\n",
"title": "Electron-correlation study of Y III-Tc VII ions using a relativistic coupled-cluster theory"
} | null | null | null | null | true | null | 20506 | null | Default | null | null |
null | {
"abstract": " This paper presents the submissions by the University of Zurich to the\nSIGMORPHON 2017 shared task on morphological reinflection. The task is to\npredict the inflected form given a lemma and a set of morpho-syntactic\nfeatures. We focus on neural network approaches that can tackle the task in a\nlimited-resource setting. As the transduction of the lemma into the inflected\nform is dominated by copying over lemma characters, we propose two recurrent\nneural network architectures with hard monotonic attention that are strong at\ncopying and, yet, substantially different in how they achieve this. The first\napproach is an encoder-decoder model with a copy mechanism. The second approach\nis a neural state-transition system over a set of explicit edit actions,\nincluding a designated COPY action. We experiment with character alignment and\nfind that naive, greedy alignment consistently produces strong results for some\nlanguages. Our best system combination is the overall winner of the SIGMORPHON\n2017 Shared Task 1 without external resources. At a setting with 100 training\nsamples, both our approaches, as ensembles of models, outperform the next best\ncompetitor.\n",
"title": "Align and Copy: UZH at SIGMORPHON 2017 Shared Task for Morphological Reinflection"
} | null | null | [
"Computer Science"
]
| null | true | null | 20507 | null | Validated | null | null |
null | {
"abstract": " We consider generalizations of the Sylvester matrix equation, consisting of\nthe sum of a Sylvester operator and a linear operator $\\Pi$ with a particular\nstructure. More precisely, the commutator of the matrix coefficients of the\noperator $\\Pi$ and the Sylvester operator coefficients are assumed to be\nmatrices with low rank. We show (under certain additional conditions) low-rank\napproximability of this problem, i.e., the solution to this matrix equation can\nbe approximated with a low-rank matrix. Projection methods have successfully\nbeen used to solve other matrix equations with low-rank approximability. We\npropose a new projection method for this class of matrix equations. The choice\nof subspace is a crucial ingredient for any projection method for matrix\nequations. Our method is based on an adaption and extension of the extended\nKrylov subspace method for Sylvester equations. A constructive choice of the\nstarting vector/block is derived from the low-rank commutators. We illustrate\nthe effectiveness of our method by solving large-scale matrix equations arising\nfrom applications in control theory and the discretization of PDEs. The\nadvantages of our approach in comparison to other methods are also illustrated.\n",
"title": "Krylov methods for low-rank commuting generalized Sylvester equations"
} | null | null | null | null | true | null | 20508 | null | Default | null | null |
null | {
"abstract": " Viscoelasticity has been described since the time of Maxwell as an\ninterpolation of purely viscous and purely elastic response, but its\nmicroscopic atomic-level mechanism in solids has remained elusive. We studied\nthree model disordered solids: a random lattice, the bond-depleted fcc lattice,\nand the fcc lattice with vacancies. Within the harmonic approximation for\ncentral-force lattices, we applied sum-rules for viscoelastic response derived\non the basis of non-affine atomic motions. The latter motions are a direct\nresult of local structural disorder, and in particular, of the lack of\ninversion-symmetry in disordered lattices. By defining a suitable quantitative\nand general atomic-level measure of nonaffinity and inversion-symmetry, we show\nthat the viscoelastic responses of all three systems collapse onto a master\ncurve upon normalizing by the overall strength of inversion-symmetry breaking\nin each system. Close to the isostatic point for central-force lattices,\npower-law creep $G(t)\\sim t^{-1/2}$ emerges as a consequence of the interplay\nbetween soft vibrational modes and non-affine dynamics, and various analytical\nscalings, supported by numerical calculations, are predicted by the theory.\n",
"title": "Atomic-scale origin of dynamic viscoelastic response and creep in disordered solids"
} | null | null | null | null | true | null | 20509 | null | Default | null | null |
null | {
"abstract": " Adaptive optic (AO) systems delivering high levels of wavefront correction\nare now common at observatories. One of the main limitations to image quality\nafter wavefront correction comes from atmospheric refraction. An Atmospheric\ndispersion compensator (ADC) is employed to correct for atmospheric refraction.\nThe correction is applied based on a look-up table consisting of dispersion\nvalues as a function of telescope elevation angle. The look-up table based\ncorrection of atmospheric dispersion results in imperfect compensation leading\nto the presence of residual dispersion in the point-spread function (PSF) and\nis insufficient when sub-milliarcsecond precision is required. The presence of\nresidual dispersion can limit the achievable contrast while employing\nhigh-performance coronagraphs or can compromise high-precision astrometric\nmeasurements. In this paper, we present the first on-sky closed-loop correction\nof atmospheric dispersion by directly using science path images. The concept\nbehind the measurement of dispersion utilizes the chromatic scaling of focal\nplane speckles. An adaptive speckle grid generated with a deformable mirror\n(DM) that has a sufficiently large number of actuators is used to accurately\nmeasure the residual dispersion and subsequently correct it by driving the ADC.\nWe have demonstrated with the Subaru Coronagraphic Extreme AO (SCExAO) system\non-sky closed-loop correction of residual dispersion to < 1 mas across H-band.\nThis work will aid in the direct detection of habitable exoplanets with\nupcoming extremely large telescopes (ELTs) and also provide a diagnostic tool\nto test the performance of instruments which require sub-milliarcsecond\ncorrection.\n",
"title": "On-sky closed loop correction of atmospheric dispersion for high-contrast coronagraphy and astrometry"
} | null | null | null | null | true | null | 20510 | null | Default | null | null |
null | {
"abstract": " The progress made in understanding spontaneous toroidal rotation reversals in\ntokamaks is reviewed and current ideas to solve this ten-year-old puzzle are\nexplored. The paper includes a summarial synthesis of the experimental\nobservations in AUG, C-Mod, KSTAR, MAST and TCV tokamaks, reasons why turbulent\nmomentum transport is thought to be responsible for the reversals, a review of\nthe theory of turbulent momentum transport and suggestions for future\ninvestigations.\n",
"title": "Experimental observations and modelling of intrinsic rotation reversals in tokamaks"
} | null | null | null | null | true | null | 20511 | null | Default | null | null |
null | {
"abstract": " Wireless engineers and business planners commonly raise the question on\nwhere, when, and how millimeter-wave (mmWave) will be used in 5G and beyond.\nSince the next generation network is not just a new radio access standard, but\ninstead an integration of networks for vertical markets with diverse\napplications, answers to the question depend on scenarios and use cases to be\ndeployed. This paper gives four 5G mmWave deployment examples and describes in\nchronological order the scenarios and use cases of their probable deployment,\nincluding expected system architectures and hardware prototypes. The paper\nstarts with 28 GHz outdoor backhauling for fixed wireless access and moving\nhotspots, which will be demonstrated at the PyeongChang winter Olympic games in\n2018. The second deployment example is a 60 GHz unlicensed indoor access system\nat the Tokyo-Narita airport, which is combined with Mobile Edge Computing (MEC)\nto enable ultra-high speed content download with low latency. The third example\nis mmWave mesh network to be used as a micro Radio Access Network ({\\mu}-RAN),\nfor cost-effective backhauling of small-cell Base Stations (BSs) in dense urban\nscenarios. The last example is mmWave based Vehicular-to-Vehicular (V2V) and\nVehicular-to-Everything (V2X) communications system, which enables automated\ndriving by exchanging High Definition (HD) dynamic map information between cars\nand Roadside Units (RSUs). For 5G and beyond, mmWave and MEC will play\nimportant roles for a diverse set of applications that require both ultra-high\ndata rate and low latency communications.\n",
"title": "Where, When, and How mmWave is Used in 5G and Beyon"
} | null | null | null | null | true | null | 20512 | null | Default | null | null |
null | {
"abstract": " The scripting language described in this document is (in the first place)\nintended to be used on robots developed by Anja M{\\o}lle Lindelof and Henning\nChristiansen as part of a research project about robots performing on stage.\nThe target robots are expected to appear as familiar domestic objects that\ntake their own life, so to speak, and perhaps perform together with human\nplayers, creating at illusion of a communication between them. In the current\nversion, these robots' common behaviour is determined uniquely by a script\nwritten in the language described here -- the only possible autonomy for the\nrobots is action to correct dynamically for inaccuracies that arise during a\nperformance.\nThe present work is preliminary and has not been compared to properly to\nother research work in this area, and the testing is still limited. A first\nimplementation on small Lego Mindstorms based robots is under development by\nMads Saustrup Fox as part of his master thesis work.\n",
"title": "A simple script language for choreography of multiple, synchronizing non-anthropomorphic robots"
} | null | null | null | null | true | null | 20513 | null | Default | null | null |
null | {
"abstract": " We consider abstract evolution equations with on-off time delay feedback.\nWithout the time delay term, the model is described by an exponentially stable\nsemigroup. We show that, under appropriate conditions involving the delay term,\nthe system remains asymptotically stable. Under additional assumptions\nexponential stability results are also obtained. Concrete examples illustrating\nthe abstract results are finally given.\n",
"title": "Stability results for abstract evolution equations with intermittent time-delay feedback"
} | null | null | null | null | true | null | 20514 | null | Default | null | null |
null | {
"abstract": " Spectral properties of the turbulent cascade from fluid to kinetic scales in\ncollisionless plasmas are investigated by means of large-size three-dimensional\n(3D) hybrid (fluid electrons, kinetic protons) particle-in-cell simulations.\nInitially isotropic Alfvènic fluctuations rapidly develop a strongly\nanisotropic turbulent cascade, mainly in the direction perpendicular to the\nambient magnetic field. The omnidirectional magnetic field spectrum shows a\ndouble power-law behavior over almost two decades in wavenumber, with a\nKolmogorov-like index at large scales, a spectral break around ion scales, and\na steepening at sub-ion scales. Power laws are also observed in the spectra of\nthe ion bulk velocity, density, and electric field, both at magnetohydrodynamic\n(MHD) and at kinetic scales. Despite the complex structure, the omnidirectional\nspectra of all fields at ion and sub-ion scales are in remarkable quantitative\nagreement with those of a two-dimensional (2D) simulation with similar physical\nparameters. This provides a partial, a-posteriori validation of the 2D\napproximation at kinetic scales. Conversely, at MHD scales, the spectra of the\ndensity and of the velocity (and, consequently, of the electric field) exhibit\ndifferences between the 2D and 3D cases. Although they can be partly ascribed\nto the lower spatial resolution, the main reason is likely the larger\nimportance of compressible effects in a full geometry. Our findings are also in\nremarkable quantitative agreement with solar wind observations.\n",
"title": "Solar wind turbulent cascade from MHD to sub-ion scales: large-size 3D hybrid particle-in-cell simulations"
} | null | null | [
"Physics"
]
| null | true | null | 20515 | null | Validated | null | null |
null | {
"abstract": " The blind source separation model for multivariate time series generally\nassumes that the observed series is a linear transformation of an unobserved\nseries with temporally uncorrelated or independent components. Given the\nobservations, the objective is to find a linear transformation that recovers\nthe latent series. Several methods for accomplishing this exist and three\nparticular ones are the classic SOBI and the recently proposed generalized FOBI\n(gFOBI) and generalized JADE (gJADE), each based on the use of joint lagged\nmoments. In this paper we generalize the methodologies behind these algorithms\nfor tensor-valued time series. We assume that our data consists of a tensor\nobserved at each time point and that the observations are linear\ntransformations of latent tensors we wish to estimate. The tensorial\ngeneralizations are shown to have particularly elegant forms and we show that\neach of them is Fisher consistent and orthogonal equivariant. Comparing the new\nmethods with the original ones in various settings shows that the tensorial\nextensions are superior to both their vector-valued counterparts and to two\nexisting tensorial dimension reduction methods for i.i.d. data. Finally,\napplications to fMRI-data and video processing show that the methods are\ncapable of extracting relevant information from noisy high-dimensional data.\n",
"title": "Blind source separation of tensor-valued time series"
} | null | null | null | null | true | null | 20516 | null | Default | null | null |
null | {
"abstract": " We introduce the notion of a $[z, r; g]$-mixed cage. A $[z, r; g]$-mixed cage\nis a mixed graph $G$, $z$-regular by arcs, $r$-regular by edges, with girth $g$\nand minimum order. In this paper we prove the existence of $[z, r ;g]$-mixed\ncages and exhibit families of mixed cages for some specific values. We also\ngive lower and upper bounds for some choices of $z, r$ and $g$. In particular\nwe present the first results on $[z,r;g]$- mixed cages for $z=1$ and any $r\\geq\n1$ and $g\\geq 3$, and for any $z\\geq 1$, $r=1$ and $g=4$.\n",
"title": "Mixed Cages"
} | null | null | null | null | true | null | 20517 | null | Default | null | null |
null | {
"abstract": " A technique to levitate and measure the three-dimensional position of\nmicrometer-sized dielectric spheres with heterodyne detection is presented. The\ntwo radial degrees of freedom are measured by interfering light transmitted\nthrough the microsphere with a reference wavefront, while the axial degree of\nfreedom is measured from the phase of the light reflected from the surface of\nthe microsphere. This method pairs the simplicity and accessibility of single\nbeam optical traps to a measurement of displacement that is intrinsically\ncalibrated by the wavelength of the trapping light and has exceptional immunity\nto stray light. A theoretical shot noise limit of\n$1.3\\times10^{-13}\\,\\text{m}/\\sqrt{\\text{Hz}}$ for the radial degrees of\nfreedom, and $3.0\\times10^{-15} \\, \\text{m}/\\sqrt{\\text{Hz}}$ for the axial\ndegree of freedom can be obtained in the system described. The measured\nacceleration noise in the radial direction is $7.5\\times10^{-5} \\,\n(\\text{m/s}^2)/\\sqrt{\\text{Hz}}$.\n",
"title": "Single-beam dielectric-microsphere trapping with optical heterodyne detection"
} | null | null | null | null | true | null | 20518 | null | Default | null | null |
null | {
"abstract": " Monero is a privacy-centric cryptocurrency that allows users to obscure their\ntransactions by including chaff coins, called \"mixins,\" along with the actual\ncoins they spend. In this paper, we empirically evaluate two weaknesses in\nMonero's mixin sampling strategy. First, about 62% of transaction inputs with\none or more mixins are vulnerable to \"chain-reaction\" analysis -- that is, the\nreal input can be deduced by elimination. Second, Monero mixins are sampled in\nsuch a way that they can be easily distinguished from the real coins by their\nage distribution; in short, the real input is usually the \"newest\" input. We\nestimate that this heuristic can be used to guess the real input with 80%\naccuracy over all transactions with 1 or more mixins. Next, we turn to the\nMonero ecosystem and study the importance of mining pools and the former\nanonymous marketplace AlphaBay on the transaction volume. We find that after\nremoving mining pool activity, there remains a large amount of potentially\nprivacy-sensitive transactions that are affected by these weaknesses. We\npropose and evaluate two countermeasures that can improve the privacy of future\ntransactions.\n",
"title": "An Empirical Analysis of Traceability in the Monero Blockchain"
} | null | null | null | null | true | null | 20519 | null | Default | null | null |
null | {
"abstract": " Face image quality can be defined as a measure of the utility of a face image\nto automatic face recognition. In this work, we propose (and compare) two\nmethods for automatic face image quality based on target face quality values\nfrom (i) human assessments of face image quality (matcher-independent), and\n(ii) quality values computed from similarity scores (matcher-dependent). A\nsupport vector regression model trained on face features extracted using a deep\nconvolutional neural network (ConvNet) is used to predict the quality of a face\nimage. The proposed methods are evaluated on two unconstrained face image\ndatabases, LFW and IJB-A, which both contain facial variations with multiple\nquality factors. Evaluation of the proposed automatic face image quality\nmeasures shows we are able to reduce the FNMR at 1% FMR by at least 13% for two\nface matchers (a COTS matcher and a ConvNet matcher) by using the proposed face\nquality to select subsets of face images and video frames for matching\ntemplates (i.e., multiple faces per subject) in the IJB-A protocol. To our\nknowledge, this is the first work to utilize human assessments of face image\nquality in designing a predictor of unconstrained face quality that is shown to\nbe effective in cross-database evaluation.\n",
"title": "Automatic Face Image Quality Prediction"
} | null | null | null | null | true | null | 20520 | null | Default | null | null |
null | {
"abstract": " Networks, which represent agents and interactions between them, arise in\nmyriad applications throughout the sciences, engineering, and even the\nhumanities. To understand large-scale structure in a network, a common task is\nto cluster a network's nodes into sets called \"communities\" such that there are\ndense connections within communities but sparse connections between them. A\npopular and statistically principled method to perform such clustering is to\nuse a family of generative models known as stochastic block models (SBMs). In\nthis paper, we show that maximum likelihood estimation in an SBM is a network\nanalog of a well-known continuum surface-tension problem that arises from an\napplication in metallurgy. To illustrate the utility of this bridge, we\nimplement network analogs of three surface-tension algorithms, with which we\nsuccessfully recover planted community structure in synthetic networks and\nwhich yield fascinating insights on empirical networks from the field of\nhyperspectral video segmentation.\n",
"title": "Stochastic Block Models are a Discrete Surface Tension"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20521 | null | Validated | null | null |
null | {
"abstract": " The pairing symmetry of the newly proposed cobalt high temperature\n(high-$T_c$) superconductors formed by vertex shared cation-anion tetrahedral\ncomplexes is studied by the methods of mean field, random phase approximation\n(RPA) and functional renormalization group (FRG) analysis. The results of all\nthese methods show that the $d_{x^2-y^2}$ pairing symmetry is robustly favored\nnear half filling. The RPA and FRG methods, which are valid in weak interaction\nregions, predict that the superconducting state is also strongly orbital\nselective, namely the $d_{x^2-y^2}$ orbital that has the largest density near\nhalf filling among the three $t_{2g}$ orbitals dominates superconducting\npairing. These results suggest that the new materials, if synthesized, can\nprovide indisputable test to high-$T_c$ pairing mechanism and the validity of\ndifferent theoretical methods.\n",
"title": "Robust d-wave pairing symmetry in multi-orbital cobalt high temperature superconductors"
} | null | null | null | null | true | null | 20522 | null | Default | null | null |
null | {
"abstract": " The search for flat-band solid-state realizations is a crucial issue to\nverify or to challenge theoretical predictions for quantum many-body flat-band\nsystems. For frustrated quantum magnets flat bands lead to various\nunconventional properties related to the existence of localized many-magnon\nstates. The recently synthesized magnetic compound Ba$_2$CoSi$_2$O$_6$Cl$_2$\nseems to be an almost perfect candidate to observe these features in\nexperiments. We develop a theory for Ba$_2$CoSi$_2$O$_6$Cl$_2$ by adapting the\nlocalized-magnon concept to this compound. We first show that our theory\ndescribes the known experimental facts and then we propose new experimental\nstudies to detect a field-driven phase transition related to a\nWigner-crystal-like ordering of localized magnons at low temperatures.\n",
"title": "Thermodynamic properties of Ba$_2$CoSi$_2$O$_6$Cl$_2$ in strong magnetic field: Realization of flat-band physics in a highly frustrated quantum magnet"
} | null | null | null | null | true | null | 20523 | null | Default | null | null |
null | {
"abstract": " A deep neural network is a hierarchical nonlinear model transforming input\nsignals to output signals. Its input-output relation is considered to be\nstochastic, being described for a given input by a parameterized conditional\nprobability distribution of outputs. The space of parameters consisting of\nweights and biases is a Riemannian manifold, where the metric is defined by the\nFisher information matrix. The natural gradient method uses the steepest\ndescent direction in a Riemannian manifold, so it is effective in learning,\navoiding plateaus. It requires inversion of the Fisher information matrix,\nhowever, which is practically impossible when the matrix has a huge number of\ndimensions. Many methods for approximating the natural gradient have therefore\nbeen introduced. The present paper uses statistical neurodynamical method to\nreveal the properties of the Fisher information matrix in a net of random\nconnections under the mean field approximation. We prove that the Fisher\ninformation matrix is unit-wise block diagonal supplemented by small order\nterms of off-block-diagonal elements, which provides a justification for the\nquasi-diagonal natural gradient method by Y. Ollivier. A unitwise\nblock-diagonal Fisher metrix reduces to the tensor product of the Fisher\ninformation matrices of single units. We further prove that the Fisher\ninformation matrix of a single unit has a simple reduced form, a sum of a\ndiagonal matrix and a rank 2 matrix of weight-bias correlations. We obtain the\ninverse of Fisher information explicitly. We then have an explicit form of the\nnatural gradient, without relying on the numerical matrix inversion, which\ndrastically speeds up stochastic gradient learning.\n",
"title": "Fisher Information and Natural Gradient Learning of Random Deep Networks"
} | null | null | null | null | true | null | 20524 | null | Default | null | null |
null | {
"abstract": " We identify and describe the main dynamic regimes occurring during the\nmelting of the PCM n-octadecane in horizontal layers of several sizes heated\nfrom below. This configuration allows to cover a wide range of effective\nRayleigh numbers on the liquid PCM phase, up to $\\sim 10^9$, without changing\nany external parameter control. We identify four different regimes as time\nevolves: (i) the conductive regime, (ii) linear regime, (iii) coarsening regime\nand (iv) turbulent regime. The first two regimes appear at all domain sizes.\nHowever the third and fourth regime require a minimum advance of the\nsolid/liquid interface to develop, and we observe them only for large enough\ndomains.\nThe transition to turbulence takes places after a secondary instability that\nforces the coarsening of the thermal plumes. Each one of the melting regimes\ncreates a distinct solid/liquid front that characterizes the internal state of\nthe melting process. We observe that most of the magnitudes of the melting\nprocess are ruled by power laws, although not all of them. Thus the number of\nplumes, some regimes of the Rayleigh number as a function of time, the number\nof plumes after the primary and secondary instability, the thermal and kinetic\nboundary layers follow simple power laws. In particular, we find that the\nNusselt number scales with the Rayleigh number as $Nu \\sim Ra^{0.29}$ in the\nturbulent regime, consistent with theories and experiments on Rayleigh-Bénard\nconvection that show an exponent $2/7$.\n",
"title": "Dynamic of plumes and scaling during the melting of a Phase Change Material heated from below"
} | null | null | null | null | true | null | 20525 | null | Default | null | null |
null | {
"abstract": " We survey different classification results for surfaces with parallel mean\ncurvature immersed into some Riemannian homogeneous four-manifolds, including\nreal and complex space forms, and product spaces. We provide a common framework\nfor this problem, with special attention to the existence of holomorphic\nquadratic differentials on such surfaces. The case of spheres with parallel\nmean curvature is also explained in detail, as well as the state-of-the-art\nadvances in the general problem.\n",
"title": "Parallel mean curvature surfaces in four-dimensional homogeneous spaces"
} | null | null | null | null | true | null | 20526 | null | Default | null | null |
null | {
"abstract": " Since introduction [A. Knyazev, Toward the optimal preconditioned\neigensolver: Locally optimal block preconditioned conjugate gradient method,\nSISC (2001) DOI:10.1137/S1064827500366124] and efficient parallel\nimplementation [A. Knyazev et al., Block locally optimal preconditioned\neigenvalue xolvers (BLOPEX) in HYPRE and PETSc, SISC (2007)\nDOI:10.1137/060661624], LOBPCG has been used is a wide range of applications in\nmechanics, material sciences, and data sciences. We review its recent\nimplementations and applications, as well as extensions of the local optimality\nidea beyond standard eigenvalue problems.\n",
"title": "Recent implementations, applications, and extensions of the Locally Optimal Block Preconditioned Conjugate Gradient method (LOBPCG)"
} | null | null | null | null | true | null | 20527 | null | Default | null | null |
null | {
"abstract": " We compare a large suite of theoretical cosmological models to observational\ndata from the cosmic microwave background, baryon acoustic oscillation\nmeasurements of expansion, Type Ia SNe measurements of expansion, redshift\nspace distortion measurements of the growth of structure, and the local Hubble\nconstant. Our theoretical models include parametrizations of dark energy as\nwell as physical models of dark energy and modified gravity. We determine the\nconstraints on the model parameters, incorporating the redshift space\ndistortion data directly in the analysis. To determine whether models can be\nruled out, we evaluate the $p$ value (the probability under the model of\nobtaining data as bad or worse than the observed data). In our comparison, we\nfind the well known tension of H$_0$ with the other data; no model resolves\nthis tension successfully. Among the models we consider, the large scale growth\nof structure data does not affect the modified gravity models as a category\nparticularly differently than dark energy models; it matters for some modified\ngravity models but not others, and the same is true for dark energy models. We\ncompute predicted observables for each model under current observational\nconstraints, and identify models for which future observational constraints\nwill be particularly informative.\n",
"title": "An evaluation of cosmological models from expansion and growth of structure measurements"
} | null | null | [
"Physics"
]
| null | true | null | 20528 | null | Validated | null | null |
null | {
"abstract": " The work is devoted to the variety of $2$-dimensional algebras over an\nalgebraically closed field. Firstly, we classify such algebras modulo\nisomorphism. Then we describe the degenerations and the closures of principal\nalgebra series in the variety under consideration. Finally, we apply our\nresults to obtain analogous descriptions for the subvarieties of flexible, and\nbicommutative algebras. In particular, we describe rigid algebras and\nirreducible components for these subvarieties.\n",
"title": "The variety of $2$-dimensional algebras over an algebraically closed field"
} | null | null | null | null | true | null | 20529 | null | Default | null | null |
null | {
"abstract": " We present a Bayesian object observation model for complete probabilistic\nsemantic SLAM. Recent studies on object detection and feature extraction have\nbecome important for scene understanding and 3D mapping. However, 3D shape of\nthe object is too complex to formulate the probabilistic observation model;\ntherefore, performing the Bayesian inference of the object-oriented features as\nwell as their pose is less considered. Besides, when the robot equipped with an\nRGB mono camera only observes the projected single view of an object, a\nsignificant amount of the 3D shape information is abandoned. Due to these\nlimitations, semantic SLAM and viewpoint-independent loop closure using\nvolumetric 3D object shape is challenging. In order to enable the complete\nformulation of probabilistic semantic SLAM, we approximate the observation\nmodel of a 3D object with a tractable distribution. We also estimate the\nvariational likelihood from the 2D image of the object to exploit its observed\nsingle view. In order to evaluate the proposed method, we perform pose and\nfeature estimation, and demonstrate that the automatic loop closure works\nseamlessly without additional loop detector in various environments.\n",
"title": "A Variational Observation Model of 3D Object for Probabilistic Semantic SLAM"
} | null | null | null | null | true | null | 20530 | null | Default | null | null |
null | {
"abstract": " We investigate the generation of optical frequency combs through a cascade of\nfour-wave mixing processes in nonlinear fibres with optimised parameters. The\ninitial optical field consists of two continuous-wave lasers with frequency\nseparation larger than 40 GHz (312.7 pm at 1531 nm). It propagates through\nthree nonlinear fibres. The first fibre serves to pulse shape the initial\nsinusoidal-square pulse, while a strong pulse compression down to sub-100 fs\ntakes place in the second fibre which is an amplifying erbium-doped fibre. The\nlast stage is a low-dispersion highly nonlinear fibre where the frequency comb\nbandwidth is increased and the line intensity is equalised. We model this\nsystem using the generalised nonlinear Schrödinger equation and investigate\nit in terms of fibre lengths, fibre dispersion, laser frequency separation and\ninput powers with the aim to minimise the frequency comb noise. With the\nsupport of the numerical results, a frequency comb is experimentally generated,\nfirst in the near infra-red and then it is frequency-doubled into the visible\nspectral range. Using a MUSE-type spectrograph, we evaluate the comb\nperformance for astronomical wavelength calibration in terms of equidistancy of\nthe comb lines and their stability.\n",
"title": "Generation of optical frequency combs via four-wave mixing processes for low- and medium-resolution astronomy"
} | null | null | [
"Physics"
]
| null | true | null | 20531 | null | Validated | null | null |
null | {
"abstract": " Physical media (like surveillance cameras) and social media (like Instagram\nand Twitter) may both be useful in attaining on-the-ground information during\nan emergency or disaster situation. However, the intersection and reliability\nof both surveillance cameras and social media during a natural disaster are not\nfully understood. To address this gap, we tested whether social media is of\nutility when physical surveillance cameras went off-line during Hurricane Irma\nin 2017. Specifically, we collected and compared geo-tagged Instagram and\nTwitter posts in the state of Florida during times and in areas where public\nsurveillance cameras went off-line. We report social media content and\nfrequency and content to determine the utility for emergency managers or first\nresponders during a natural disaster.\n",
"title": "Cross-referencing Social Media and Public Surveillance Camera Data for Disaster Response"
} | null | null | [
"Computer Science"
]
| null | true | null | 20532 | null | Validated | null | null |
null | {
"abstract": " A physical model of a three-dimensional flow of a viscous bubbly fluid in an\nintermediate regime between bubble formation and breakage is presented. The\nmodel is based on mechanics and thermodynamics of a single bubble coupled to\nthe dynamics of a viscous fluid as a whole, and takes into account multiple\nphysical effects, including gravity, viscosity, and surface tension.\nDimensionless versions of the resulting nonlinear model are obtained, and\nvalues of dimensionless parameters are estimated for typical magma flows in\nhorizontal subaerial lava fields and vertical volcanic conduits.\nExact solutions of the resulting system of nonlinear equations corresponding\nto equilibrium flows and traveling waves are analyzed in the one-dimensional\nsetting. Generalized Su-Gardner-type perturbation analysis is employed to study\napproximate solutions of the model in the long-wave ansatz. Simplified\nnonlinear partial differential equations (PDE) satisfied by the leading terms\nof the perturbation solutions are systematically derived. It is shown that for\nspecific classes of perturbations, approximate solutions of the bubbly fluid\nmodel arise from solutions of the classical diffusion, Burgers,\nvariable-coefficient Burgers, and Korteweg-de Vries equations.\n",
"title": "Nonlinear Dynamics of a Viscous Bubbly Fluid"
} | null | null | null | null | true | null | 20533 | null | Default | null | null |
null | {
"abstract": " Localization of anatomical structures is a prerequisite for many tasks in\nmedical image analysis. We propose a method for automatic localization of one\nor more anatomical structures in 3D medical images through detection of their\npresence in 2D image slices using a convolutional neural network (ConvNet).\nA single ConvNet is trained to detect presence of the anatomical structure of\ninterest in axial, coronal, and sagittal slices extracted from a 3D image. To\nallow the ConvNet to analyze slices of different sizes, spatial pyramid pooling\nis applied. After detection, 3D bounding boxes are created by combining the\noutput of the ConvNet in all slices.\nIn the experiments 200 chest CT, 100 cardiac CT angiography (CTA), and 100\nabdomen CT scans were used. The heart, ascending aorta, aortic arch, and\ndescending aorta were localized in chest CT scans, the left cardiac ventricle\nin cardiac CTA scans, and the liver in abdomen CT scans. Localization was\nevaluated using the distances between automatically and manually defined\nreference bounding box centroids and walls.\nThe best results were achieved in localization of structures with clearly\ndefined boundaries (e.g. aortic arch) and the worst when the structure boundary\nwas not clearly visible (e.g. liver). The method was more robust and accurate\nin localization multiple structures.\n",
"title": "ConvNet-Based Localization of Anatomical Structures in 3D Medical Images"
} | null | null | null | null | true | null | 20534 | null | Default | null | null |
null | {
"abstract": " Let $(X,d,\\mu)$ be a doubling metric measure space endowed with a Dirichlet\nform $\\E$ deriving from a \"carré du champ\". Assume that $(X,d,\\mu,\\E)$\nsupports a scale-invariant $L^2$-Poincaré inequality. In this article, we\nstudy the following properties of harmonic functions, heat kernels and Riesz\ntransforms for $p\\in (2,\\infty]$:\n(i) $(G_p)$: $L^p$-estimate for the gradient of the associated heat\nsemigroup;\n(ii) $(RH_p)$: $L^p$-reverse Hölder inequality for the gradients of\nharmonic functions;\n(iii) $(R_p)$: $L^p$-boundedness of the Riesz transform ($p<\\infty$);\n(iv) $(GBE)$: a generalised Bakry-Émery condition.\nWe show that, for $p\\in (2,\\infty)$, (i), (ii) (iii) are equivalent, while\nfor $p=\\infty$, (i), (ii), (iv) are equivalent.\nMoreover, some of these equivalences still hold under weaker conditions than\nthe $L^2$-Poincaré inequality.\nOur result gives a characterisation of Li-Yau's gradient estimate of heat\nkernels for $p=\\infty$, while for $p\\in (2,\\infty)$ it is a substantial\nimprovement as well as a generalisation of earlier results by\nAuscher-Coulhon-Duong-Hofmann [7] and Auscher-Coulhon [6]. Applications to\nisoperimetric inequalities and Sobolev inequalities are given. Our results\napply to Riemannian and sub-Riemannian manifolds as well as to non-smooth\nspaces, and to degenerate elliptic/parabolic equations in these settings.\n",
"title": "Gradient estimates for heat kernels and harmonic functions"
} | null | null | null | null | true | null | 20535 | null | Default | null | null |
null | {
"abstract": " The last decade was remarkable for neutrino physics. In particular, the\nphenomenon of neutrino flavor oscillations has been firmly established by a\nseries of independent measurements. All parameters of the neutrino mixing are\nnow known and we have elements to plan a judicious exploration of new scenarios\nthat are opened by these recent advances. With precise measurements, we can\ntest the 3-neutrino paradigm, neutrino mass hierarchy and CP asymmetry in the\nlepton sector. The future long-baseline experiments are considered to be a\nfundamental tool to deepen our knowledge of electroweak interactions. The Deep\nUnderground Neutrino Experiment -- DUNE will detect a broad-band neutrino beam\nfrom Fermilab in an underground massive Liquid Argon Time-Projection Chamber at\nan L/E of about $10^3$ km / GeV to reach good sensitivity for CP-phase\nmeasurements and the determination of the mass hierarchy. The dimensions and\nthe depth of the Far Detector also create an excellent opportunity to look for\nrare signals like proton decay to study violation of baryonic number, as well\nas supernova neutrino bursts, broadening the scope of the experiment to\nastrophysics and associated impacts in cosmology. In this presentation, we will\ndiscuss the physics motivations and the main experimental features of the DUNE\nproject required to reach its scientific goals.\n",
"title": "The Deep Underground Neutrino Experiment -- DUNE: the precision era of neutrino physics"
} | null | null | null | null | true | null | 20536 | null | Default | null | null |
null | {
"abstract": " We study several variants of q-Garnier system corresponding to various\ndirections of discrete time evolutions. We also investigate a relation between\nthe $q$-Garnier system and Suzuki's higher order $q$-Painlev/'e system by using\na duality of the $q$-KP system.\n",
"title": "Variations of $q$-Garnier system"
} | null | null | null | null | true | null | 20537 | null | Default | null | null |
null | {
"abstract": " We investigate, in the Luttinger model with fixed box potential, the time\nevolution of an inhomogeneous state prepared as a localized fermion added to\nthe noninteracting ground state. We proved that, if the state is evolved with\nthe interacting Hamiltonian, the averaged density has two peaks moving in\nopposite directions, with a constant but renormalized velocity. We also proved\nthat a dynamical `Landau quasi-particle weight' appears in the oscillating part\nof the averaged density, asymptotically vanishing with large time. The results\nare proved with the Mattis-Lieb diagonalization method. A simpler proof with\nthe exact Bosonization formulas is also provided.\n",
"title": "Quantum Quench dynamics in Non-local Luttinger Model: Rigorous Results"
} | null | null | [
"Mathematics"
]
| null | true | null | 20538 | null | Validated | null | null |
null | {
"abstract": " The cosmological relaxation of the electroweak scale has been proposed as a\nmechanism to address the hierarchy problem of the Standard Model. A field, the\nrelaxion, rolls down its potential and, in doing so, scans the squared mass\nparameter of the Higgs, relaxing it to a parametrically small value. In this\nwork, we promote the relaxion to an inflaton. We couple it to Abelian gauge\nbosons, thereby introducing the necessary dissipation mechanism which slows\ndown the field in the last stages. We describe a novel reheating mechanism,\nwhich relies on the gauge-boson production leading to strong electromagnetic\nfields, and proceeds via the vacuum production of electron-positron pairs\nthrough the Schwinger effect. We refer to this mechanism as Schwinger\nreheating. We discuss the cosmological dynamics of the model and the\nphenomenological constraints from CMB and other experiments. We find that a\ncutoff close to the Planck scale may be achieved. In its minimal form, the\nmodel does not generate sufficient curvature perturbations and additional\ningredients, such as a curvaton field, are needed.\n",
"title": "Dynamics of Relaxed Inflation"
} | null | null | [
"Physics"
]
| null | true | null | 20539 | null | Validated | null | null |
null | {
"abstract": " In solving hard computational problems, semidefinite program (SDP)\nrelaxations often play an important role because they come with a guarantee of\noptimality. Here, we focus on a popular semidefinite relaxation of K-means\nclustering which yields the same solution as the non-convex original\nformulation for well segregated datasets. We report an unexpected finding: when\ndata contains (greater than zero-dimensional) manifolds, the SDP solution\ncaptures such geometrical structures. Unlike traditional manifold embedding\ntechniques, our approach does not rely on manually defining a kernel but rather\nenforces locality via a nonnegativity constraint. We thus call our approach\nNOnnegative MAnifold Disentangling, or NOMAD. To build an intuitive\nunderstanding of its manifold learning capabilities, we develop a theoretical\nanalysis of NOMAD on idealized datasets. While NOMAD is convex and the globally\noptimal solution can be found by generic SDP solvers with polynomial time\ncomplexity, they are too slow for modern datasets. To address this problem, we\nanalyze a non-convex heuristic and present a new, convex and yet efficient,\nalgorithm, based on the conditional gradient method. Our results render NOMAD a\nversatile, understandable, and powerful tool for manifold learning.\n",
"title": "Clustering is semidefinitely not that hard: Nonnegative SDP for manifold disentangling"
} | null | null | [
"Computer Science"
]
| null | true | null | 20540 | null | Validated | null | null |
null | {
"abstract": " Magnetically tunable Feshbach resonances in ultracold atomic systems are\nchiefly identified and characterized through time consuming atom loss\nspectroscopy. We describe an off-resonant dispersive optical probing technique\nto rapidly locate Feshbach resonances and demonstrate the method by locating\nfour resonances of $^{87}$Rb, between the $|\\rm{F} = 1, \\rm{m_F}=1 \\rangle$ and\n$|\\rm{F} = 2, \\rm{m_F}=0 \\rangle$ states. Despite the loss features being\n$\\lesssim0.1$ G wide, we require only 21 experimental runs to explore a\nmagnetic field range >18 G, where $1~\\rm{G}=10^{-4}$ T. The resonances consist\nof two known s-wave features in the vicinity of 9 G and 18 G and two previously\nunobserved p-wave features near 5 G and 10 G. We further utilize the dispersive\napproach to directly characterize the two-body loss dynamics for each Feshbach\nresonance.\n",
"title": "Dispersive optical detection of magnetic Feshbach resonances in ultracold gases"
} | null | null | null | null | true | null | 20541 | null | Default | null | null |
null | {
"abstract": " In Compressed Sensing, a real-valued sparse vector has to be recovered from\nan underdetermined system of linear equations. In many applications, however,\nthe elements of the sparse vector are drawn from a finite set. Adapted\nalgorithms incorporating this additional knowledge are required for the\ndiscrete-valued setup. In this paper, turbo-based algorithms for both cases are\nelucidated and analyzed from a communications engineering perspective, leading\nto a deeper understanding of the algorithm. In particular, we gain the\nintriguing insight that the calculation of extrinsic values is equal to the\nunbiasing of a biased estimate and present an improved algorithm.\n",
"title": "Unveiling Bias Compensation in Turbo-Based Algorithms for (Discrete) Compressed Sensing"
} | null | null | null | null | true | null | 20542 | null | Default | null | null |
null | {
"abstract": " Momentum methods such as Polyak's heavy ball (HB) method, Nesterov's\naccelerated gradient (AG) as well as accelerated projected gradient (APG)\nmethod have been commonly used in machine learning practice, but their\nperformance is quite sensitive to noise in the gradients. We study these\nmethods under a first-order stochastic oracle model where noisy estimates of\nthe gradients are available. For strongly convex problems, we show that the\ndistribution of the iterates of AG converges with the accelerated\n$O(\\sqrt{\\kappa}\\log(1/\\varepsilon))$ linear rate to a ball of radius\n$\\varepsilon$ centered at a unique invariant distribution in the 1-Wasserstein\nmetric where $\\kappa$ is the condition number as long as the noise variance is\nsmaller than an explicit upper bound we can provide. Our analysis also\ncertifies linear convergence rates as a function of the stepsize, momentum\nparameter and the noise variance; recovering the accelerated rates in the\nnoiseless case and quantifying the level of noise that can be tolerated to\nachieve a given performance. In the special case of strongly convex quadratic\nobjectives, we can show accelerated linear rates in the $p$-Wasserstein metric\nfor any $p\\geq 1$ with improved sensitivity to noise for both AG and HB through\na non-asymptotic analysis under some additional assumptions on the noise\nstructure. Our analysis for HB and AG also leads to improved non-asymptotic\nconvergence bounds in suboptimality for both deterministic and stochastic\nsettings which is of independent interest. To the best of our knowledge, these\nare the first linear convergence results for stochastic momentum methods under\nthe stochastic oracle model. We also extend our results to the APG method and\nweakly convex functions showing accelerated rates when the noise magnitude is\nsufficiently small.\n",
"title": "Accelerated Linear Convergence of Stochastic Momentum Methods in Wasserstein Distances"
} | null | null | null | null | true | null | 20543 | null | Default | null | null |
null | {
"abstract": " The recovery of approximately sparse or compressible coefficients in a\nPolynomial Chaos Expansion is a common goal in modern parametric uncertainty\nquantification (UQ). However, relatively little effort in UQ has been directed\ntoward theoretical and computational strategies for addressing the sparse\ncorruptions problem, where a small number of measurements are highly corrupted.\nSuch a situation has become pertinent today since modern computational\nframeworks are sufficiently complex with many interdependent components that\nmay introduce hardware and software failures, some of which can be difficult to\ndetect and result in a highly polluted simulation result.\nIn this paper we present a novel compressive sampling-based theoretical\nanalysis for a regularized $\\ell^1$ minimization algorithm that aims to recover\nsparse expansion coefficients in the presence of measurement corruptions. Our\nrecovery results are uniform, and prescribe algorithmic regularization\nparameters in terms of a user-defined a priori estimate on the ratio of\nmeasurements that are believed to be corrupted. We also propose an iteratively\nreweighted optimization algorithm that automatically refines the value of the\nregularization parameter, and empirically produces superior results. Our\nnumerical results test our framework on several medium-to-high dimensional\nexamples of solutions to parameterized differential equations, and demonstrate\nthe effectiveness of our approach.\n",
"title": "Compressed sensing with sparse corruptions: Fault-tolerant sparse collocation approximations"
} | null | null | null | null | true | null | 20544 | null | Default | null | null |
null | {
"abstract": " Despite recent advances, memory-augmented deep neural networks are still\nlimited when it comes to life-long and one-shot learning, especially in\nremembering rare events. We present a large-scale life-long memory module for\nuse in deep learning. The module exploits fast nearest-neighbor algorithms for\nefficiency and thus scales to large memory sizes. Except for the\nnearest-neighbor query, the module is fully differentiable and trained\nend-to-end with no extra supervision. It operates in a life-long manner, i.e.,\nwithout the need to reset it during training.\nOur memory module can be easily added to any part of a supervised neural\nnetwork. To show its versatility we add it to a number of networks, from simple\nconvolutional ones tested on image classification to deep sequence-to-sequence\nand recurrent-convolutional models. In all cases, the enhanced network gains\nthe ability to remember and do life-long one-shot learning. Our module\nremembers training examples shown many thousands of steps in the past and it\ncan successfully generalize from them. We set new state-of-the-art for one-shot\nlearning on the Omniglot dataset and demonstrate, for the first time, life-long\none-shot learning in recurrent neural networks on a large-scale machine\ntranslation task.\n",
"title": "Learning to Remember Rare Events"
} | null | null | [
"Computer Science"
]
| null | true | null | 20545 | null | Validated | null | null |
null | {
"abstract": " The crystallographic stacking order in multilayer graphene plays an important\nrole in determining its electronic structure. In trilayer graphene,\nrhombohedral stacking (ABC) is particularly intriguing, exhibiting a flat band\nwith an electric-field tunable band gap. Such electronic structure is distinct\nfrom simple hexagonal stacking (AAA) or typical Bernal stacking (ABA), and is\npromising for nanoscale electronics, optoelectronics applications. So far clean\nexperimental electronic spectra on the first two stackings are missing because\nthe samples are usually too small in size (um or nm scale) to be resolved by\nconventional angle-resolved photoemission spectroscopy (ARPES). Here by using\nARPES with nanospot beam size (NanoARPES), we provide direct experimental\nevidence for the coexistence of three different stackings of trilayer graphene\nand reveal their distinctive electronic structures directly. By fitting the\nexperimental data, we provide important experimental band parameters for\ndescribing the electronic structure of trilayer graphene with different\nstackings.\n",
"title": "Stacking-dependent electronic structure of trilayer graphene resolved by nanospot angle-resolved photoemission spectroscopy"
} | null | null | null | null | true | null | 20546 | null | Default | null | null |
null | {
"abstract": " We study a variational Ginzburg-Landau type model depending on a small\nparameter $\\epsilon>0$ for (tangent) vector fields on a $2$-dimensional\nRiemannian surface. As $\\epsilon\\to 0$, the vector fields tend to be of unit\nlength and will have singular points of a (non-zero) index, called vortices.\nOur main result determines the interaction energy between these vortices as a\n$\\Gamma$-limit (at the second order) as $\\epsilon\\to 0$.\n",
"title": "Interaction energy between vortices of vector fields on Riemannian surfaces"
} | null | null | null | null | true | null | 20547 | null | Default | null | null |
null | {
"abstract": " We present a multimodal non-linear optical (NLO) laser-scanning microscope,\nbased on a compact fiber-format excitation laser and integrating coherent\nanti-Stokes Raman scattering (CARS), stimulated Raman scattering (SRS) and\ntwo-photon-excitation fluorescence (TPEF) on a single platform. We demonstrate\nits capabilities in simultaneously acquiring CARS and SRS images of a blend of\n6-{\\mu}m poly(methyl methacrylate) beads and 3-{\\mu}m polystyrene beads. We\nthen apply it to visualize cell walls and chloroplast of an unprocessed fresh\nleaf of Elodea aquatic plant via SRS and TPEF modalities, respectively. The\npresented NLO microscope, developed in house using off-the-shelf components,\noffers full accessibility to the optical path and ensures its easy\nre-configurability and flexibility.\n",
"title": "Multimodal Nonlinear Microscope based on a Compact Fiber-format Laser Source"
} | null | null | null | null | true | null | 20548 | null | Default | null | null |
null | {
"abstract": " In 1965, the discovery of a new type of uniform radiation, located between\nradiowaves and infrared light, was accidental. Known today as Cosmic Microwave\nbackground (CMB), this diffuse radiation is commonly interpreted as a fossil\nlight released in an early hot and dense universe and constitutes today the\nmain 'pilar' of the big bang cosmology. Considerable efforts have been devoted\nto derive fundamental cosmological parameters from the characteristics of this\nradiation that led to a surprising universe that is shaped by at least three\nmajor unknown components: inflation, dark matter and dark energy. This is an\nimportant weakness of the present consensus cosmological model that justifies\nraising several questions on the CMB interpretation. Can we consider its\ncosmological nature as undisputable? Do other possible interpretations exist in\nthe context of other cosmological theories or simply as a result of other\nphysical mechanisms that could account for it? In an effort to questioning the\nvalidity of scientific hypotheses and the under-determination of theories\ncompared to observations, we examine here the difficulties that still exist on\nthe interpretation of this diffuse radiation and explore other proposed tracks\nto explain its origin. We discuss previous historical concepts of diffuse\nradiation before and after the CMB discovery and underline the limit of our\npresent understanding.\n",
"title": "The Diffuse Light of the Universe - On the microwave background before and after its discovery: open questions"
} | null | null | null | null | true | null | 20549 | null | Default | null | null |
null | {
"abstract": " Leibniz algebras are certain generalization of Lie algebras. In this paper we\ngive the classification of $5-$dimensional complex non-Lie nilpotent Leibniz\nalgebras. We use the canonical forms for the congruence classes of matrices of\nbilinear forms to classify the case $\\dim(A^2)=3$ and $\\dim(Leib(A))=1$ which\ncan be applied to higher dimensions. The remaining cases are classified via\ndirect method.\n",
"title": "Classification of $5$-Dimensional Complex Nilpotent Leibniz Algebras"
} | null | null | null | null | true | null | 20550 | null | Default | null | null |
null | {
"abstract": " Reconstruction of skilled humans sensation and control system often leads to\na development of robust control for the robots. We are developing an unscrewing\nrobot for the automated disassembly which requires a comprehensive control\nsystem, but unscrewing experiments with robots are often limited to several\nconditions. On the contrary, humans typically have a broad range of screwing\nexperiences and sensations throughout their lives, and we conducted an\nexperiment to find these haptic patterns. Results show that people apply axial\nforce to the screws to avoid screwdriver slippage (cam-outs), which is one of\nthe key problems during screwing and unscrewing, and this axial force is\nproportional to the torque which is required for screwing. We have found that\ntype of the screw head influences the amount of axial force applied. Using this\nknowledge an unscrewing robot for the smart disassembly factory RecyBot is\ndeveloped, and experiments confirm the optimality of the strategy, used by\nhumans. Finally, a methodology for robust unscrewing algorithm design is\npresented as a generalization of the findings. It can seriously speed up the\ndevelopment of the screwing and unscrewing robots and tools.\n",
"title": "Haptics of Screwing and Unscrewing for its Application in Smart Factories for Disassembly"
} | null | null | null | null | true | null | 20551 | null | Default | null | null |
null | {
"abstract": " Let $(X, T^{1,0}X)$ be a compact connected orientable CR manifold of\ndimension $2n+1$ with non-degenerate Levi curvature. Assume that $X$ admits a\nconnected compact Lie group action $G$. Under certain natural assumptions about\nthe group action $G$, we show that the $G$-invariant Szegö kernel for $(0,q)$\nforms is a complex Fourier integral operator, smoothing away $\\mu^{-1}(0)$ and\nthere is a precise description of the singularity near $\\mu^{-1}(0)$, where\n$\\mu$ denotes the CR moment map. We apply our result to the case when $X$\nadmits a transversal CR $S^1$ action and deduce an asymptotic expansion for the\n$m$-th Fourier component of the $G$-invariant Szegö kernel for $(0,q)$ forms\nas $m \\to+\\infty$. As an application, we show that if $m$ large enough,\nquantization commutes with reduction.\n",
"title": "$G$-invariant Szegö kernel asymptotics and CR reduction"
} | null | null | null | null | true | null | 20552 | null | Default | null | null |
null | {
"abstract": " In the present work we explore resistive circuits where the individual\nresistors are arranged in fractal-like patterns. These circuits have some of\nthe characteristics typically found in geometric fractals, namely\nself-similarity and scale invariance. Considering resistive circuits as graphs,\nwe propose a definition of self-similar circuits which mimics a self-similar\nfractal. General properties of the resistive circuits generated by this\napproach are investigated, and interesting examples are commented in detail.\nSpecifically, we consider self-similar resistive series, tree-like resistive\nnetworks and Sierpinski's configurations with resistors.\n",
"title": "Self-similar resistive circuits as fractal-like structures"
} | null | null | null | null | true | null | 20553 | null | Default | null | null |
null | {
"abstract": " We have developed a computational code, DynaPhoPy, that allow us to extract\nthe microscopic anharmonic phonon properties from molecular dynamics (MD)\nsimulations using the normal-mode-decomposition technique as presented by Sun\net al. [T. Sun, D. Zhang, R. Wentzcovitch, 2014]. Using this code we calculated\nthe quasiparticle phonon frequencies and linewidths of crystalline silicon at\ndifferent temperatures using both of first-principles and the Tersoff empirical\npotential approaches. In this work we show the dependence of these properties\non the temperature using both approaches and compare them with reported\nexperimental data obtained by Raman spectroscopy [M. Balkanski, R. Wallis, E.\nHaro, 1983 and R. Tsu, J. G. Hernandez, 1982].\n",
"title": "DynaPhoPy: A code for extracting phonon quasiparticles from molecular dynamics simulations"
} | null | null | null | null | true | null | 20554 | null | Default | null | null |
null | {
"abstract": " The change detection problem is to determine if the Markov network structures\nof two Markov random fields differ from one another given two sets of samples\ndrawn from the respective underlying distributions. We study the trade-off\nbetween the sample sizes and the reliability of change detection, measured as a\nminimax risk, for the important cases of the Ising models and the Gaussian\nMarkov random fields restricted to the models which have network structures\nwith $p$ nodes and degree at most $d$, and obtain information-theoretic lower\nbounds for reliable change detection over these models. We show that for the\nIsing model, $\\Omega\\left(\\frac{d^2}{(\\log d)^2}\\log p\\right)$ samples are\nrequired from each dataset to detect even the sparsest possible changes, and\nthat for the Gaussian, $\\Omega\\left( \\gamma^{-2} \\log(p)\\right)$ samples are\nrequired from each dataset to detect change, where $\\gamma$ is the smallest\nratio of off-diagonal to diagonal terms in the precision matrices of the\ndistributions. These bounds are compared to the corresponding results in\nstructure learning, and closely match them under mild conditions on the model\nparameters. Thus, our change detection bounds inherit partial tightness from\nthe structure learning schemes in previous literature, demonstrating that in\ncertain parameter regimes, the naive structure learning based approach to\nchange detection is minimax optimal up to constant factors.\n",
"title": "Lower Bounds for Two-Sample Structural Change Detection in Ising and Gaussian Models"
} | null | null | null | null | true | null | 20555 | null | Default | null | null |
null | {
"abstract": " Context: Poor usability of cryptographic APIs is a severe source of\nvulnerabilities. Aim: We wanted to find out what kind of cryptographic\nlibraries are present in Rust and how usable they are. Method: We explored\nRust's cryptographic libraries through a systematic search, conducted an\nexploratory study on the major libraries and a controlled experiment on two of\nthese libraries with 28 student participants. Results: Only half of the major\nlibraries explicitly focus on usability and misuse resistance, which is\nreflected in their current APIs. We found that participants were more\nsuccessful using rust-crypto which we considered less usable than ring before\nthe experiment. Conclusion: We discuss API design insights and make\nrecommendations for the design of crypto libraries in Rust regarding the detail\nand structure of the documentation, higher-level APIs as wrappers for the\nexisting low-level libraries, and selected, good-quality example code to\nimprove the emerging cryptographic libraries of Rust.\n",
"title": "How Usable are Rust Cryptography APIs?"
} | null | null | null | null | true | null | 20556 | null | Default | null | null |
null | {
"abstract": " Building dialog agents that can converse naturally with humans is a\nchallenging yet intriguing problem of artificial intelligence. In open-domain\nhuman-computer conversation, where the conversational agent is expected to\nrespond to human responses in an interesting and engaging way, commonsense\nknowledge has to be integrated into the model effectively. In this paper, we\ninvestigate the impact of providing commonsense knowledge about the concepts\ncovered in the dialog. Our model represents the first attempt to integrating a\nlarge commonsense knowledge base into end-to-end conversational models. In the\nretrieval-based scenario, we propose the Tri-LSTM model to jointly take into\naccount message and commonsense for selecting an appropriate response. Our\nexperiments suggest that the knowledge-augmented models are superior to their\nknowledge-free counterparts in automatic evaluation.\n",
"title": "Augmenting End-to-End Dialog Systems with Commonsense Knowledge"
} | null | null | null | null | true | null | 20557 | null | Default | null | null |
null | {
"abstract": " We propose an alternative framework to existing setups for controlling false\nalarms when multiple A/B tests are run over time. This setup arises in many\npractical applications, e.g. when pharmaceutical companies test new treatment\noptions against control pills for different diseases, or when internet\ncompanies test their default webpages versus various alternatives over time.\nOur framework proposes to replace a sequence of A/B tests by a sequence of\nbest-arm MAB instances, which can be continuously monitored by the data\nscientist. When interleaving the MAB tests with an an online false discovery\nrate (FDR) algorithm, we can obtain the best of both worlds: low sample\ncomplexity and any time online FDR control. Our main contributions are: (i) to\npropose reasonable definitions of a null hypothesis for MAB instances; (ii) to\ndemonstrate how one can derive an always-valid sequential p-value that allows\ncontinuous monitoring of each MAB test; and (iii) to show that using rejection\nthresholds of online-FDR algorithms as the confidence levels for the MAB\nalgorithms results in both sample-optimality, high power and low FDR at any\npoint in time. We run extensive simulations to verify our claims, and also\nreport results on real data collected from the New Yorker Cartoon Caption\ncontest.\n",
"title": "A framework for Multi-A(rmed)/B(andit) testing with online FDR control"
} | null | null | null | null | true | null | 20558 | null | Default | null | null |
null | {
"abstract": " This paper presents a statistical method of single-channel speech enhancement\nthat uses a variational autoencoder (VAE) as a prior distribution on clean\nspeech. A standard approach to speech enhancement is to train a deep neural\nnetwork (DNN) to take noisy speech as input and output clean speech. Although\nthis supervised approach requires a very large amount of pair data for\ntraining, it is not robust against unknown environments. Another approach is to\nuse non-negative matrix factorization (NMF) based on basis spectra trained on\nclean speech in advance and those adapted to noise on the fly. This\nsemi-supervised approach, however, causes considerable signal distortion in\nenhanced speech due to the unrealistic assumption that speech spectrograms are\nlinear combinations of the basis spectra. Replacing the poor linear generative\nmodel of clean speech in NMF with a VAE---a powerful nonlinear deep generative\nmodel---trained on clean speech, we formulate a unified probabilistic\ngenerative model of noisy speech. Given noisy speech as observed data, we can\nsample clean speech from its posterior distribution. The proposed method\noutperformed the conventional DNN-based method in unseen noisy environments.\n",
"title": "Statistical Speech Enhancement Based on Probabilistic Integration of Variational Autoencoder and Non-Negative Matrix Factorization"
} | null | null | null | null | true | null | 20559 | null | Default | null | null |
null | {
"abstract": " Understanding the global optimality in deep learning (DL) has been attracting\nmore and more attention recently. Conventional DL solvers, however, have not\nbeen developed intentionally to seek for such global optimality. In this paper\nwe propose a novel approximation algorithm, BPGrad, towards optimizing deep\nmodels globally via branch and pruning. Our BPGrad algorithm is based on the\nassumption of Lipschitz continuity in DL, and as a result it can adaptively\ndetermine the step size for current gradient given the history of previous\nupdates, wherein theoretically no smaller steps can achieve the global\noptimality. We prove that, by repeating such branch-and-pruning procedure, we\ncan locate the global optimality within finite iterations. Empirically an\nefficient solver based on BPGrad for DL is proposed as well, and it outperforms\nconventional DL solvers such as Adagrad, Adadelta, RMSProp, and Adam in the\ntasks of object recognition, detection, and segmentation.\n",
"title": "BPGrad: Towards Global Optimality in Deep Learning via Branch and Pruning"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20560 | null | Validated | null | null |
null | {
"abstract": " Concepts and tools from network theory, the so-called Lagrangian Flow Network\nframework, have been successfully used to obtain a coarse-grained description\nof transport by closed fluid flows. Here we explore the application of this\nmethodology to open chaotic flows, and check it with numerical results for a\nmodel open flow, namely a jet with a localized wave perturbation. We find that\nnetwork nodes with high values of out-degree and of finite-time entropy in the\nforward-in-time direction identify the location of the chaotic saddle and its\nstable manifold, whereas nodes with high in-degree and backwards finite-time\nentropy highlight the location of the saddle and its unstable manifold. The\ncyclic clustering coefficient, associated to the presence of periodic orbits,\ntakes non-vanishing values at the location of the saddle itself.\n",
"title": "Lagrangian Flow Network approach to an open flow model"
} | null | null | null | null | true | null | 20561 | null | Default | null | null |
null | {
"abstract": " In this paper, by maximum principle and cutoff function, we investigate\ngradient estimates for positive solutions to two nonlinear parabolic equations\nunder Ricci flow. The related Harnack inequalities are deduced. An result about\npositive solutions on closed manifolds under Ricci flow is abtained. As\napplications, gradient estimates and Harnack inequalities for positive\nsolutions to the heat equation under Ricci flow are derived. These results in\nthe paper can be regard as generalizing the gradient estimates of Li-Yau, J. Y.\nLi, Hamilton and Li-Xu to the Ricci flow. Our results also improve the\nestimates of S. P. Liu and J. Sun to the nonlinear parabolic equation under\nRicci flow.\n",
"title": "Some new gradient estimates for two nonlinear parabolic equations under Ricci flow"
} | null | null | null | null | true | null | 20562 | null | Default | null | null |
null | {
"abstract": " In this paper, we extend the improved pointwise iteration-complexity result\nof a dynamic regularized alternating direction method of multipliers (ADMM) for\na new stepsize domain. In this complexity analysis, the stepsize parameter can\neven be chosen in the interval $(0,2)$ instead of interval\n$(0,(1+\\sqrt{5})/2)$. As usual, our analysis is established by interpreting\nthis ADMM variant as an instance of a hybrid proximal extragradient framework\napplied to a specific monotone inclusion problem.\n",
"title": "On the pointwise iteration-complexity of a dynamic regularized ADMM with over-relaxation stepsize"
} | null | null | null | null | true | null | 20563 | null | Default | null | null |
null | {
"abstract": " This paper presents a novel deep learning-based method for learning a\nfunctional representation of mammalian neural images. The method uses a deep\nconvolutional denoising autoencoder (CDAE) for generating an invariant, compact\nrepresentation of in situ hybridization (ISH) images. While most existing\nmethods for bio-imaging analysis were not developed to handle images with\nhighly complex anatomical structures, the results presented in this paper show\nthat functional representation extracted by CDAE can help learn features of\nfunctional gene ontology categories for their classification in a highly\naccurate manner. Using this CDAE representation, our method outperforms the\nprevious state-of-the-art classification rate, by improving the average AUC\nfrom 0.92 to 0.98, i.e., achieving 75% reduction in error. The method operates\non input images that were downsampled significantly with respect to the\noriginal ones to make it computationally feasible.\n",
"title": "DeepBrain: Functional Representation of Neural In-Situ Hybridization Images for Gene Ontology Classification Using Deep Convolutional Autoencoders"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20564 | null | Validated | null | null |
null | {
"abstract": " We study the following nonlocal diffusion equation in the Heisenberg group\n$\\mathbb{H}_n$, \\[ u_t(z,s,t)=J\\ast u(z,s,t)-u(z,s,t), \\] where $\\ast$ denote\nconvolution product and $J$ satisfies appropriated hypothesis. For the Cauchy\nproblem we obtain that the asymptotic behavior of the solutions is the same\nform that the one for the heat equation in the Heisenberg group. To obtain this\nresult we use the spherical transform related to the pair\n$(U(n),\\mathbb{H}_n)$. Finally we prove that solutions of properly rescaled\nnonlocal Dirichlet problem converge uniformly to the solution of the\ncorresponding Dirichlet problem for the classical heat equation in the\nHeisenberg group.\n",
"title": "Nonlocal heat equations in the Heisenberg group"
} | null | null | null | null | true | null | 20565 | null | Default | null | null |
null | {
"abstract": " Fast and efficient motion planning algorithms are crucial for many\nstate-of-the-art robotics applications such as self-driving cars. Existing\nmotion planning methods such as RRT*, A*, and D*, become ineffective as their\ncomputational complexity increases exponentially with the dimensionality of the\nmotion planning problem. To address this issue, we present a neural\nnetwork-based novel planning algorithm which generates end-to-end\ncollision-free paths irrespective of the obstacles' geometry. The proposed\nmethod, called MPNet (Motion Planning Network), comprises of a Contractive\nAutoencoder which encodes the given workspaces directly from a point cloud\nmeasurement, and a deep feedforward neural network which takes the workspace\nencoding, start and goal configuration, and generates end-to-end feasible\nmotion trajectories for the robot to follow. We evaluate MPNet on multiple\nplanning problems such as planning of a point-mass robot, rigid-body, and 7 DOF\nBaxter robot manipulators in various 2D and 3D environments. The results show\nthat MPNet is not only consistently computationally efficient in all 2D and 3D\nenvironments but also show remarkable generalization to completely unseen\nenvironments. The results also show that computation time of MPNet consistently\nremains less than 1 second which is significantly lower than existing\nstate-of-the-art motion planning algorithms. Furthermore, through transfer\nlearning, the MPNet trained in one scenario (e.g., indoor living places) can\nalso quickly adapt to new scenarios (e.g., factory floors) with a little amount\nof data.\n",
"title": "Motion Planning Networks"
} | null | null | null | null | true | null | 20566 | null | Default | null | null |
null | {
"abstract": " With the ever increasing size of web, relevant information extraction on the\nInternet with a query formed by a few keywords has become a big challenge. To\novercome this, query expansion (QE) plays a crucial role in improving the\nInternet searches, where the user's initial query is reformulated to a new\nquery by adding new meaningful terms with similar significance. QE -- as part\nof information retrieval (IR) -- has long attracted researchers' attention. It\nhas also become very influential in the field of personalized social document,\nQuestion Answering over Linked Data (QALD), and, Text Retrieval Conference\n(TREC) and REAL sets. This paper surveys QE techniques in IR from 1960 to 2017\nwith respect to core techniques, data sources used, weighting and ranking\nmethodologies, user participation and applications (of QE techniques) --\nbringing out similarities and differences.\n",
"title": "Query Expansion Techniques for Information Retrieval: a Survey"
} | null | null | null | null | true | null | 20567 | null | Default | null | null |
null | {
"abstract": " This paper fills a gap in aspect-based sentiment analysis and aims to present\na new method for preparing and analysing texts concerning opinion and\ngenerating user-friendly descriptive reports in natural language. We present a\ncomprehensive set of techniques derived from Rhetorical Structure Theory and\nsentiment analysis to extract aspects from textual opinions and then build an\nabstractive summary of a set of opinions. Moreover, we propose aspect-aspect\ngraphs to evaluate the importance of aspects and to filter out unimportant ones\nfrom the summary. Additionally, the paper presents a prototype solution of data\nflow with interesting and valuable results. The proposed method's results\nproved the high accuracy of aspect detection when applied to the gold standard\ndataset.\n",
"title": "Method for Aspect-Based Sentiment Annotation Using Rhetorical Analysis"
} | null | null | [
"Computer Science"
]
| null | true | null | 20568 | null | Validated | null | null |
null | {
"abstract": " Let $(X, \\mathscr{L}, \\lambda)$ and $(Y, \\mathscr{M}, \\mu)$ be finite measure\nspaces for which there exist $A \\in \\mathscr{L}$ and $B \\in \\mathscr{M}$ with\n$0 < \\lambda(A) < \\lambda(X)$ and $0 < \\mu(B) < \\mu(Y)$, and let $I\\subseteq\n\\mathbf{R}$ be a non-empty interval. We prove that, if $f$ and $g$ are\ncontinuous bijections $I \\to \\mathbf{R}^+$, then the equation $$\nf^{-1}\\!\\left(\\int_X f\\!\\left(g^{-1}\\!\\left(\\int_Y g \\circ\nh\\;d\\mu\\right)\\right)d \\lambda\\right)\\! = g^{-1}\\!\\left(\\int_Y\ng\\!\\left(f^{-1}\\!\\left(\\int_X f \\circ h\\;d\\lambda\\right)\\right)d \\mu\\right)$$\nis satisfied by every $\\mathscr{L} \\otimes \\mathscr{M}$-measurable simple\nfunction $h: X \\times Y \\to I$ if and only if $f=c g$ for some $c \\in\n\\mathbf{R}^+$ (it is easy to see that the equation is well posed). An\nanalogous, but essentially different, result, with $f$ and $g$ replaced by\ncontinuous injections $I \\to \\mathbf R$ and $\\lambda(X)=\\mu(Y)=1$, was recently\nobtained in [Indag. Math. 27 (2016), 945-953].\n",
"title": "Commutativity of integral quasi-arithmetic means on measure spaces"
} | null | null | null | null | true | null | 20569 | null | Default | null | null |
null | {
"abstract": " Partial differential equations with random inputs have become popular models\nto characterize physical systems with uncertainty coming from, e.g., imprecise\nmeasurement and intrinsic randomness. In this paper, we perform asymptotic rare\nevent analysis for such elliptic PDEs with random inputs. In particular, we\nconsider the asymptotic regime that the noise level converges to zero\nsuggesting that the system uncertainty is low, but does exists. We develop\nsharp approximations of the probability of a large class of rare events.\n",
"title": "Moderate Deviation for Random Elliptic PDEs with Small Noise"
} | null | null | null | null | true | null | 20570 | null | Default | null | null |
null | {
"abstract": " Music relies heavily on repetition to build structure and meaning.\nSelf-reference occurs on multiple timescales, from motifs to phrases to reusing\nof entire sections of music, such as in pieces with ABA structure. The\nTransformer (Vaswani et al., 2017), a sequence model based on self-attention,\nhas achieved compelling results in many generation tasks that require\nmaintaining long-range coherence. This suggests that self-attention might also\nbe well-suited to modeling music. In musical composition and performance,\nhowever, relative timing is critically important. Existing approaches for\nrepresenting relative positional information in the Transformer modulate\nattention based on pairwise distance (Shaw et al., 2018). This is impractical\nfor long sequences such as musical compositions since their memory complexity\nfor intermediate relative information is quadratic in the sequence length. We\npropose an algorithm that reduces their intermediate memory requirement to\nlinear in the sequence length. This enables us to demonstrate that a\nTransformer with our modified relative attention mechanism can generate\nminute-long compositions (thousands of steps, four times the length modeled in\nOore et al., 2018) with compelling structure, generate continuations that\ncoherently elaborate on a given motif, and in a seq2seq setup generate\naccompaniments conditioned on melodies. We evaluate the Transformer with our\nrelative attention mechanism on two datasets, JSB Chorales and\nPiano-e-Competition, and obtain state-of-the-art results on the latter.\n",
"title": "Music Transformer"
} | null | null | null | null | true | null | 20571 | null | Default | null | null |
null | {
"abstract": " Massive multiple-input multiple-output (M-MIMO) technique brings better\nenergy efficiency and coverage but higher computational complexity than\nsmall-scale MIMO. For linear detections such as minimum mean square error\n(MMSE), prohibitive complexity lies in solving large-scale linear equations.\nFor a better trade-off between bit-error-rate (BER) performance and\ncomputational complexity, iterative linear algorithms like conjugate gradient\n(CG) have been applied and have shown their feasibility in recent years. In\nthis paper, residual-based detection (RBD) algorithms are proposed for M-MIMO\ndetection, including minimal residual (MINRES) algorithm, generalized minimal\nresidual (GMRES) algorithm, and conjugate residual (CR) algorithm. RBD\nalgorithms focus on the minimization of residual norm per iteration, whereas\nmost existing algorithms focus on the approximation of exact signal. Numerical\nresults have shown that, for $64$-QAM $128\\times 8$ MIMO, RBD algorithms are\nonly $0.13$ dB away from the exact matrix inversion method when BER$=10^{-4}$.\nStability of RBD algorithms has also been verified in various correlation\nconditions. Complexity comparison has shown that, CR algorithm require $87\\%$\nless complexity than the traditional method for $128\\times 60$ MIMO. The\nunified hardware architecture is proposed with flexibility, which guarantees a\nlow-complexity implementation for a family of RBD M-MIMO detectors.\n",
"title": "Residual-Based Detections and Unified Architecture for Massive MIMO Uplink"
} | null | null | [
"Computer Science"
]
| null | true | null | 20572 | null | Validated | null | null |
null | {
"abstract": " A split feasibility formulation for the inverse problem of\nintensity-modulated radiation therapy (IMRT) treatment planning with\ndose-volume constraints (DVCs) included in the planning algorithm is presented.\nIt involves a new type of sparsity constraint that enables the inclusion of a\npercentage-violation constraint in the model problem and its handling by\ncontinuous (as opposed to integer) methods. We propose an iterative algorithmic\nframework for solving such a problem by applying the feasibility-seeking\nCQ-algorithm of Byrne combined with the automatic relaxation method (ARM) that\nuses cyclic projections. Detailed implementation instructions are furnished.\nFunctionality of the algorithm was demonstrated through the creation of an\nintensity-modulated proton therapy plan for a simple 2D C-shaped geometry and\nalso for a realistic base-of-skull chordoma treatment site. Monte Carlo\nsimulations of proton pencil beams of varying energy were conducted to obtain\ndose distributions for the 2D test case. A research release of the Pinnacle3\nproton treatment planning system was used to extract pencil beam doses for a\nclinical base-of-skull chordoma case. In both cases the beamlet doses were\ncalculated to satisfy dose-volume constraints according to our new algorithm.\nExamination of the dose-volume histograms following inverse planning with our\nalgorithm demonstrated that it performed as intended. The application of our\nproposed algorithm to dose-volume constraint inverse planning was successfully\ndemonstrated. Comparison with optimized dose distributions from the research\nrelease of the Pinnacle3 treatment planning system showed the algorithm could\nachieve equivalent or superior results.\n",
"title": "Sparsity constrained split feasibility for dose-volume constraints in inverse planning of intensity-modulated photon or proton therapy"
} | null | null | [
"Physics",
"Mathematics"
]
| null | true | null | 20573 | null | Validated | null | null |
null | {
"abstract": " In this paper we study the superalgebra $A_n$, introduced by the authors in\nprevious work on categorification of Verma modules for quantum\n$\\mathfrak{sl}_2$. The superalgebra $A_n$ is akin to the nilHecke algebra, and\nshares similar properties. In particular, we prove a uniqueness result about\n2-Verma modules on $\\Bbbk$-linear 2-categories.\n",
"title": "On 2-Verma modules for quantum $\\mathfrak{sl}_2$"
} | null | null | null | null | true | null | 20574 | null | Default | null | null |
null | {
"abstract": " We investigate preference profiles for a set $\\mathcal{V}$ of voters, where\neach voter $i$ has a preference order $\\succ_i$ on a finite set $A$ of\nalternatives (that is, a linear order on $A$) such that for each two\nalternatives $a,b\\in A$, voter $i$ prefers $a$ to $b$ if $a\\succ_i b$. Such a\nprofile is narcissistic if each alternative $a$ is preferred the most by at\nleast one voter. It is single-peaked if there is a linear order\n$\\triangleright^{\\text{sp}}$ on the alternatives such that each voter's\npreferences on the alternatives along the order $\\triangleright^{\\text{sp}}$\nare either strictly increasing, or strictly decreasing, or first strictly\nincreasing and then strictly decreasing. It is single-crossing if there is a\nlinear order $\\triangleright^{\\text{sc}}$ on the voters such that each pair of\nalternatives divides the order $\\triangleright^{\\text{sc}}$ into at most two\nsuborders, where in each suborder, all voters have the same linear order on\nthis pair.\nWe show that for $n$ voters and $n$ alternatives,the number of single-peaked\nnarcissistic profiles is $\\prod_{i=2}^{n-1} \\binom{n-1}{i-1}$ while the number\nof single-crossing narcissistic profiles is $2^{\\binom{n-1}{2}}$.\n",
"title": "On the Number of Single-Peaked Narcissistic or Single-Crossing Narcissistic Preference Profiles"
} | null | null | null | null | true | null | 20575 | null | Default | null | null |
null | {
"abstract": " This review is based on lectures given at the 45th Saas-Fee Advanced Course\n'From Protoplanetary Disks to Planet Formation' held in March 2015 in Les\nDiablerets, Switzerland. Starting with an overview of the main characterictics\nof the Solar System and extrasolar planets, we describe the planet formation\nprocess in terms of the sequential accretion scenario. First the growth\nprocesses of dust particles to planetesimals and subsequently to terrestrial\nplanets or planetary cores are presented. This is followed by the formation\nprocess of the giant planets either by core accretion or gravitational\ninstability. Finally, the dynamical evolution of the orbital elements as driven\nby disk-planet interaction and the overall evolution of multi-object systems is\npresented.\n",
"title": "Planet formation and disk-planet interactions"
} | null | null | [
"Physics"
]
| null | true | null | 20576 | null | Validated | null | null |
null | {
"abstract": " Using the representation theory of Cherednik algebras at $t=0$ and a Galois\ncovering of the Calogero-Moser space, we define the notions of left, right and\ntwo-sided Calogero-Moser cells for any finite complex reflection group. To each\nCaloger-Moser two-sided cell is associated a Calogero-Moser family, while to\neach Calogero-Moser left cell is associated a Calogero-Moser cellular\nrepresentation. We study properties of these objects and we conjecture that,\nwhenever the reflection group is real (i.e. is a Coxeter group), these notions\ncoincide with the one of Kazhdan-Lusztig left, right and two-sided cells,\nKazhdan-Lusztig families and Kazhdan-Lusztig cellular representations.\n",
"title": "Cherednik algebras and Calogero-Moser cells"
} | null | null | null | null | true | null | 20577 | null | Default | null | null |
null | {
"abstract": " We show that the Khovanov complex of a rational tangle has a very simple\nrepresentative whose backbone of non-zero morphisms forms a zig-zag.\nFurthermore, this minimal complex can be computed quickly by an inductive\nalgorithm. (For example, we calculate $Kh(8_2)$ by hand.) We find that the\nbigradings of the subobjects in these minimal complexes can be described by\nmatrix actions, which after a change of basis is the reduced Burau\nrepresentation of $B_3$.\n",
"title": "Khovanov complexes of rational tangles"
} | null | null | null | null | true | null | 20578 | null | Default | null | null |
null | {
"abstract": " We study how shocks to the forward-looking expectations of investors buying\ncall and put options transmit across the financial system. We introduce a new\ncontagion measure, called asymmetric fear connectedness (AFC), which captures\nthe information related to \"fear\" on the two sides of the options market and\ncan be used as a forward-looking systemic risk monitoring tool. The decomposed\nconnectedness measures provide timely predictive information for near-future\nmacroeconomic conditions and uncertainty indicators, and they contain\nadditional valuable information that is not included in the aggregate\nconnectedness measure. The role of a positive/negative \"fear\"\ntransmitter/receiver emerges clearly when we focus more closely on\nidiosyncratic events for financial institutions. We identify banks that are\npredominantly positive/negative receivers of \"fear\", as well as banks that\npositively/negatively transmit \"fear\" in the financial system.\n",
"title": "Asymmetric Connectedness of Fears in the U.S. Financial Sector"
} | null | null | [
"Quantitative Finance"
]
| null | true | null | 20579 | null | Validated | null | null |
null | {
"abstract": " This work presents a novel framework based on feed-forward neural network for\ntext-independent speaker classification and verification, two related systems\nof speaker recognition. With optimized features and model training, it achieves\n100% classification rate in classification and less than 6% Equal Error Rate\n(ERR), using merely about 1 second and 5 seconds of data respectively. Features\nwith stricter Voice Active Detection (VAD) than the regular one for speech\nrecognition ensure extracting stronger voiced portion for speaker recognition,\nspeaker-level mean and variance normalization helps to eliminate the\ndiscrepancy between samples from the same speaker. Both are proven to improve\nthe system performance. In building the neural network speaker classifier, the\nnetwork structure parameters are optimized with grid search and dynamically\nreduced regularization parameters are used to avoid training terminated in\nlocal minimum. It enables the training goes further with lower cost. In speaker\nverification, performance is improved with prediction score normalization,\nwhich rewards the speaker identity indices with distinct peaks and penalizes\nthe weak ones with high scores but more competitors, and speaker-specific\nthresholding, which significantly reduces ERR in the ROC curve. TIMIT corpus\nwith 8K sampling rate is used here. First 200 male speakers are used to train\nand test the classification performance. The testing files of them are used as\nin-domain registered speakers, while data from the remaining 126 male speakers\nare used as out-of-domain speakers, i.e. imposters in speaker verification.\n",
"title": "Neural Network Based Speaker Classification and Verification Systems with Enhanced Features"
} | null | null | null | null | true | null | 20580 | null | Default | null | null |
null | {
"abstract": " Slimness of a graph measures the local deviation of its metric from a tree\nmetric. In a graph $G=(V,E)$, a geodesic triangle $\\bigtriangleup(x,y,z)$ with\n$x, y, z\\in V$ is the union $P(x,y) \\cup P(x,z) \\cup P(y,z)$ of three shortest\npaths connecting these vertices. A geodesic triangle $\\bigtriangleup(x,y,z)$ is\ncalled $\\delta$-slim if for any vertex $u\\in V$ on any side $P(x,y)$ the\ndistance from $u$ to $P(x,z) \\cup P(y,z)$ is at most $\\delta$, i.e. each path\nis contained in the union of the $\\delta$-neighborhoods of two others. A graph\n$G$ is called $\\delta$-slim, if all geodesic triangles in $G$ are\n$\\delta$-slim. The smallest value $\\delta$ for which $G$ is $\\delta$-slim is\ncalled the slimness of $G$. In this paper, using the layering partition\ntechnique, we obtain sharp bounds on slimness of such families of graphs as (1)\ngraphs with cluster-diameter $\\Delta(G)$ of a layering partition of $G$, (2)\ngraphs with tree-length $\\lambda$, (3) graphs with tree-breadth $\\rho$, (4)\n$k$-chordal graphs, AT-free graphs and HHD-free graphs. Additionally, we show\nthat the slimness of every 4-chordal graph is at most 2 and characterize those\n4-chordal graphs for which the slimness of every of its induced subgraph is at\nmost 1.\n",
"title": "Slimness of graphs"
} | null | null | null | null | true | null | 20581 | null | Default | null | null |
null | {
"abstract": " Despite their immense popularity, deep learning-based acoustic systems are\ninherently vulnerable to adversarial attacks, wherein maliciously crafted\naudios trigger target systems to misbehave. In this paper, we present\nSirenAttack, a new class of attacks to generate adversarial audios. Compared\nwith existing attacks, SirenAttack highlights with a set of significant\nfeatures: (i) versatile -- it is able to deceive a range of end-to-end acoustic\nsystems under both white-box and black-box settings; (ii) effective -- it is\nable to generate adversarial audios that can be recognized as specific phrases\nby target acoustic systems; and (iii) stealthy -- it is able to generate\nadversarial audios indistinguishable from their benign counterparts to human\nperception. We empirically evaluate SirenAttack on a set of state-of-the-art\ndeep learning-based acoustic systems (including speech command recognition,\nspeaker recognition and sound event classification), with results showing the\nversatility, effectiveness, and stealthiness of SirenAttack. For instance, it\nachieves 99.45% attack success rate on the IEMOCAP dataset against the ResNet18\nmodel, while the generated adversarial audios are also misinterpreted by\nmultiple popular ASR platforms, including Google Cloud Speech, Microsoft Bing\nVoice, and IBM Speech-to-Text. We further evaluate three potential defense\nmethods to mitigate such attacks, including adversarial training, audio\ndownsampling, and moving average filtering, which leads to promising directions\nfor further research.\n",
"title": "SirenAttack: Generating Adversarial Audio for End-to-End Acoustic Systems"
} | null | null | null | null | true | null | 20582 | null | Default | null | null |
null | {
"abstract": " We present an $O((\\log k)^2)$-competitive randomized algorithm for the\n$k$-server problem on hierarchically separated trees (HSTs). This is the first\n$o(k)$-competitive randomized algorithm for which the competitive ratio is\nindependent of the size of the underlying HST. Our algorithm is designed in the\nframework of online mirror descent where the mirror map is a multiscale\nentropy. When combined with Bartal's static HST embedding reduction, this leads\nto an $O((\\log k)^2 \\log n)$-competitive algorithm on any $n$-point metric\nspace. We give a new dynamic HST embedding that yields an $O((\\log k)^3 \\log\n\\Delta)$-competitive algorithm on any metric space where the ratio of the\nlargest to smallest non-zero distance is at most $\\Delta$.\n",
"title": "k-server via multiscale entropic regularization"
} | null | null | null | null | true | null | 20583 | null | Default | null | null |
null | {
"abstract": " We present a general framework for studying regularized estimators; i.e.,\nestimation problems wherein \"plug-in\" type estimators are either ill-defined or\nill-behaved. We derive primitive conditions that imply consistency and\nasymptotic linear representation for regularized estimators, allowing for\nslower than $\\sqrt{n}$ estimators as well as infinite dimensional parameters.\nWe also provide data-driven methods for choosing tuning parameters that, under\nsome conditions, achieve the aforementioned results. We illustrate the scope of\nour approach by studying a wide range of applications, revisiting known results\nand deriving new ones.\n",
"title": "Some Large Sample Results for the Method of Regularized Estimators"
} | null | null | null | null | true | null | 20584 | null | Default | null | null |
null | {
"abstract": " 'Sharing of statistical strength' is a phrase often employed in machine\nlearning and signal processing. In sensor networks, for example, missing\nsignals from certain sensors may be predicted by exploiting their correlation\nwith observed signals acquired from other sensors. For humans, our hands move\nsynchronously with our legs, and we can exploit these implicit correlations for\npredicting new poses and for generating new natural-looking walking sequences.\nWe can also go much further and exploit this form of transfer learning, to\ndevelop new control schemas for robust control of rehabilitation robots. In\nthis short paper we introduce coregionalised locomotion envelopes - a method\nfor multi-dimensional manifold regression, on human locomotion variates. Herein\nwe render a qualitative description of this method.\n",
"title": "Coregionalised Locomotion Envelopes - A Qualitative Approach"
} | null | null | null | null | true | null | 20585 | null | Default | null | null |
null | {
"abstract": " This paper uses model symmetries in the instrumental variable (IV) regression\nto derive an invariant test for the causal structural parameter. Contrary to\npopular belief, we show there exist model symmetries when equation errors are\nheteroskedastic and autocorrelated (HAC). Our theory is consistent with\nexisting results for the homoskedastic model (Andrews, Moreira and Stock(2006}\nand Chamberlain (2007}), but in general uses information on the structural\nparameter beyond the Anderson-Rubin, score, and rank statistics. This suggests\nthat tests based only the Anderson-Rubin and score statistics discard\ninformation on the causal parameter of interest. We apply our theory to\nconstruct designs in which these tests indeed have power arbitrarily close to\nsize. Other tests, including other adaptations to the CLR test, do not suffer\nthe same deficiencies. Finally, we use the model symmetries to propose novel\nweighted-average power tests for the HAC-IV model.\n",
"title": "Optimal Invariant Tests in an Instrumental Variables Regression With Heteroskedastic and Autocorrelated Errors"
} | null | null | [
"Mathematics",
"Statistics"
]
| null | true | null | 20586 | null | Validated | null | null |
null | {
"abstract": " In this paper we use the classical electrodynamics to show that the Lorenz\ngauge can be incompatible with some particular solutions of the d Alembert\nequations for electromagnetic potentials. In its turn, the d Alembert equations\nfor the elec- tromagnetic potentials is the result of application of the Lorenz\ngauge to general equations for the potentials. The last ones is the\nstraightforward consequence of Maxwell equations. Since the d Alembert\nequations and the electromagnetic poten- tials are necessary for quantum\nelectrodynamics formulation, one should oblige to satisfy these equations also\nin classical case. The solution of d Alembert equations, which modifies\nlongitudinal electric field is found. The requirement of this modifi- cation\nfollows from the necessity to satisfy the physical condition of impossibility\nof instantaneous transferring of interaction in space.\n",
"title": "Longitudinal electric field: from Maxwell equation to non-locality in time and space"
} | null | null | null | null | true | null | 20587 | null | Default | null | null |
null | {
"abstract": " Assume that we observe a sample of size n composed of p-dimensional signals,\neach signal having independent entries drawn from a scaled Poisson distribution\nwith an unknown intensity. We are interested in estimating the sum of the n\nunknown intensity vectors, under the assumption that most of them coincide with\na given 'background' signal. The number s of p-dimensional signals different\nfrom the background signal plays the role of sparsity and the goal is to\nleverage this sparsity assumption in order to improve the quality of estimation\nas compared to the naive estimator that computes the sum of the observed\nsignals. We first introduce the group hard thresholding estimator and analyze\nits mean squared error measured by the squared Euclidean norm. We establish a\nnonasymptotic upper bound showing that the risk is at most of the order of\n{\\sigma}^2(sp + s^2sqrt(p)) log^3/2(np). We then establish lower bounds on the\nminimax risk over a properly defined class of collections of s-sparse signals.\nThese lower bounds match with the upper bound, up to logarithmic terms, when\nthe dimension p is fixed or of larger order than s^2. In the case where the\ndimension p increases but remains of smaller order than s^2, our results show a\ngap between the lower and the upper bounds, which can be up to order sqrt(p).\n",
"title": "Estimating linear functionals of a sparse family of Poisson means"
} | null | null | null | null | true | null | 20588 | null | Default | null | null |
null | {
"abstract": " Pulsed-laser dry printing of noble-metal microrings with a tunable internal\nporous structure, which can be revealed via an ion-beam etching post-procedure,\nwas demonstrated. Abundance and average size of the pores inside the microrings\nwere shown to be tuned in a wide range by varying incident pulse energy and a\nnitrogen doping level controlled in the process of magnetron deposition of the\ngold film in the appropriate gaseous environment. The fabricated porous\nmicrorings were shown to provide many-fold near-field enhancement of incident\nelectromagnetic fields, which was confirmed by mapping of the characteristic\nRaman band of a nanometer-thick covering layer of Rhodamine 6G dye molecules\nand supporting finite-difference time-domain calculations. The proposed laser\nprinting/ion-beam etching approach is demonstrated to be a unique tool aimed at\ndesigning and fabricating multifunctional plasmonic structures and metasurfaces\nfor spectroscopic bioidentification based on surface-enhanced infrared\nabsorption, Raman scattering and photoluminescence detection schemes.\n",
"title": "Fabrication of porous microrings via laser printing and ion-beam post-etching"
} | null | null | [
"Physics"
]
| null | true | null | 20589 | null | Validated | null | null |
null | {
"abstract": " Increasing amounts of data from varied sources, particularly in the fields of\nmachine learning and graph analytics, are causing storage requirements to grow\nrapidly. A variety of technologies exist for storing and sharing these data,\nranging from parallel file systems used by supercomputers to distributed block\nstorage systems found in clouds. Relatively few comparative measurements exist\nto inform decisions about which storage systems are best suited for particular\ntasks. This work provides these measurements for two of the most popular\nstorage technologies: Lustre and Amazon S3. Lustre is an open-source, high\nperformance, parallel file system used by many of the largest supercomputers in\nthe world. Amazon's Simple Storage Service, or S3, is part of the Amazon Web\nServices offering, and offers a scalable, distributed option to store and\nretrieve data from anywhere on the Internet. Parallel processing is essential\nfor achieving high performance on modern storage systems. The performance tests\nused span the gamut of parallel I/O scenarios, ranging from single-client,\nsingle-node Amazon S3 and Lustre performance to a large-scale, multi-client\ntest designed to demonstrate the capabilities of a modern storage appliance\nunder heavy load. These results show that, when parallel I/O is used correctly\n(i.e., many simultaneous read or write processes), full network bandwidth\nperformance is achievable and ranged from 10 gigabits/s over a 10 GigE S3\nconnection to 0.35 terabits/s using Lustre on a 1200 port 10 GigE switch. These\nresults demonstrate that S3 is well-suited to sharing vast quantities of data\nover the Internet, while Lustre is well-suited to processing large quantities\nof data locally.\n",
"title": "Performance Measurements of Supercomputing and Cloud Storage Solutions"
} | null | null | null | null | true | null | 20590 | null | Default | null | null |
null | {
"abstract": " This paper proposes a distributed consensus algorithm for linear event-based\nheterogeneous multi-agent systems (MAS). The proposed scheme is event-triggered\nin the sense that an agent selectively transmits its information within its\nlocal neighbourhood based on a directed network topology under the fulfillment\nof certain conditions. Using the Lyapunov stability theorem, the system\nconstraints and event-triggering condition are expressed in terms of several\nlinear matrix inequalities (LMIs) to derive the consensus parameters. The\nobjective is to design the transmission threshold and minimum-norm\nheterogeneous control gains which collectively ensure an exponential consensus\nconvergence rate for the closed-loop systems. The LMI computed control gains\nare robust to uncertainty with some deviation from their nominal values\nallowed. The practicability of the proposed event-based framework is further\nstudied by proving the Zeno behaviour exclusion. Numerical simulations quantify\nthe advantages of our event-triggered consensus approach in second-order,\nlinear and heterogeneous multi-agent systems.\n",
"title": "Multi-Objective Event-triggered Consensus of Linear Multi-agent Systems"
} | null | null | [
"Computer Science"
]
| null | true | null | 20591 | null | Validated | null | null |
null | {
"abstract": " A secure and private framework for inter-agent communication and coordination\nis developed. This allows an agent, in our case a fleet owner, to ask questions\nor submit queries in an encrypted fashion using semi-homomorphic encryption.\nThe submitted query can be about the interest of the other fleet owners for\nusing a road at a specific time of the day, for instance, for the purpose of\ncollaborative vehicle platooning. The other agents can then provide appropriate\nresponses without knowing the content of the questions or the queries. Strong\nprivacy and security guarantees are provided for the agent who is submitting\nthe queries. It is also shown that the amount of the information that this\nagent can extract from the other agent is bounded. In fact, with submitting one\nquery, a sophisticated agent can at most extract the answer to two queries.\nThis secure communication platform is used subsequently to develop a\ndistributed coordination mechanisms among fleet owners.\n",
"title": "Private and Secure Coordination of Match-Making for Heavy-Duty Vehicle Platooning"
} | null | null | null | null | true | null | 20592 | null | Default | null | null |
null | {
"abstract": " Under covariate shift, training (source) data and testing (target) data\ndiffer in input space distribution, but share the same conditional label\ndistribution. This poses a challenging machine learning task. Robust Bias-Aware\n(RBA) prediction provides the conditional label distribution that is robust to\nthe worstcase logarithmic loss for the target distribution while matching\nfeature expectation constraints from the source distribution. However,\nemploying RBA with insufficient feature constraints may result in high\ncertainty predictions for much of the source data, while leaving too much\nuncertainty for target data predictions. To overcome this issue, we extend the\nrepresenter theorem to the RBA setting, enabling minimization of regularized\nexpected target risk by a reweighted kernel expectation under the source\ndistribution. By applying kernel methods, we establish consistency guarantees\nand demonstrate better performance of the RBA classifier than competing methods\non synthetically biased UCI datasets as well as datasets that have natural\ncovariate shift.\n",
"title": "Kernel Robust Bias-Aware Prediction under Covariate Shift"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20593 | null | Validated | null | null |
null | {
"abstract": " While enormous progress has been made to Variational Autoencoder (VAE) in\nrecent years, similar to other deep networks, VAE with deep networks suffers\nfrom the problem of degeneration, which seriously weakens the correlation\nbetween the input and the corresponding latent codes, deviating from the goal\nof the representation learning. To investigate how degeneration affects VAE\nfrom a theoretical perspective, we illustrate the information transmission in\nVAE and analyze the intermediate layers of the encoders/decoders. Specifically,\nwe propose a Fisher Information measure for the layer-wise analysis. With such\nmeasure, we demonstrate that information loss is ineluctable in feed-forward\nnetworks and causes the degeneration in VAE. We show that skip connections in\nVAE enable the preservation of information without changing the model\narchitecture. We call this class of VAE equipped with skip connections as SCVAE\nand perform a range of experiments to show its advantages in information\npreservation and degeneration mitigation.\n",
"title": "Degeneration in VAE: in the Light of Fisher Information Loss"
} | null | null | null | null | true | null | 20594 | null | Default | null | null |
null | {
"abstract": " We demonsrtate electrical spin injection and detection in $n$-type Ge\n($n$-Ge) at room temperature using four-terminal nonlocal spin-valve and\nHanle-effect measurements in lateral spin-valve (LSV) devices with\nHeusler-alloy Schottky tunnel contacts. The spin diffusion length\n($\\lambda$$_{\\rm Ge}$) of the Ge layer used ($n \\sim$ 1 $\\times$ 10$^{19}$\ncm$^{-3}$) at 296 K is estimated to be $\\sim$ 0.44 $\\pm$ 0.02 $\\mu$m.\nRoom-temperature spin signals can be observed reproducibly at the low bias\nvoltage range ($\\le$ 0.7 V) for LSVs with relatively low resistance-area\nproduct ($RA$) values ($\\le$ 1 k$\\Omega$$\\mu$m$^{2}$). This means that the\nSchottky tunnel contacts used here are more suitable than ferromagnet/MgO\ntunnel contacts ($RA \\ge$ 100 k$\\Omega$$\\mu$m$^{2}$) for developing Ge\nspintronic applications.\n",
"title": "Room-temperature spin transport in n-Ge probed by four-terminal nonlocal measurements"
} | null | null | null | null | true | null | 20595 | null | Default | null | null |
null | {
"abstract": " We present an algorithm for rapidly learning controllers for robotics\nsystems. The algorithm follows the model-based reinforcement learning paradigm,\nand improves upon existing algorithms; namely Probabilistic learning in Control\n(PILCO) and a sample-based version of PILCO with neural network dynamics\n(Deep-PILCO). We propose training a neural network dynamics model using\nvariational dropout with truncated Log-Normal noise. This allows us to obtain a\ndynamics model with calibrated uncertainty, which can be used to simulate\ncontroller executions via rollouts. We also describe set of techniques,\ninspired by viewing PILCO as a recurrent neural network model, that are crucial\nto improve the convergence of the method. We test our method on a variety of\nbenchmark tasks, demonstrating data-efficiency that is competitive with PILCO,\nwhile being able to optimize complex neural network controllers. Finally, we\nassess the performance of the algorithm for learning motor controllers for a\nsix legged autonomous underwater vehicle. This demonstrates the potential of\nthe algorithm for scaling up the dimensionality and dataset sizes, in more\ncomplex control tasks.\n",
"title": "Synthesizing Neural Network Controllers with Probabilistic Model based Reinforcement Learning"
} | null | null | null | null | true | null | 20596 | null | Default | null | null |
null | {
"abstract": " SCADA protocols for Industrial Control Systems (ICS) are vulnerable to\nnetwork attacks such as session hijacking. Hence, research focuses on network\nanomaly detection based on meta--data (message sizes, timing, command\nsequence), or on the state values of the physical process. In this work we\npresent a class of semantic network-based attacks against SCADA systems that\nare undetectable by the above mentioned anomaly detection. After hijacking the\ncommunication channels between the Human Machine Interface (HMI) and\nProgrammable Logic Controllers (PLCs), our attacks cause the HMI to present a\nfake view of the industrial process, deceiving the human operator into taking\nmanual actions. Our most advanced attack also manipulates the messages\ngenerated by the operator's actions, reversing their semantic meaning while\ncausing the HMI to present a view that is consistent with the attempted human\nactions. The attacks are totaly stealthy because the message sizes and timing,\nthe command sequences, and the data values of the ICS's state all remain\nlegitimate.\nWe implemented and tested several attack scenarios in the test lab of our\nlocal electric company, against a real HMI and real PLCs, separated by a\ncommercial-grade firewall. We developed a real-time security assessment tool,\nthat can simultaneously manipulate the communication to multiple PLCs and cause\nthe HMI to display a coherent system--wide fake view. Our tool is configured\nwith message-manipulating rules written in an ICS Attack Markup Language (IAML)\nwe designed, which may be of independent interest. Our semantic attacks all\nsuccessfully fooled the operator and brought the system to states of blackout\nand possible equipment damage.\n",
"title": "Stealthy Deception Attacks Against SCADA Systems"
} | null | null | null | null | true | null | 20597 | null | Default | null | null |
null | {
"abstract": " Despite their significant functional roles, beta-band oscillations are least\nunderstood. Synchronization in neuronal networks have attracted much attention\nin recent years with the main focus on transition type. Whether one obtains\nexplosive transition or a continuous transition is an important feature of the\nneuronal network which can depend on network structure as well as synaptic\ntypes. In this study we consider the effect of synaptic interaction (electrical\nand chemical) as well as structural connectivity on synchronization transition\nin network models of Izhikevich neurons which spike regularly with beta\nrhythms. We find a wide range of behavior including continuous transition,\nexplosive transition, as well as lack of global order. The stronger electrical\nsynapses are more conducive to synchronization and can even lead to explosive\nsynchronization. The key network element which determines the order of\ntransition is found to be the clustering coefficient and not the small world\neffect, or the existence of hubs in a network. These results are in contrast to\nprevious results which use phase oscillator models such as the Kuramoto model.\nFurthermore, we show that the patterns of synchronization changes when one goes\nto the gamma band. We attribute such a change to the change in the refractory\nperiod of Izhikevich neurons which changes significantly with frequency.\n",
"title": "Beta-rhythm oscillations and synchronization transition in network models of Izhikevich neurons: effect of topology and synaptic type"
} | null | null | [
"Quantitative Biology"
]
| null | true | null | 20598 | null | Validated | null | null |
null | {
"abstract": " We propose a natural relaxation of differential privacy based on the Renyi\ndivergence. Closely related notions have appeared in several recent papers that\nanalyzed composition of differentially private mechanisms. We argue that the\nuseful analytical tool can be used as a privacy definition, compactly and\naccurately representing guarantees on the tails of the privacy loss.\nWe demonstrate that the new definition shares many important properties with\nthe standard definition of differential privacy, while additionally allowing\ntighter analysis of composite heterogeneous mechanisms.\n",
"title": "Renyi Differential Privacy"
} | null | null | null | null | true | null | 20599 | null | Default | null | null |
null | {
"abstract": " In this paper, we propose a quality enhancement network for Versatile Video\nCoding (VVC) compressed videos by jointly exploiting spatial details and\ntemporal structure (SDTS). The network consists of a temporal structure\nprediction subnet and a spatial detail enhancement subnet. The former subnet is\nused to estimate and compensate the temporal motion across frames, and the\nspatial detail subnet is used to reduce the compression artifacts and enhance\nthe reconstruction quality of the VVC compressed video. Experimental results\ndemonstrate the effectiveness of our SDTS-based approach. It offers over\n7.82$\\%$ BD-rate saving on the common test video sequences and achieves the\nstate-of-the-art performance.\n",
"title": "Enhancing Quality for VVC Compressed Videos by Jointly Exploiting Spatial Details and Temporal Structure"
} | null | null | [
"Computer Science"
]
| null | true | null | 20600 | null | Validated | null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.