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": " Brain-computer interfaces (BCIs) can provide an alternative means of\ncommunication for individuals with severe neuromuscular limitations. The\nP300-based BCI speller relies on eliciting and detecting transient\nevent-related potentials (ERPs) in electroencephalography (EEG) data, in\nresponse to a user attending to rarely occurring target stimuli amongst a\nseries of non-target stimuli. However, in most P300 speller implementations,\nthe stimuli to be presented are randomly selected from a limited set of options\nand stimulus selection and presentation are not optimized based on previous\nuser data. In this work, we propose a data-driven method for stimulus selection\nbased on the expected discrimination gain metric. The data-driven approach\nselects stimuli based on previously observed stimulus responses, with the aim\nof choosing a set of stimuli that will provide the most information about the\nuser's intended target character. Our approach incorporates knowledge of\nphysiological and system constraints imposed due to real-time BCI\nimplementation. Simulations were performed to compare our stimulus selection\napproach to the row-column paradigm, the conventional stimulus selection method\nfor P300 spellers. Results from the simulations demonstrated that our adaptive\nstimulus selection approach has the potential to significantly improve\nperformance from the conventional method: up to 34% improvement in accuracy and\n43% reduction in the mean number of stimulus presentations required to spell a\ncharacter in a 72-character grid. In addition, our greedy approach to stimulus\nselection provides the flexibility to accommodate design constraints.\n",
"title": "Adaptive Stimulus Selection in ERP-Based Brain-Computer Interfaces by Maximizing Expected Discrimination Gain"
}
| null | null | null | null | true | null |
11001
| null |
Default
| null | null |
null |
{
"abstract": " We experimentally investigate the bursting dynamics of confined liquid film\nsuspended in air and find a viscous dynamics distinctly different from the\nnon-confined counterpart, due to lack of circular symmetry in the shape of\nexpanding hole: the novel confined-viscous bursting proceeds at a constant\nspeed and a rim formed at the bursting tip does not grow. We find a\nconfined-viscous to confined-inertial crossover, as well as a\nnonconfined-inertial to confined-inertial crossover, at which bursting speed\ndoes not change although the circular symmetry in the hole shape breaks\ndynamically.\n",
"title": "Bursting dynamics of viscous film without circular symmetry: the effect of confinement"
}
| null | null | null | null | true | null |
11002
| null |
Default
| null | null |
null |
{
"abstract": " We discuss a low energy $e^+e^-$ collider for production of the not yet\nobserved ($\\mu^+\\mu^-$) bound system (dimuonium). Collider with large crossing\nangle for $e^+e^-$ beams intersection produces dimuonium with non-zero\nmomentum, therefore, its decay point is shifted from the beam collision area\nproviding effective suppression of the elastic $e^+e^-$ scattering background.\nThe experimental constraints define subsequent collider specifications. We show\npreliminary layout of the accelerator and obtained main parameters. High\nluminosity in chosen beam energy range allows to study $\\pi^\\pm$ and $\\eta$\n-mesons.\n",
"title": "Low-energy electron-positron collider to search and study (μ^+μ^-) bound state"
}
| null | null | null | null | true | null |
11003
| null |
Default
| null | null |
null |
{
"abstract": " We introduce and investigate different definitions of effective amenability,\nin terms of computability of F{\\o}lner sets, Reiter functions, and F{\\o}lner\nfunctions. As a consequence, we prove that recursively presented amenable\ngroups have subrecursive F{\\o}lner function, answering a question of Gromov,\nfor the same class of groups we prove that solvability of the Equality Problem\non a generic set (generic EP) is equivalent to solvability of the Word Problem\non the whole group (WP), thus providing the first examples of finitely\npresented groups with unsolvable generic EP. In particular, we prove that for\nfinitely presented groups, solvability of generic WP doesn't imply solvability\nof generic EP.\n",
"title": "Følner functions and the generic Word Problem for finitely generated amenable groups"
}
| null | null | null | null | true | null |
11004
| null |
Default
| null | null |
null |
{
"abstract": " In this study, we examine a collection of health-related news articles\npublished by reliable and unreliable media outlets. Our analysis shows that\nthere are structural, topical, and semantic differences in the way reliable and\nunreliable media outlets conduct health journalism. We argue that the findings\nfrom this study will be useful for combating health disinformation problem.\n",
"title": "Differences between Health Related News Articles from Reliable and Unreliable Media"
}
| null | null | null | null | true | null |
11005
| null |
Default
| null | null |
null |
{
"abstract": " In their seminal work `Robust Replication of Volatility Derivatives,' Carr\nand Lee show how to robustly price and replicate a variety of claims written on\nthe quadratic variation of a risky asset under the assumption that the asset's\nvolatility process is independent of the Brownian motion that drives the\nasset's price. Additionally, they propose a correlation immunization method\nthat minimizes the pricing and hedging error that results when the correlation\nbetween the risky asset's price and volatility is nonzero. In this paper, we\nperform a number of Monte Carlo experiments to test the effectiveness of Carr\nand Lee's immunization strategy. Our results indicate that the correlation\nimmunization method is an effective means of reducing pricing and hedging\nerrors that result from nonzero correlation.\n",
"title": "A Numerical Study of Carr and Lee's Correlation Immunization Strategy for Volatility Derivatives"
}
| null | null |
[
"Quantitative Finance"
] | null | true | null |
11006
| null |
Validated
| null | null |
null |
{
"abstract": " We explore the effects of asymmetry of hopping parameters between double\nparallel quantum dots and the leads on the conductance and a possibility of\nlocal magnetic moment formation in this system using functional renormalization\ngroup approach with the counterterm. We demonstrate a possibility of a quantum\nphase transition to a local moment regime (so called singular Fermi liquid\n(SFL) state) for various types of hopping asymmetries and discuss respective\ngate voltage dependences of the conductance. It is shown, that depending on the\ntype of the asymmetry, the system can demonstrate either a first order quantum\nphase transition to SFL state, accompanied by a discontinuous change of the\nconductance, similarly to the symmetric case, or the second order quantum phase\ntransition, in which the conductance is continuous and exhibits Fano-type\nasymmetric resonance near the transition point. A semi-analytical explanation\nof these different types of conductance behavior is presented.\n",
"title": "Functional renormalization group study of parallel double quantum dots: Effects of asymmetric dot-lead couplings"
}
| null | null | null | null | true | null |
11007
| null |
Default
| null | null |
null |
{
"abstract": " Purpose: We propose a phenotype-based artificial intelligence system that can\nself-learn and is accurate for screening purposes, and test it on a Level IV\nmonitoring system. Methods: Based on the physiological knowledge, we\nhypothesize that the phenotype information will allow us to find subjects from\na well-annotated database that share similar sleep apnea patterns. Therefore,\nfor a new-arriving subject, we can establish a prediction model from the\nexisting database that is adaptive to the subject. We test the proposed\nalgorithm on a database consisting of 62 subjects with the signals recorded\nfrom a Level IV wearable device measuring the thoracic and abdominal movements\nand the SpO2. Results: With the leave-one cross validation, the accuracy of the\nproposed algorithm to screen subjects with an apnea-hypopnea index greater or\nequal to 15 is 93.6%, the positive likelihood ratio is 6.8, and the negative\nlikelihood ratio is 0.03. Conclusion: The results confirm the hypothesis and\nshow that the proposed algorithm has great potential to screen patients with\nSAS.\n",
"title": "Phenotype-based and Self-learning Inter-individual Sleep Apnea Screening with a Level IV Monitoring System"
}
| null | null |
[
"Statistics"
] | null | true | null |
11008
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper, we address the problem of reconstructing a time-domain signal\n(or a phase spectrogram) solely from a magnitude spectrogram. Since magnitude\nspectrograms do not contain phase information, we must restore or infer phase\ninformation to reconstruct a time-domain signal. One widely used approach for\ndealing with the signal reconstruction problem was proposed by Griffin and Lim.\nThis method usually requires many iterations for the signal reconstruction\nprocess and depending on the inputs, it does not always produce high-quality\naudio signals. To overcome these shortcomings, we apply a learning-based\napproach to the signal reconstruction problem by modeling the signal\nreconstruction process using a deep neural network and training it using the\nidea of a generative adversarial network. Experimental evaluations revealed\nthat our method was able to reconstruct signals faster with higher quality than\nthe Griffin-Lim method.\n",
"title": "Generative adversarial network-based approach to signal reconstruction from magnitude spectrograms"
}
| null | null | null | null | true | null |
11009
| null |
Default
| null | null |
null |
{
"abstract": " The braids of $B\\_\\infty$ can be equipped with a selfdistributive operation\n$\\mathbin{\\triangleright}$ enjoying a number of deep properties. This text is a\nsurvey of known properties and open questions involving this structure, its\nquotients, and its extensions.\n",
"title": "The Braid Shelf"
}
| null | null |
[
"Mathematics"
] | null | true | null |
11010
| null |
Validated
| null | null |
null |
{
"abstract": " The exciton relaxation dynamics of photoexcited electronic states in\npoly($p$-phenylenevinylene) (PPV) are theoretically investigated within a\ncoarse-grained model, in which both the exciton and nuclear degrees of freedom\nare treated quantum mechanically. The Frenkel-Holstein Hamiltonian is used to\ndescribe the strong exciton-phonon coupling present in the system, while\nexternal damping of the internal nuclear degrees of freedom are accounted for\nby a Lindblad master equation. Numerically, the dynamics are computed using the\ntime evolving block decimation (TEBD) and quantum jump trajectory techniques.\nThe values of the model parameters physically relevant to polymer systems\nnaturally lead to a separation of time scales, with the ultra-fast dynamics\ncorresponding to energy transfer from the exciton to the internal phonon modes\n(i.e., the C-C bond oscillations), while the longer time dynamics correspond to\ndamping of these phonon modes by the external dissipation. Associated with\nthese time scales, we investigate the following processes that are indicative\nof the system relaxing onto the emissive chromophores of the polymer: 1)\nExciton-polaron formation occurs on an ultra-fast time scale, with the\nassociated exciton-phonon correlations present within half a vibrational time\nperiod of the C-C bond oscillations. 2) Exciton decoherence is driven by the\ndecay in the vibrational overlaps associated with exciton-polaron formation,\noccurring on the same time scale. 3) Exciton density localization is driven by\nthe external dissipation, arising from `wavefunction collapse' occurring as a\nresult of the system-environment interactions. Finally, we show how\nfluorescence anisotropy measurements can be used to investigate the exciton\ndecoherence process during the relaxation dynamics.\n",
"title": "Ultra-Fast Relaxation, Decoherence and Localization of Photoexcited States in $π$-Conjugated Polymers: A TEBD Study"
}
| null | null | null | null | true | null |
11011
| null |
Default
| null | null |
null |
{
"abstract": " Many deployed learned models are black boxes: given input, returns output.\nInternal information about the model, such as the architecture, optimisation\nprocedure, or training data, is not disclosed explicitly as it might contain\nproprietary information or make the system more vulnerable. This work shows\nthat such attributes of neural networks can be exposed from a sequence of\nqueries. This has multiple implications. On the one hand, our work exposes the\nvulnerability of black-box neural networks to different types of attacks -- we\nshow that the revealed internal information helps generate more effective\nadversarial examples against the black box model. On the other hand, this\ntechnique can be used for better protection of private content from automatic\nrecognition models using adversarial examples. Our paper suggests that it is\nactually hard to draw a line between white box and black box models.\n",
"title": "Towards Reverse-Engineering Black-Box Neural Networks"
}
| null | null | null | null | true | null |
11012
| null |
Default
| null | null |
null |
{
"abstract": " We use reinforcement learning (RL) to learn dexterous in-hand manipulation\npolicies which can perform vision-based object reorientation on a physical\nShadow Dexterous Hand. The training is performed in a simulated environment in\nwhich we randomize many of the physical properties of the system like friction\ncoefficients and an object's appearance. Our policies transfer to the physical\nrobot despite being trained entirely in simulation. Our method does not rely on\nany human demonstrations, but many behaviors found in human manipulation emerge\nnaturally, including finger gaiting, multi-finger coordination, and the\ncontrolled use of gravity. Our results were obtained using the same distributed\nRL system that was used to train OpenAI Five. We also include a video of our\nresults: this https URL\n",
"title": "Learning Dexterous In-Hand Manipulation"
}
| null | null | null | null | true | null |
11013
| null |
Default
| null | null |
null |
{
"abstract": " New index transforms with Weber type kernels, consisting of products of\nBessel functions of the first and second kind are investigated. Mapping\nproperties and inversion formulas are established for these transforms in\nLebesgue spaces. The results are applied to solve a boundary value problem on\nthe wedge for a fourth order partial differential equation.\n",
"title": "Index transforms with Weber type kernels"
}
| null | null | null | null | true | null |
11014
| null |
Default
| null | null |
null |
{
"abstract": " Many conventional statistical procedures are extremely sensitive to seemingly\nminor deviations from modeling assumptions. This problem is exacerbated in\nmodern high-dimensional settings, where the problem dimension can grow with and\npossibly exceed the sample size. We consider the problem of robust estimation\nof sparse functionals, and provide a computationally and statistically\nefficient algorithm in the high-dimensional setting. Our theory identifies a\nunified set of deterministic conditions under which our algorithm guarantees\naccurate recovery. By further establishing that these deterministic conditions\nhold with high-probability for a wide range of statistical models, our theory\napplies to many problems of considerable interest including sparse mean and\ncovariance estimation; sparse linear regression; and sparse generalized linear\nmodels.\n",
"title": "Computationally Efficient Robust Estimation of Sparse Functionals"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
11015
| null |
Validated
| null | null |
null |
{
"abstract": " With the increasing interest in the use of millimeter wave bands for 5G\ncellular systems comes renewed interest in resource sharing. Properties of\nmillimeter wave bands such as massive bandwidth, highly directional antennas,\nhigh penetration loss, and susceptibility to shadowing, suggest technical\nadvantages to spectrum and infrastructure sharing in millimeter wave cellular\nnetworks. However, technical advantages do not necessarily translate to\nincreased profit for service providers, or increased consumer surplus. In this\npaper, detailed network simulations are used to better understand the economic\nimplications of resource sharing in a vertically differentiated duopoly market\nfor cellular service. The results suggest that resource sharing is less often\nprofitable for millimeter wave service providers compared to microwave cellular\nservice providers, and does not necessarily increase consumer surplus.\n",
"title": "Resource Sharing Among mmWave Cellular Service Providers in a Vertically Differentiated Duopoly"
}
| null | null | null | null | true | null |
11016
| null |
Default
| null | null |
null |
{
"abstract": " Recent results by Alagic and Russell have given some evidence that the\nEven-Mansour cipher may be secure against quantum adversaries with quantum\nqueries, if considered over other groups than $(\\mathbb{Z}/2)^n$. This prompts\nthe question as to whether or not other classical schemes may be generalized to\narbitrary groups and whether classical results still apply to those generalized\nschemes.\nIn this paper, we generalize the Even-Mansour cipher and the Feistel cipher.\nWe show that Even and Mansour's original notions of secrecy are obtained on a\none-key, group variant of the Even-Mansour cipher. We generalize the result by\nKilian and Rogaway, that the Even-Mansour cipher is pseudorandom, to super\npseudorandomness, also in the one-key, group case. Using a Slide Attack we\nmatch the bound found above. After generalizing the Feistel cipher to arbitrary\ngroups we resolve an open problem of Patel, Ramzan, and Sundaram by showing\nthat the $3$-round Feistel cipher over an arbitrary group is not super\npseudorandom. Finally, we generalize a result by Gentry and Ramzan showing that\nthe Even-Mansour cipher can be implemented using the Feistel cipher as the\npublic permutation. In this last result, we also consider the one-key case over\na group and generalize their bound.\n",
"title": "How to Generate Pseudorandom Permutations Over Other Groups: Even-Mansour and Feistel Revisited"
}
| null | null | null | null | true | null |
11017
| null |
Default
| null | null |
null |
{
"abstract": " We examine the asymptotics of the spectral counting function of a compact\nRiemannian manifold by V.G.~Avakumovic \\cite{Avakumovic} and L.~Hörmander\n\\cite{Hormander-eigen} and show that for the scale of orthogonal and unitary\ngroups ${\\bf SO}(N)$, ${\\bf SU}(N)$, ${\\bf U}(N)$ and ${\\bf Spin}(N)$ it is not\nsharp. While for negative sectional curvature improvements are possible and\nknown, {\\it cf.} e.g., J.J.~Duistermaat $\\&$ V.~Guillemin \\cite{Duist-Guill},\nhere, we give sharp and contrasting examples in the positive Ricci curvature\ncase [non-negative for ${\\bf U}(N)$]. Furthermore here the improvements are\nsharp and quantitative relating to the dimension and {\\it rank} of the group.\nWe discuss the implications of these results on the closely related problem of\nclosed geodesics and the length spectrum.\n",
"title": "On Weyl's asymptotics and remainder term for the orthogonal and unitary groups"
}
| null | null |
[
"Mathematics"
] | null | true | null |
11018
| null |
Validated
| null | null |
null |
{
"abstract": " Marchenko redatuming is a novel scheme used to retrieve up- and down-going\nGreen's functions in an unknown medium. Marchenko equations are based on\nreciprocity theorems and are derived on the assumption of the existence of so\ncalled focusing functions, i.e. functions which exhibit time-space focusing\nproperties once injected in the subsurface. In contrast to interferometry but\nsimilarly to standard migration methods, Marchenko redatuming only requires an\nestimate of the direct wave from the virtual source (or to the virtual\nreceiver), illumination from only one side of the medium, and no physical\nsources (or receivers) inside the medium. In this contribution we consider a\ndifferent time-focusing condition within the frame of Marchenko redatuming and\nshow how this can lead to the retrieval of virtual plane-wave responses, thus\nallowing multiple-free imaging using only a 1 dimensional sampling of the\ntargeted model. The potential of the new method is demonstrated on a 2D\nsynthetic model.\n",
"title": "Virtual plane-wave imaging via Marchenko redatuming"
}
| null | null | null | null | true | null |
11019
| null |
Default
| null | null |
null |
{
"abstract": " Excitons and plasmons are the two most fundamental types of collective\nelectronic excitations occurring in solids. Traditionally, they have been\nstudied separately using bulk techniques that probe their average energetic\nstructure over large spatial regions. However, as the dimensions of materials\nand devices continue to shrink, it becomes crucial to understand how these\nexcitations depend on local variations in the crystal- and chemical structure\non the atomic scale. Here we use monochromated low-loss\nscanning-transmission-electron-microscopy electron-energy-loss (LL-STEM-EEL)\nspectroscopy, providing the best simultaneous energy and spatial resolution\nachieved to-date to unravel the full set of electronic excitations in few-layer\nMoS2 nanosheets over a wide energy range. Using first-principles many-body\ncalculations we confirm the excitonic nature of the peaks at ~2eV and ~3eV in\nthe experimental EEL spectrum and the plasmonic nature of higher energy-loss\npeaks. We also rationalise the non-trivial dependence of the EEL spectrum on\nbeam and sample geometry such as the number of atomic layers and distance to\nsteps and edges. Moreover, we show that the excitonic features are dominated by\nthe long wavelength (q=0) components of the probing field, while the plasmonic\nfeatures are sensitive to a much broader range of q-vectors, indicating a\nqualitative difference in the spatial character of the two types of collective\nexcitations. Our work provides a template protocol for mapping the local nature\nof electronic excitations that open new possibilities for studying\nphoto-absorption and energy transfer processes on a nanometer scale.\n",
"title": "Probing the local nature of excitons and plasmons in few-layer MoS2"
}
| null | null |
[
"Physics"
] | null | true | null |
11020
| null |
Validated
| null | null |
null |
{
"abstract": " We investigate the evolution of the flavour composition of the cosmic\nneutrino background from neutrino decoupling until today. The decoherence of\nneutrino mass states is described by means of Lindblad operators. Decoherence\ngoes along with the increase of neutrino family entropy, which we obtain as a\nfunction of initial spectral distortions, mixing angles and CP-violation phase.\nWe also present the expected flavour composition of the cosmic neutrino\nbackground after decoherence is completed. Decoherence is proposed to happen\nafter the two heaviest neutrino mass states become non-relativistic. We discuss\nhow the associated increase of entropy could be observed (in principle). The\nphysics of two- or three-flavour oscillation of cosmological neutrinos\nresembles in many aspects two- or three-level systems in atomic clocks, which\nwere recently proposed by Weinberg for the study of decoherence phenomena.\n",
"title": "Flavour composition and entropy increase of cosmological neutrinos after decoherence"
}
| null | null | null | null | true | null |
11021
| null |
Default
| null | null |
null |
{
"abstract": " We analyse the limiting behavior of the eigenvalue and singular value\ndistribution for random convolution operators on large (not necessarily\nAbelian) groups, extending the results by M. Meckes for the Abelian case. We\nshow that for regular sequences of groups the limiting distribution of\neigenvalues (resp. singular values) is a mixture of eigenvalue (resp. singular\nvalue) distributions of Ginibre matrices with the directing measure being\nrelated to the limiting behavior of the Plancherel measure of the sequence of\ngroups. In particular for the sequence of symmetric groups, the limiting\ndistributions are just the circular and quarter circular laws, whereas e.g. for\nthe dihedral groups the limiting distributions have unbounded supports but are\ndifferent than in the Abelian case.\nWe also prove that under additional assumptions on the sequence of groups (in\nparticular for symmetric groups of increasing order) families of stochastically\nindependent random projection operators converge in moments to free circular\nelements.\nFinally, in the Gaussian case we provide Central Limit Theorems for linear\neigenvalue statistics.\n",
"title": "Random non-Abelian G-circulant matrices. Spectrum of random convolution operators on large finite groups"
}
| null | null | null | null | true | null |
11022
| null |
Default
| null | null |
null |
{
"abstract": " Active communication between robots and humans is essential for effective\nhuman-robot interaction. To accomplish this objective, Cloud Robotics (CR) was\nintroduced to make robots enhance their capabilities. It enables robots to\nperform extensive computations in the cloud by sharing their outcomes. Outcomes\ninclude maps, images, processing power, data, activities, and other robot\nresources. But due to the colossal growth of data and traffic, CR suffers from\nserious latency issues. Therefore, it is unlikely to scale a large number of\nrobots particularly in human-robot interaction scenarios, where responsiveness\nis paramount. Furthermore, other issues related to security such as privacy\nbreaches and ransomware attacks can increase. To address these problems, in\nthis paper, we have envisioned the next generation of social robotic\narchitectures based on Fog Robotics (FR) that inherits the strengths of Fog\nComputing to augment the future social robotic systems. These new architectures\ncan escalate the dexterity of robots by shoving the data closer to the robot.\nAdditionally, they can ensure that human-robot interaction is more responsive\nby resolving the problems of CR. Moreover, experimental results are further\ndiscussed by considering a scenario of FR and latency as a primary factor\ncomparing to CR models.\n",
"title": "Fog Robotics for Efficient, Fluent and Robust Human-Robot Interaction"
}
| null | null | null | null | true | null |
11023
| null |
Default
| null | null |
null |
{
"abstract": " We present the results of our investigation of the star-forming potential in\nthe Perseus star-forming complex. We build on previous starless core,\nprotostellar core, and young stellar object (YSO) catalogs from Spitzer,\nHerschel, and SCUBA observations in the literature. We place the cores and YSOs\nwithin seven star-forming clumps based on column densities greater than 5x10^21\ncm^-2. We calculate the mean density and free-fall time for 69 starless cores\nas 5.55x10^-19 gcm^-3 and 0.1 Myr,respectively, and we estimate the star\nformation rate for the near future as 150 Msun Myr^-1. According to Bonnor\nEbert stability analysis, we find that majority of starless cores in Perseus\nare unstable. Broadly, these cores can collapse to form the next generation of\nstars. We found a relation between starless cores and YSOs, where the numbers\nof young protostars (Class 0 + Class I) are similar to the numbers of starless\ncores. This similarity, which shows a one-to-one relation, suggests that these\nstarless cores may form the next generation of stars with approximately the\nsame formation rate as the current generation, as identified by the Class 0 and\nClass I protostars. It follows that if such a relation between starless cores\nand any YSO stage exists, the SFR values of these two populations must be\nnearly constant. In brief, we propose that this one-to-one relation is an\nimportant factor in better understanding the star formation process within a\ncloud.\n",
"title": "An Estimation of the Star Formation Rate in the Perseus Complex"
}
| null | null |
[
"Physics"
] | null | true | null |
11024
| null |
Validated
| null | null |
null |
{
"abstract": " Introduction to deep neural networks and their history.\n",
"title": "Deep Neural Networks - A Brief History"
}
| null | null | null | null | true | null |
11025
| null |
Default
| null | null |
null |
{
"abstract": " Commercial photon-counting modules based on actively quenched solid-state\navalanche photodiode sensors are used in a wide variety of applications.\nManufacturers characterize their detectors by specifying a small set of\nparameters, such as detection efficiency, dead time, dark counts rate,\nafterpulsing probability and single-photon arrival-time resolution (jitter).\nHowever, they usually do not specify the range of conditions over which these\nparameters are constant or present a sufficient description of the\ncharacterization process. In this work, we perform a few novel tests on two\ncommercial detectors and identify an additional set of imperfections that must\nbe specified to sufficiently characterize their behavior. These include\nrate-dependence of the dead time and jitter, detection delay shift, and\n\"twilighting.\" We find that these additional non-ideal behaviors can lead to\nunexpected effects or strong deterioration of the performance of a system using\nthese devices. We explain their origin by an in-depth analysis of the active\nquenching process. To mitigate the effects of these imperfections, a\ncustom-built detection system is designed using a novel active quenching\ncircuit. Its performance is compared against two commercial detectors in a fast\nquantum key distribution system with hyper-entangled photons and a random\nnumber generator.\n",
"title": "An advanced active quenching circuit for ultra-fast quantum cryptography"
}
| null | null | null | null | true | null |
11026
| null |
Default
| null | null |
null |
{
"abstract": " The Shannon entropy in the atomic, molecular and chemical physics context is\npresented by using as test cases the hydrogenic-like atoms $H_c$, ${He_c}^+$\nand ${Li_c}^{2+}$ confined by an impenetrable spherical box. Novel expressions\nfor entropic uncertainty relation and Shannon entropies $S_r$ and $S_p$ are\nproposed to ensure their physical dimensionless characteristic. The electronic\nground state energy and the quantities $S_r$, $S_p$ and $S_t$ are calculated\nfor the hydrogenic-like atoms to different confinement radii by using a\nvariational method. The global behavior of these quantities and different\nconjectures are analyzed. The results are compared, when available, with those\npreviously published.\n",
"title": "Shannon entropy: a study of confined hydrogenic-like atoms"
}
| null | null | null | null | true | null |
11027
| null |
Default
| null | null |
null |
{
"abstract": " Thanks to modern sky surveys, over twenty stellar streams and overdensity\nstructures have been discovered in the halo of the Milky Way. In this paper, we\npresent an analysis of spectroscopic observations of individual stars from one\nsuch structure, \"A13\", first identified as an overdensity using the M giant\ncatalog from the Two Micron All-Sky Survey. Our spectroscopic observations show\nthat stars identified with A13 have a velocity dispersion of $\\lesssim$ 40\n$\\mathrm{km~s^{-1}}$, implying that it is a genuine coherent structure rather\nthan a chance super-position of random halo stars. From its position on the\nsky, distance ($\\sim$15~kpc heliocentric), and kinematical properties, A13 is\nlikely to be an extension of another low Galactic latitude substructure -- the\nGalactic Anticenter Stellar Structure (also known as the Monoceros Ring) --\ntowards smaller Galactic longitude and farther distance. Furthermore, the\nkinematics of A13 also connect it with another structure in the southern\nGalactic hemisphere -- the Triangulum-Andromeda overdensity. We discuss these\nthree connected structures within the context of a previously proposed scenario\nthat one or all of these features originate from the disk of the Milky Way.\n",
"title": "Exploring Halo Substructure with Giant Stars. XV. Discovery of a Connection between the Monoceros Ring and the Triangulum-Andromeda Overdensity?"
}
| null | null | null | null | true | null |
11028
| null |
Default
| null | null |
null |
{
"abstract": " The fitness of a species determines its abundance and survival in an\necosystem. At the same time, species take up resources for growth, so their\nabundance affects the availability of resources in an ecosystem. We show here\nthat such species-resource coupling can be used to assign a quantitative metric\nfor fitness to each species. This fitness metric also allows for the modeling\nof drift in species composition, and hence ecosystem evolution through\nspeciation and adaptation. Our results provide a foundation for an entirely\ncomputational exploration of evolutionary ecosystem dynamics on any length or\ntime scale. For example, we can evolve ecosystem dynamics even by initiating\ndynamics out of a single primordial ancestor and show that there exists a well\ndefined ecosystem-averaged fitness dynamics that is resilient against resource\nshocks.\n",
"title": "Universal fitness dynamics through an adaptive resource utilization model"
}
| null | null |
[
"Quantitative Biology"
] | null | true | null |
11029
| null |
Validated
| null | null |
null |
{
"abstract": " We first derive a general integral-turnpike property around a set for\ninfinite-dimensional non-autonomous optimal control problems with any possible\nterminal state constraints, under some appropriate assumptions. Roughly\nspeaking, the integral-turnpike property means that the time average of the\ndistance from any optimal trajectory to the turnpike set con- verges to zero,\nas the time horizon tends to infinity. Then, we establish the measure-turnpike\nproperty for strictly dissipative optimal control systems, with state and\ncontrol constraints. The measure-turnpike property, which is slightly stronger\nthan the integral-turnpike property, means that any optimal (state and control)\nsolution remains essentially, along the time frame, close to an optimal\nsolution of an associated static optimal control problem, except along a subset\nof times that is of small relative Lebesgue measure as the time horizon is\nlarge. Next, we prove that strict strong duality, which is a classical notion\nin optimization, implies strict dissipativity, and measure-turnpike. Finally,\nwe conclude the paper with several comments and open problems.\n",
"title": "Integral and measure-turnpike properties for infinite-dimensional optimal control systems"
}
| null | null | null | null | true | null |
11030
| null |
Default
| null | null |
null |
{
"abstract": " We study the vortex patch problem for $2d-$stratified Navier-Stokes system.\nWe aim at extending several results obtained in\n\\cite{ad,danchinpoche,hmidipoche} for standard Euler and Navier-Stokes systems.\nWe shall deal with smooth initial patches and establish global strong estimates\nuniformly with respect to the viscosity in the spirit of \\cite{HZ-poche,\nZ-poche}. This allows to prove the convergence of the viscous solutions towards\nthe inviscid one. In the setting of a Rankine vortex, we show that the rate of\nconvergence for the vortices is optimal in $L^p$ space and is given by $(\\mu\nt)^{\\frac{1}{2p}}$. This generalizes the result of \\cite{ad} obtained for $L^2$\nspace.\n",
"title": "Optimal rate of convergence in Stratified Boussinesq system"
}
| null | null | null | null | true | null |
11031
| null |
Default
| null | null |
null |
{
"abstract": " A one-dimensional, unsteady nozzle flow is modelled to identify the sources\nof indirect noise in multicomponent gases. First, from non-equilibrium\nthermodynamics relations, it is shown that a compositional inhomogeneity\nadvected in an accelerating flow is a source of sound induced by\ninhomogeneities in the mixture (i) chemical potentials and (ii) specific heat\ncapacities. Second, it is shown that the acoustic, entropy and compositional\nlinear perturbations evolve independently from each other and they become\ncoupled through mean-flow gradients and/or at the boundaries. Third, the\nequations are cast in invariant formulation and a mathematical solution is\nfound by asymptotic expansion of path-ordered integrals with an infinite radius\nof convergence. Finally, the transfer functions are calculated for a supersonic\nnozzle with finite spatial extent perturbed by a methane-air compositional\ninhomogeneity. The proposed framework will help identify and quantify the\nsources of sound in nozzles with relevance, for example, to aeronautical gas\nturbines.\n",
"title": "On indirect noise in multicomponent nozzle flows"
}
| null | null | null | null | true | null |
11032
| null |
Default
| null | null |
null |
{
"abstract": " In this short note we provide an analytical formula for the conditional\ncovariance matrices of the elliptically distributed random vectors, when the\nconditioning is based on the values of any linear combination of the marginal\nrandom variables. We show that one could introduce the univariate invariant\ndepending solely on the conditioning set, which greatly simplifies the\ncalculations. As an application, we show that one could define uniquely defined\nquantile-based sets on which conditional covariance matrices must be equal to\neach other if only the vector is multivariate normal. The similar results are\nobtained for conditional correlation matrices of the general elliptic case.\n",
"title": "A note on conditional covariance matrices for elliptical distributions"
}
| null | null | null | null | true | null |
11033
| null |
Default
| null | null |
null |
{
"abstract": " Channel-reciprocity based key generation (CRKG) has gained significant\nimportance as it has recently been proposed as a potential lightweight security\nsolution for IoT devices. However, the impact of the attacker's position in\nclose range has only rarely been evaluated in practice, posing an open research\nproblem about the security of real-world realizations. Furthermore, this would\nfurther bridge the gap between theoretical channel models and their\npractice-oriented realizations. For security metrics, we utilize\ncross-correlation, mutual information, and a lower bound on secret-key\ncapacity. We design a practical setup of three parties such that the channel\nstatistics, although based on joint randomness, are always reproducible. We run\nexperiments to obtain channel states and evaluate the aforementioned metrics\nfor the impact of an attacker depending on his position. It turns out the\nattacker himself affects the outcome, which has not been adequately regarded\nyet in standard channel models.\n",
"title": "The Passive Eavesdropper Affects my Channel: Secret-Key Rates under Real-World Conditions (Extended Version)"
}
| null | null | null | null | true | null |
11034
| null |
Default
| null | null |
null |
{
"abstract": " A rational projective plane ($\\mathbb{QP}^2$) is a simply connected, smooth,\nclosed manifold $M$ such that $H^*(M;\\mathbb{Q}) \\cong\n\\mathbb{Q}[\\alpha]/\\langle \\alpha^3 \\rangle$. An open problem is to classify\nthe dimensions at which such a manifold exists. The Barge-Sullivan rational\nsurgery realization theorem provides necessary and sufficient conditions that\ninclude the Hattori-Stong integrality conditions on the Pontryagin numbers. In\nthis article, we simplify these conditions and combine them with the signature\nequation to give a single quadratic residue equation that determines whether a\ngiven dimension supports a $\\mathbb{QP}^2$. We then confirm existence of a\n$\\mathbb{QP}^2$ in two new dimensions and prove several non-existence results\nusing factorizations of numerators of divided Bernoulli numbers. We also\nresolve the existence question in the Spin case, and we discuss existence\nresults for the more general class of rational projective spaces.\n",
"title": "On dimensions supporting a rational projective plane"
}
| null | null | null | null | true | null |
11035
| null |
Default
| null | null |
null |
{
"abstract": " Progress in probabilistic generative models has accelerated, developing\nricher models with neural architectures, implicit densities, and with scalable\nalgorithms for their Bayesian inference. However, there has been limited\nprogress in models that capture causal relationships, for example, how\nindividual genetic factors cause major human diseases. In this work, we focus\non two challenges in particular: How do we build richer causal models, which\ncan capture highly nonlinear relationships and interactions between multiple\ncauses? How do we adjust for latent confounders, which are variables\ninfluencing both cause and effect and which prevent learning of causal\nrelationships? To address these challenges, we synthesize ideas from causality\nand modern probabilistic modeling. For the first, we describe implicit causal\nmodels, a class of causal models that leverages neural architectures with an\nimplicit density. For the second, we describe an implicit causal model that\nadjusts for confounders by sharing strength across examples. In experiments, we\nscale Bayesian inference on up to a billion genetic measurements. We achieve\nstate of the art accuracy for identifying causal factors: we significantly\noutperform existing genetics methods by an absolute difference of 15-45.3%.\n",
"title": "Implicit Causal Models for Genome-wide Association Studies"
}
| null | null | null | null | true | null |
11036
| null |
Default
| null | null |
null |
{
"abstract": " Vector quantization aims to form new vectors/matrices with shared values\nclose to the original. It could compress data with acceptable information loss\nand could be of great usefulness in areas like Image Processing, Pattern\nRecognition, and Machine Learning. In this paper, the problem of vector\nquantization is examined from a new perspective, namely sparse least square\noptimization. Specifically, inspired by the property of sparsity of Lasso, a\nnovel quantization algorithm based on $l_1$ least square is proposed and\nimplemented. Similar schemes with $l_1 + l_2$ combination penalization and\n$l_0$ regularization are simultaneously proposed. In addition, to produce\nquantization results with given amount of quantized values(instead of\npenalization coefficient $\\lambda$), this paper proposed an iterative sparse\nleast square method and a cluster-based least square quantization method. It is\nalso noticed that the later method is mathematically equivalent to an improved\nversion of the existed clustering-based quantization algorithm, although the\ntwo algorithms originated from different intuitions. The algorithms proposed\nwere tested under three scenarios of data and their computational performance,\nincluding information loss, time consumption and the distribution of the value\nof sparse vectors were compared and analyzed. The paper offers a new\nperspective to probe the area of vector quantization, and the algorithms\nproposed could offer better performance especially when the required\npost-quantization value amounts are not on a tiny scale.\n",
"title": "Vector Quantization as Sparse Least Square Optimization"
}
| null | null | null | null | true | null |
11037
| null |
Default
| null | null |
null |
{
"abstract": " The effects of the spatial scale on the results of the optimisation of\ntransmission and generation capacity in Europe are quantified under a 95% CO2\nreduction compared to 1990 levels, interpolating between one-node-per-country\nsolutions and many-nodes-per-country. The trade-offs that come with higher\nspatial detail between better exposure of transmission bottlenecks,\nexploitation of sites with good renewable resources (particularly wind power)\nand computational limitations are discussed. It is shown that solutions with no\ngrid expansion beyond today's capacities are only around 20% more expensive\nthan with cost-optimal grid expansion.\n",
"title": "The role of spatial scale in joint optimisations of generation and transmission for European highly renewable scenarios"
}
| null | null | null | null | true | null |
11038
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we consider regression problems with one-hidden-layer neural\nnetworks (1NNs). We distill some properties of activation functions that lead\nto $\\mathit{local~strong~convexity}$ in the neighborhood of the ground-truth\nparameters for the 1NN squared-loss objective. Most popular nonlinear\nactivation functions satisfy the distilled properties, including rectified\nlinear units (ReLUs), leaky ReLUs, squared ReLUs and sigmoids. For activation\nfunctions that are also smooth, we show $\\mathit{local~linear~convergence}$\nguarantees of gradient descent under a resampling rule. For homogeneous\nactivations, we show tensor methods are able to initialize the parameters to\nfall into the local strong convexity region. As a result, tensor initialization\nfollowed by gradient descent is guaranteed to recover the ground truth with\nsample complexity $ d \\cdot \\log(1/\\epsilon) \\cdot \\mathrm{poly}(k,\\lambda )$\nand computational complexity $n\\cdot d \\cdot \\mathrm{poly}(k,\\lambda) $ for\nsmooth homogeneous activations with high probability, where $d$ is the\ndimension of the input, $k$ ($k\\leq d$) is the number of hidden nodes,\n$\\lambda$ is a conditioning property of the ground-truth parameter matrix\nbetween the input layer and the hidden layer, $\\epsilon$ is the targeted\nprecision and $n$ is the number of samples. To the best of our knowledge, this\nis the first work that provides recovery guarantees for 1NNs with both sample\ncomplexity and computational complexity $\\mathit{linear}$ in the input\ndimension and $\\mathit{logarithmic}$ in the precision.\n",
"title": "Recovery Guarantees for One-hidden-layer Neural Networks"
}
| null | null | null | null | true | null |
11039
| null |
Default
| null | null |
null |
{
"abstract": " With the rising number of interconnected devices and sensors, modeling\ndistributed sensor networks is of increasing interest. Recurrent neural\nnetworks (RNN) are considered particularly well suited for modeling sensory and\nstreaming data. When predicting future behavior, incorporating information from\nneighboring sensor stations is often beneficial. We propose a new RNN based\narchitecture for context specific information fusion across multiple spatially\ndistributed sensor stations. Hereby, latent representations of multiple local\nmodels, each modeling one sensor station, are jointed and weighted, according\nto their importance for the prediction. The particular importance is assessed\ndepending on the current context using a separate attention function. We\ndemonstrate the effectiveness of our model on three different real-world sensor\nnetwork datasets.\n",
"title": "Attention-based Information Fusion using Multi-Encoder-Decoder Recurrent Neural Networks"
}
| null | null | null | null | true | null |
11040
| null |
Default
| null | null |
null |
{
"abstract": " The widespread availability of GPS information in everyday devices such as\ncars, smartphones and smart watches make it possible to collect large amount of\ngeospatial trajectory information. A particularly important, yet technically\nchallenging, application of this data is to identify the underlying road\nnetwork and keep it updated under various changes. In this paper, we propose\nefficient algorithms that can generate accurate maps in both batch and online\nsettings. Our algorithms utilize techniques from graph spanners so that they\nproduce maps can effectively handle a wide variety of road and intersection\nshapes. We conduct a rigorous evaluation of our algorithms over two real-world\ndatasets and under a wide variety of performance metrics. Our experiments show\na significant improvement over prior work. In particular, we observe an\nincrease in Biagioni f-score of up to 20% when compared to the state of the art\nwhile reducing the execution time by an order of magnitude. We also make our\nsource code open source for reproducibility and enable other researchers to\nbuild on our work.\n",
"title": "Kharita: Robust Map Inference using Graph Spanners"
}
| null | null |
[
"Computer Science"
] | null | true | null |
11041
| null |
Validated
| null | null |
null |
{
"abstract": " Information extraction and user intention identification are central topics\nin modern query understanding and recommendation systems. In this paper, we\npropose DeepProbe, a generic information-directed interaction framework which\nis built around an attention-based sequence to sequence (seq2seq) recurrent\nneural network. DeepProbe can rephrase, evaluate, and even actively ask\nquestions, leveraging the generative ability and likelihood estimation made\npossible by seq2seq models. DeepProbe makes decisions based on a derived\nuncertainty (entropy) measure conditioned on user inputs, possibly with\nmultiple rounds of interactions. Three applications, namely a rewritter, a\nrelevance scorer and a chatbot for ad recommendation, were built around\nDeepProbe, with the first two serving as precursory building blocks for the\nthird. We first use the seq2seq model in DeepProbe to rewrite a user query into\none of standard query form, which is submitted to an ordinary recommendation\nsystem. Secondly, we evaluate DeepProbe's seq2seq model-based relevance\nscoring. Finally, we build a chatbot prototype capable of making active user\ninteractions, which can ask questions that maximize information gain, allowing\nfor a more efficient user intention idenfication process. We evaluate first two\napplications by 1) comparing with baselines by BLEU and AUC, and 2) human judge\nevaluation. Both demonstrate significant improvements compared with current\nstate-of-the-art systems, proving their values as useful tools on their own,\nand at the same time laying a good foundation for the ongoing chatbot\napplication.\n",
"title": "DeepProbe: Information Directed Sequence Understanding and Chatbot Design via Recurrent Neural Networks"
}
| null | null | null | null | true | null |
11042
| null |
Default
| null | null |
null |
{
"abstract": " While cardiovascular diseases (CVDs) are prevalent across economic strata,\nthe economically disadvantaged population is disproportionately affected due to\nthe high cost of traditional CVD management. Accordingly, developing an\nultra-low-cost alternative, affordable even to groups at the bottom of the\neconomic pyramid, has emerged as a societal imperative. Against this backdrop,\nwe propose an inexpensive yet accurate home-based electrocardiogram(ECG)\nmonitoring service. Specifically, we seek to provide point-of-care monitoring\nof premature ventricular contractions (PVCs), high frequency of which could\nindicate the onset of potentially fatal arrhythmia. Note that a traditional\ntelecardiology system acquires the ECG, transmits it to a professional\ndiagnostic centre without processing, and nearly achieves the diagnostic\naccuracy of a bedside setup, albeit at high bandwidth cost. In this context, we\naim at reducing cost without significantly sacrificing reliability. To this\nend, we develop a dictionary-based algorithm that detects with high sensitivity\nthe anomalous beats only which are then transmitted. We further compress those\ntransmitted beats using class-specific dictionaries subject to suitable\nreconstruction/diagnostic fidelity. Such a scheme would not only reduce the\noverall bandwidth requirement, but also localising anomalous beats, thereby\nreducing physicians' burden. Finally, using Monte Carlo cross validation on\nMIT/BIH arrhythmia database, we evaluate the performance of the proposed\nsystem. In particular, with a sensitivity target of at most one undetected PVC\nin one hundred beats, and a percentage root mean squared difference less than\n9% (a clinically acceptable level of fidelity), we achieved about 99.15%\nreduction in bandwidth cost, equivalent to 118-fold savings over traditional\ntelecardiology.\n",
"title": "Dictionary-based Monitoring of Premature Ventricular Contractions: An Ultra-Low-Cost Point-of-Care Service"
}
| null | null | null | null | true | null |
11043
| null |
Default
| null | null |
null |
{
"abstract": " This paper addresses the question of emotion classification. The task\nconsists in predicting emotion labels (taken among a set of possible labels)\nbest describing the emotions contained in short video clips. Building on a\nstandard framework -- lying in describing videos by audio and visual features\nused by a supervised classifier to infer the labels -- this paper investigates\nseveral novel directions. First of all, improved face descriptors based on 2D\nand 3D Convo-lutional Neural Networks are proposed. Second, the paper explores\nseveral fusion methods, temporal and multimodal, including a novel hierarchical\nmethod combining features and scores. In addition, we carefully reviewed the\ndifferent stages of the pipeline and designed a CNN architecture adapted to the\ntask; this is important as the size of the training set is small compared to\nthe difficulty of the problem, making generalization difficult. The so-obtained\nmodel ranked 4th at the 2017 Emotion in the Wild challenge with the accuracy of\n58.8 %.\n",
"title": "Temporal Multimodal Fusion for Video Emotion Classification in the Wild"
}
| null | null | null | null | true | null |
11044
| null |
Default
| null | null |
null |
{
"abstract": " Charge transfer among individual atoms in a molecule is the key concept in\nthe modern electronic theory of chemical bonding. In this work, we defined an\natomic region between two atoms by Slater orbital exponents of valence\nelectrons and suggested a method for analytical calculation of charge\npenetration between all atoms in a molecule. Computation of charge penetration\namount is self-consistently performed until each orbital exponent converges to\nits certain values respectively. Charge penetration matrix was calculated for\nethylene and MgO, and bonding charge and its distribution were analyzed by\nusing the charge penetration matrix and the orbital exponents under the bonding\nstate. These results were compared with those by density function method and\nshowed that this method is a simple and direct method to obtain bonding charge\ndistribution of molecule from atomic orbital functions.\n",
"title": "Bonding charge distribution analysis of molecule by computation of interatomic charge penetration"
}
| null | null | null | null | true | null |
11045
| null |
Default
| null | null |
null |
{
"abstract": " We study how a single value of the shatter function of a set system restricts\nits asymptotic growth. Along the way, we refute a conjecture of Bondy and\nHajnal which generalizes Sauer's Lemma.\n",
"title": "Shatter functions with polynomial growth rates"
}
| null | null | null | null | true | null |
11046
| null |
Default
| null | null |
null |
{
"abstract": " Consider the moduli space of framed flat $U(2)$ connections with fixed odd\ndeterminant over a surface. Newstead combined some fundamental facts about this\nmoduli space with the Mayer-Vietoris sequence to compute its betti numbers over\nany field not of characteristic two. We adapt his method in characteristic two\nto produce conjectural recursive formulae for the mod two betti numbers of the\nframed moduli space which we partially verify. We also discuss the interplay\nwith the mod two cohomology ring structure of the unframed moduli space.\n",
"title": "On Newstead's Mayer-Vietoris argument in characteristic 2"
}
| null | null | null | null | true | null |
11047
| null |
Default
| null | null |
null |
{
"abstract": " Non-Gaussianities of dynamical origin are disentangled from primordial ones\nusing the formalism of large deviation statistics with spherical collapse\ndynamics. This is achieved by relying on accurate analytical predictions for\nthe one-point probability distribution function (PDF) and the two-point\nclustering of spherically-averaged cosmic densities (sphere bias). Sphere bias\nextends the idea of halo bias to intermediate density environments and voids as\nunderdense regions. In the presence of primordial non-Gaussianity, sphere bias\ndisplays a strong scale dependence relevant for both high and low density\nregions, which is predicted analytically. The statistics of densities in\nspheres are built to model primordial non-Gaussianity via an initial skewness\nwith a scale-dependence that depends on the bispectrum of the underlying model.\nThe analytical formulas with the measured nonlinear dark matter variance as\ninput are successfully tested against numerical simulations. For local\nnon-Gaussianity with a range from $f_{\\rm NL}=-100$ to $+100$ they are found to\nagree within 2\\% or better for densities $\\rho\\in[0.5,3]$ in spheres of radius\n15 Mpc$/h$ down to $z=0.35$. The validity of the large deviation statistics\nformalism is thereby established for all observationally relevant local-type\ndepartures from perfectly Gaussian initial conditions. The corresponding\nestimators for the amplitude of the nonlinear variance $\\sigma_8$ and\nprimordial skewness $f_{\\rm NL}$ are validated using a fiducial joint maximum\nlikelihood experiment. The influence of observational effects and the prospects\nfor a future detection of primordial non-Gaussianity from joint one- and\ntwo-point densities-in-spheres statistics are discussed.\n",
"title": "Hunting high and low: Disentangling primordial and late-time non-Gaussianity with cosmic densities in spheres"
}
| null | null | null | null | true | null |
11048
| null |
Default
| null | null |
null |
{
"abstract": " Measure Theory and Integration is exposed with the clear aim to help\nbeginning learners to perfectly master its essence. In opposition of a delivery\nof the contents in an academic and vertical course, the knowledge is broken\ninto exercises which are left to the learners for solutions. Hints are present\nat any corner to help readers to achieve the solutions. In that way, the\nknowledge is constructed by the readers by summarizing the results of one or a\ngroup of exercises.\nEach chapter is organized into Summary documents which contain the knowledge,\nDiscovery documents which give the learner the opportunity to extract the\nknowledge himself through exercises and into Solution Documents which offer\ndetailed answers for the exercises. Exceptionally, a few number of results (A\nkey lemma related the justification of definition of the integral of a\nnon-negative function, the Caratheodory's theorem and the Lebesgue-Stieljes\nmeasure on $\\mathbb{R}^d$) are presented in appendix documents and given for\nreading in small groups.\nThe full theory is presented in the described way. We highly expect that any\nstudent who goes through the materials, alone or in a small group or under the\nsupervision of an assistant will gain a very solid knowledge in the subject and\nby the way ensure a sound foundation for studying disciplines such as\nProbability Theory, Statistics, Functional Analysis, etc.\nThe materials have been successfully used as such in normal real analysis\nclasses several times.\n",
"title": "Measure Theory and Integration By and For the Learner"
}
| null | null | null | null | true | null |
11049
| null |
Default
| null | null |
null |
{
"abstract": " We propose a contextual-bandit approach for demand side management by\noffering price incentives. More precisely, a target mean consumption is set at\neach round and the mean consumption is modeled as a complex function of the\ndistribution of prices sent and of some contextual variables such as the\ntemperature, weather, and so on. The performance of our strategies is measured\nin quadratic losses through a regret criterion. We offer $\\sqrt{T}$ upper\nbounds on this regret (up to poly-logarithmic terms), for strategies inspired\nby standard strategies for contextual bandits (like LinUCB, Li et al., 2010).\nSimulations on a real data set gathered by UK Power Networks, in which price\nincentives were offered, show that our strategies are effective and may indeed\nmanage demand response by suitably picking the price levels.\n",
"title": "Target Tracking for Contextual Bandits: Application to Demand Side Management"
}
| null | null | null | null | true | null |
11050
| null |
Default
| null | null |
null |
{
"abstract": " We present an overview of techniques for quantizing convolutional neural\nnetworks for inference with integer weights and activations. Per-channel\nquantization of weights and per-layer quantization of activations to 8-bits of\nprecision post-training produces classification accuracies within 2% of\nfloating point networks for a wide variety of CNN architectures. Model sizes\ncan be reduced by a factor of 4 by quantizing weights to 8-bits, even when\n8-bit arithmetic is not supported. This can be achieved with simple, post\ntraining quantization of weights.We benchmark latencies of quantized networks\non CPUs and DSPs and observe a speedup of 2x-3x for quantized implementations\ncompared to floating point on CPUs. Speedups of up to 10x are observed on\nspecialized processors with fixed point SIMD capabilities, like the Qualcomm\nQDSPs with HVX.\nQuantization-aware training can provide further improvements, reducing the\ngap to floating point to 1% at 8-bit precision. Quantization-aware training\nalso allows for reducing the precision of weights to four bits with accuracy\nlosses ranging from 2% to 10%, with higher accuracy drop for smaller\nnetworks.We introduce tools in TensorFlow and TensorFlowLite for quantizing\nconvolutional networks and review best practices for quantization-aware\ntraining to obtain high accuracy with quantized weights and activations. We\nrecommend that per-channel quantization of weights and per-layer quantization\nof activations be the preferred quantization scheme for hardware acceleration\nand kernel optimization. We also propose that future processors and hardware\naccelerators for optimized inference support precisions of 4, 8 and 16 bits.\n",
"title": "Quantizing deep convolutional networks for efficient inference: A whitepaper"
}
| null | null | null | null | true | null |
11051
| null |
Default
| null | null |
null |
{
"abstract": " We present a new experimental approach to investigate the magnetic properties\nof the anisotropic heavy-fermion system YbRh$_2$Si$_2$ as a function of\ncrystallographic orientation. Angle-dependent electron spin resonance (ESR)\nmeasurements are performed at a low temperature of 1.6 K and at an ESR\nfrequency of 4.4 GHz utilizing a superconducting planar microwave resonator in\na $^4$He-cryostat in combination with in-situ sample rotation. The obtained ESR\ng-factor of YbRh$_2$Si$_2$ as a function of the crystallographic angle is\nconsistent with results of previous measurements using conventional ESR\nspectrometers at higher frequencies and fields. Perspectives to implement this\nexperimental approach into a dilution refrigerator and to reach the\nmagnetically ordered phase of YbRh$_2$Si$_2$ are discussed.\n",
"title": "Angle-dependent electron spin resonance of YbRh$_2$Si$_2$ measured with planar microwave resonators and in-situ rotation"
}
| null | null | null | null | true | null |
11052
| null |
Default
| null | null |
null |
{
"abstract": " The positive definite Kohn-Sham kinetic energy(KS-KE) density plays crucial\nrole in designing semilocal meta generalized gradient approximations(meta-GGAs)\nfor low dimensional quantum systems. It has been rigorously shown that near\nnucleus and at the asymptotic region, the KE-KS differ from its von\nWeizsäcker(VW) counterpart as contributions from different orbitals (i.e.,\ns and p orbitals) play important role. This has been explored using two\ndimensional isotropic quantum harmonic oscillator as a test case. Several\nmeta-GGA ingredients with different physical behaviors are also constructed and\nfurther used to design an accurate semilocal functionals at meta-GGA level. In\nthe asymptotic region, a new exchange energy functional is constructed using\nthe meta-GGA ingredients with formally exact properties of the enhancement\nfactor. Also, it has been shown that exact asymptotic behavior of the exchange\nenergy density and potential can be attained by choosing accurately the\nenhancement factor as a functional of meta-GGA ingredients.\n",
"title": "Role of Kohn-Sham Kinetic Energy Density in Designing Asymptotically Correct Semilocal Exchange-Correlation Functionals in Two Dimensions"
}
| null | null | null | null | true | null |
11053
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we describe our attempt at producing a state-of-the-art Twitter\nsentiment classifier using Convolutional Neural Networks (CNNs) and Long Short\nTerm Memory (LSTMs) networks. Our system leverages a large amount of unlabeled\ndata to pre-train word embeddings. We then use a subset of the unlabeled data\nto fine tune the embeddings using distant supervision. The final CNNs and LSTMs\nare trained on the SemEval-2017 Twitter dataset where the embeddings are fined\ntuned again. To boost performances we ensemble several CNNs and LSTMs together.\nOur approach achieved first rank on all of the five English subtasks amongst 40\nteams.\n",
"title": "BB_twtr at SemEval-2017 Task 4: Twitter Sentiment Analysis with CNNs and LSTMs"
}
| null | null | null | null | true | null |
11054
| null |
Default
| null | null |
null |
{
"abstract": " Discerning how a mutation affects the stability of a protein is central to\nthe study of a wide range of diseases. Machine learning and statistical\nanalysis techniques can inform how to allocate limited resources to the\nconsiderable time and cost associated with wet lab mutagenesis experiments. In\nthis work we explore the effectiveness of using a neural network classifier to\npredict the change in the stability of a protein due to a mutation. Assessing\nthe accuracy of our approach is dependent on the use of experimental data about\nthe effects of mutations performed in vitro. Because the experimental data is\nprone to discrepancies when similar experiments have been performed by multiple\nlaboratories, the use of the data near the juncture of stabilizing and\ndestabilizing mutations is questionable. We address this later problem via a\nsystematic approach in which we explore the use of a three-way classification\nscheme with stabilizing, destabilizing, and inconclusive labels. For a\nsystematic search of potential classification cutoff values our classifier\nachieved 68 percent accuracy on ternary classification for cutoff values of\n-0.6 and 0.7 with a low rate of classifying stabilizing as destabilizing and\nvice versa.\n",
"title": "Protein Mutation Stability Ternary Classification using Neural Networks and Rigidity Analysis"
}
| null | null | null | null | true | null |
11055
| null |
Default
| null | null |
null |
{
"abstract": " We analyze the correlation between starspots and superflares on solar-type\nstars using observations from the Kepler mission. The analysis shows that the\nobserved fraction of stars with superflares decreases as the rotation period\nincreases and as the amplitude of photometric variability associated with\nrotation decreases. We found that the fraction of stars with superflares among\nthe stars showing large-amplitude rotational variations, which are thought to\nbe the signature of the large starspots, also decreases as the rotation period\nincreases. The small fraction of superflare stars among the stars with large\nstarspots in the longer-period regime suggests that some of the stars with\nlarge starspots show a much lower flare activity than the superflare stars with\nthe same spot area. Assuming simple relations between spot area and lifetime\nand between spot temperature and photospheric temperature, we compared the size\ndistribution of large starspot groups on slowly-rotating solar-type stars with\nthat of sunspot groups. The size distribution of starspots shows the power-law\ndistribution and the size distribution of larger sunspots lies on this\npower-law line. We also found that frequency-energy distributions for flares\noriginating from spots with different sizes are the same for solar-type stars\nwith superflares and the Sun. These results suggest that the magnetic activity\nwe observe on solar-type stars with superflares and that on the Sun is caused\nby the same physical processes.\n",
"title": "Starspot activity and superflares on solar-type stars"
}
| null | null | null | null | true | null |
11056
| null |
Default
| null | null |
null |
{
"abstract": " Monotone systems preserve a partial ordering of states along system\ntrajectories and are often amenable to separable Lyapunov functions that are\neither the sum or the maximum of a collection of functions of a scalar\nargument. In this paper, we consider constructing separable Lyapunov functions\nfor monotone systems that are also contractive, that is, the distance between\nany pair of trajectories exponentially decreases. The distance is defined in\nterms of a possibly state-dependent norm. When this norm is a weighted\none-norm, we obtain conditions which lead to sum-separable Lyapunov functions,\nand when this norm is a weighted infinity-norm, symmetric conditions lead to\nmax-separable Lyapunov functions. In addition, we consider two classes of\nLyapunov functions: the first class is separable along the system's state, and\nthe second class is separable along components of the system's vector field.\nThe latter case is advantageous for many practically motivated systems for\nwhich it is difficult to measure the system's state but easier to measure the\nsystem's velocity or rate of change. In addition, we present an algorithm based\non sum-of-squares programming to compute such separable Lyapunov functions. We\nprovide several examples to demonstrate our results.\n",
"title": "A Contractive Approach to Separable Lyapunov Functions for Monotone Systems"
}
| null | null | null | null | true | null |
11057
| null |
Default
| null | null |
null |
{
"abstract": " Isotopic ratios in comets are critical to understanding the origin of\ncometary material and the physical and chemical conditions in the early solar\nnebula. Comet C/2015 ER61 (PANSTARRS) underwent an outburst with a total\nbrightness increase of 2 magnitudes on the night of 2017 April 4. The sharp\nincrease in brightness offered a rare opportunity to measure the isotopic\nratios of the light elements in the coma of this comet. We obtained two\nhigh-resolution spectra of C/2015 ER61 with UVES/VLT on the nights of 2017\nApril 13 and 17. At the time of our observations, the comet was fading\ngradually following the outburst. We measured the nitrogen and carbon isotopic\nratios from the CN violet (0,0) band and found that $^{12}$C/$^{13}$C=100 $\\pm$\n15, $^{14}$N/$^{15}$N=130 $\\pm$ 15. In addition, we determined the\n$^{14}$N/$^{15}$N ratio from four pairs of NH$_2$ isotopolog lines and measured\n$^{14}$N/$^{15}$N=140 $\\pm$ 28. The measured isotopic ratios of C/2015 ER61 do\nnot deviate significantly from those of other comets.\n",
"title": "Isotopic ratios in outbursting comet C/2015 ER61"
}
| null | null |
[
"Physics"
] | null | true | null |
11058
| null |
Validated
| null | null |
null |
{
"abstract": " The COSINE-100 dark matter search experiment has started taking physics data\nwith the goal of performing an independent measurement of the annual modulation\nsignal observed by DAMA/LIBRA. A muon detector was constructed by using plastic\nscintillator panels in the outermost layer of the shield surrounding the\nCOSINE-100 detector. It is used to detect cosmic ray muons in order to\nunderstand the impact of the muon annual modulation on dark matter analysis.\nAssembly and initial performance test of each module have been performed at a\nground laboratory. The installation of the detector in Yangyang Underground\nLaboratory (Y2L) was completed in the summer of 2016. Using three months of\ndata, the muon underground flux was measured to be 328 $\\pm$ 1(stat.)$\\pm$\n10(syst.) muons/m$^2$/day. In this report, the assembly of the muon detector\nand the results from the analysis are presented.\n",
"title": "Muon detector for the COSINE-100 experiment"
}
| null | null | null | null | true | null |
11059
| null |
Default
| null | null |
null |
{
"abstract": " In recent years quantum phenomena have been experimentally demonstrated on\nvariety of optomechanical systems ranging from micro-oscillators to photonic\ncrystals. Since single photon couplings are quite small, most experimental\napproaches rely on the realization of high finesse Fabry-Perot cavities in\norder to enhance the effective coupling. Here we show that by exploiting a,\nlong path, low finesse fiber Fabry-Perot interferometer ground state cooling\ncan be achieved. We model a 100 m long cavity with a finesse of 10 and analyze\nthe impact of additional noise sources arising from the fiber. As a mechanical\noscillator we consider a levitated microdisk but the same approach could be\napplied to other optomechanical systems.\n",
"title": "Levitated optomechanics with a fiber Fabry-Perot interferometer"
}
| null | null |
[
"Physics"
] | null | true | null |
11060
| null |
Validated
| null | null |
null |
{
"abstract": " The efficiency of intracellular cargo transport from specific source to\ntarget locations is strongly dependent upon molecular motor-assisted motion\nalong the cytoskeleton. Radial transport along microtubules and lateral\ntransport along the filaments of the actin cortex underneath the cell membrane\nare characteristic for cells with a centrosome. The interplay between the\nspecific cytoskeleton organization and the motor performance realizes a\nspatially inhomogeneous intermittent search strategy. In order to analyze the\nefficiency of such intracellular search strategies we formulate a random\nvelocity model with intermittent arrest states. We evaluate efficiency in terms\nof mean first passage times for three different, frequently encountered\nintracellular transport tasks: i) the narrow escape problem, which emerges\nduring cargo transport to a synapse or other specific region of the cell\nmembrane, ii) the reaction problem, which considers the binding time of two\nparticles within the cell, and iii) the reaction-escape problem, which arises\nwhen cargo must be released at a synapse only after pairing with another\nparticle. Our results indicate that cells are able to realize efficient search\nstrategies for various intracellular transport tasks economically through a\nspatial cytoskeleton organization that involves only a narrow actin cortex\nrather than a cell body filled with randomly oriented actin filaments.\n",
"title": "Spatial cytoskeleton organization supports targeted intracellular transport"
}
| null | null | null | null | true | null |
11061
| null |
Default
| null | null |
null |
{
"abstract": " Deep learning models have consistently outperformed traditional machine\nlearning models in various classification tasks, including image\nclassification. As such, they have become increasingly prevalent in many real\nworld applications including those where security is of great concern. Such\npopularity, however, may attract attackers to exploit the vulnerabilities of\nthe deployed deep learning models and launch attacks against security-sensitive\napplications. In this paper, we focus on a specific type of data poisoning\nattack, which we refer to as a {\\em backdoor injection attack}. The main goal\nof the adversary performing such attack is to generate and inject a backdoor\ninto a deep learning model that can be triggered to recognize certain embedded\npatterns with a target label of the attacker's choice. Additionally, a backdoor\ninjection attack should occur in a stealthy manner, without undermining the\nefficacy of the victim model. Specifically, we propose two approaches for\ngenerating a backdoor that is hardly perceptible yet effective in poisoning the\nmodel. We consider two attack settings, with backdoor injection carried out\neither before model training or during model updating. We carry out extensive\nexperimental evaluations under various assumptions on the adversary model, and\ndemonstrate that such attacks can be effective and achieve a high attack\nsuccess rate (above $90\\%$) at a small cost of model accuracy loss (below\n$1\\%$) with a small injection rate (around $1\\%$), even under the weakest\nassumption wherein the adversary has no knowledge either of the original\ntraining data or the classifier model.\n",
"title": "Backdoor Embedding in Convolutional Neural Network Models via Invisible Perturbation"
}
| null | null | null | null | true | null |
11062
| null |
Default
| null | null |
null |
{
"abstract": " Using a semiclassical Green's function formalism, we discover the emergence\nof chiral and topological orbital magnetism in two-dimensional chiral spin\ntextures by explicitly finding the corrections to the orbital magnetization,\nproportional to the powers of the gradients of the texture. We show that in the\nabsence of spin-orbit coupling, the resulting orbital moment can be understood\nas the electronic response to the emergent magnetic field associated with the\nreal-space Berry curvature. By referring to the Rashba model, we demonstrate\nthat by tuning the parameters of surface systems the engineering of emergent\norbital magnetism in spin textures can pave the way to novel concepts in\norbitronics.\n",
"title": "Chiral and Topological Orbital Magnetism of Spin Textures"
}
| null | null |
[
"Physics"
] | null | true | null |
11063
| null |
Validated
| null | null |
null |
{
"abstract": " We introduce a new cosmic emulator for the matter power spectrum covering\neight cosmological parameters. Targeted at optical surveys, the emulator\nprovides accurate predictions out to a wavenumber k~5/Mpc and redshift z<=2.\nBesides covering the standard set of LCDM parameters, massive neutrinos and a\ndynamical dark energy of state are included. The emulator is built on a sample\nset of 36 cosmological models, carefully chosen to provide accurate predictions\nover the wide and large parameter space. For each model, we have performed a\nhigh-resolution simulation, augmented with sixteen medium-resolution\nsimulations and TimeRG perturbation theory results to provide accurate coverage\nof a wide k-range; the dataset generated as part of this project is more than\n1.2Pbyte. With the current set of simulated models, we achieve an accuracy of\napproximately 4%. Because the sampling approach used here has established\nconvergence and error-control properties, follow-on results with more than a\nhundred cosmological models will soon achieve ~1% accuracy. We compare our\napproach with other prediction schemes that are based on halo model ideas and\nremapping approaches. The new emulator code is publicly available.\n",
"title": "The Mira-Titan Universe II: Matter Power Spectrum Emulation"
}
| null | null | null | null | true | null |
11064
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we compare the performance of various homomorphic encryption\nmethods on a private search scheme that can achieve $k$-anonymity privacy. To\nmake our benchmarking fair, we use open sourced cryptographic libraries which\nare written by experts and well scrutinized. We find that Goldwasser-Micali\nencryption achieves good enough performance for practical use, whereas fully\nhomomorphic encryptions are much slower than partial ones like\nGoldwasser-Micali and Paillier.\n",
"title": "k-Anonymously Private Search over Encrypted Data"
}
| null | null | null | null | true | null |
11065
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we consider Burgers' equation with uncertain boundary and\ninitial conditions. The polynomial chaos (PC) approach yields a hyperbolic\nsystem of deterministic equations, which can be solved by several numerical\nmethods. Here, we apply the correction procedure via reconstruction (CPR) using\nsummation-by-parts operators. We focus especially on stability, which is proven\nfor CPR methods and the systems arising from the PC approach. Due to the usage\nof split-forms, the major challenge is to construct entropy stable numerical\nfluxes. For the first time, such numerical fluxes are constructed for all\nsystems resulting from the PC approach for Burgers' equation. In numerical\ntests, we verify our results and show also the advantage of the given ansatz\nusing CPR methods. Moreover, one of the simulations, i.e. Burgers' equation\nequipped with an initial shock, demonstrates quite fascinating observations.\nThe behaviour of the numerical solutions from several methods (finite volume,\nfinite difference, CPR) differ significantly from each other. Through careful\ninvestigations, we conclude that the reason for this is the high sensitivity of\nthe system to varying dissipation. Furthermore, it should be stressed that the\nsystem is not strictly hyperbolic with genuinely nonlinear or linearly\ndegenerate fields.\n",
"title": "Stability of Correction Procedure via Reconstruction With Summation-by-Parts Operators for Burgers' Equation Using a Polynomial Chaos Approach"
}
| null | null | null | null | true | null |
11066
| null |
Default
| null | null |
null |
{
"abstract": " Delay-differential equations are functional differential equations that\ninvolve shifts and derivatives with respect to a single independent variable.\nSome integrability candidates in this class have been identified by various\nmeans. For three of these equations we consider their elliptic and soliton-type\nsolutions. Using Hirota's bilinear method, we find that two of our equations\npossess three-soliton-type solutions.\n",
"title": "Special Solutions of Bi-Riccati Delay-Differential Equations"
}
| null | null | null | null | true | null |
11067
| null |
Default
| null | null |
null |
{
"abstract": " VSe2 is a transition metal dichaclogenide which has a charge-density wave\ntransition that has been well studied. We report on a low-temperature upturn in\nthe resistivity and, at temperatures below this resistivity minimum, an unusual\nmagnetoresistance which is negative at low fields and positive at higher\nfields, in single crystals of VSe2. The negative magnetoresistance has a\nparabolic dependence on the magnetic field and shows little angular dependence.\nThe magnetoresistance at temperatures above the resistivity minimum is always\npositive. We interpret these results as signatures of the Kondo effect in VSe2.\nAn upturn in the susceptibility indicates the presence of interlayer V ions\nwhich can provide the localized magnetic moments required for scattering the\nconduction electrons in the Kondo effect. The low-temperature behaviour of the\nheat capacity, including a high value of gamma, along with a deviation from a\nCurie-Weiss law observed in the low-temperature magnetic susceptibility, are\nconsistent with the presence of magnetic interactions between the paramagnetic\ninterlayer V ions and a Kondo screening of these V moments.\n",
"title": "Signatures of the Kondo effect in VSe2"
}
| null | null |
[
"Physics"
] | null | true | null |
11068
| null |
Validated
| null | null |
null |
{
"abstract": " Procedural textures are normally generated from mathematical models with\nparameters carefully selected by experienced users. However, for naive users,\nthe intuitive way to obtain a desired texture is to provide semantic\ndescriptions such as \"regular,\" \"lacelike,\" and \"repetitive\" and then a\nprocedural model with proper parameters will be automatically suggested to\ngenerate the corresponding textures. By contrast, it is less practical for\nusers to learn mathematical models and tune parameters based on multiple\nexaminations of large numbers of generated textures. In this study, we propose\na novel framework that generates procedural textures according to user-defined\nsemantic descriptions, and we establish a mapping between procedural models and\nsemantic texture descriptions. First, based on a vocabulary of semantic\nattributes collected from psychophysical experiments, a multi-label learning\nmethod is employed to annotate a large number of textures with semantic\nattributes to form a semantic procedural texture dataset. Then, we derive a low\ndimensional semantic space in which the semantic descriptions can be separated\nfrom one other. Finally, given a set of semantic descriptions, the diverse\nproperties of the samples in the semantic space can lead the framework to find\nan appropriate generation model that uses appropriate parameters to produce a\ndesired texture. The experimental results show that the proposed framework is\neffective and that the generated textures closely correlate with the input\nsemantic descriptions.\n",
"title": "A Procedural Texture Generation Framework Based on Semantic Descriptions"
}
| null | null | null | null | true | null |
11069
| null |
Default
| null | null |
null |
{
"abstract": " Multiple planet systems provide an ideal laboratory for probing exoplanet\ncomposition, formation history and potential habitability. For the TRAPPIST-1\nplanets, the planetary radii are well established from transits (Gillon et al.,\n2016, Gillon et al., 2017), with reasonable mass estimates coming from transit\ntiming variations (Gillon et al., 2017, Wang et al., 2017) and dynamical\nmodeling (Quarles et al., 2017). The low bulk densities of the TRAPPIST-1\nplanets demand significant volatile content. Here we show using\nmass-radius-composition models, that TRAPPIST-1f and g likely contain\nsubstantial ($\\geq50$ wt\\%) water/ice, with b and c being significantly drier\n($\\leq15$ wt\\%). We propose this gradient of water mass fractions implies\nplanets f and g formed outside the primordial snow line whereas b and c formed\ninside. We find that compared to planets in our solar system that also formed\nwithin the snow line, TRAPPIST-1b and c contain hundreds more oceans worth of\nwater. We demonstrate the extent and timescale of migration in the TRAPPIST-1\nsystem depends on how rapidly the planets formed and the relative location of\nthe primordial snow line. This work provides a framework for understanding the\ndifferences between the protoplanetary disks of our solar system versus M\ndwarfs. Our results provide key insights into the volatile budgets, timescales\nof planet formation, and migration history of likely the most common planetary\nhost in the Galaxy.\n",
"title": "Inward Migration of the TRAPPIST-1 Planets as Inferred From Their Water-Rich Compositions"
}
| null | null | null | null | true | null |
11070
| null |
Default
| null | null |
null |
{
"abstract": " Conditional Generative Adversarial Networks (cGANs) are generative models\nthat can produce data samples ($x$) conditioned on both latent variables ($z$)\nand known auxiliary information ($c$). We propose the Bidirectional cGAN\n(BiCoGAN), which effectively disentangles $z$ and $c$ in the generation process\nand provides an encoder that learns inverse mappings from $x$ to both $z$ and\n$c$, trained jointly with the generator and the discriminator. We present\ncrucial techniques for training BiCoGANs, which involve an extrinsic factor\nloss along with an associated dynamically-tuned importance weight. As compared\nto other encoder-based cGANs, BiCoGANs encode $c$ more accurately, and utilize\n$z$ and $c$ more effectively and in a more disentangled way to generate\nsamples.\n",
"title": "Bidirectional Conditional Generative Adversarial Networks"
}
| null | null | null | null | true | null |
11071
| null |
Default
| null | null |
null |
{
"abstract": " The attainability of modification of the apparent magnetic anisotropy in\n(III,Mn)V ferromagnetic semiconductors is probed by means of the\nfinite-elements-based modelling. The most representative case of (Ga,Mn)As and\nits in-plane uniaxial anisotropy is investigated. The hysteresis loops of the\ncontinuous films of a ferromagnetic semiconductor as well as films structured\nwith the elliptic antidots are modelled for various eccentricity, orientation,\nand separation of the anti dots. The effect of anti-dots on the magnetic\nanisotropy is confirmed but overall is found to be very weak. The subsequent\nmodelling for (Ga,Mn)As film with the elliptic dots comprising of metallic NiFe\nshows much stronger effect, revealing switching of the magnetic moment in the\nferromagnetic semiconductor governed by the switching behavior of the metallic\ninclusions.\n",
"title": "Micromagnetic study of a feasibility of the magnetic anisotropy engineering in nano-structured epitaxial films of (III,Mn)V ferromagnetic semiconductors"
}
| null | null | null | null | true | null |
11072
| null |
Default
| null | null |
null |
{
"abstract": " We complement the argument of M. Z. Garaev (2009) with several other ideas to\nobtain a stronger version of the large sieve inequality with sparse exponential\nsequences of the form $\\lambda^{s_n}$. In particular, we obtain a result which\nis non-trivial for monotonically increasing sequences $\\cal{S}=\\{s_n\n\\}_{n=1}^{\\infty}$ provided $s_n\\le n^{2+o(1)}$, whereas the original argument\nof M. Z. Garaev requires $s_n \\le n^{15/14 +o(1)}$ in the same setting. We also\ngive an application of our result to arithmetic properties of integers with\nalmost all digits prescribed.\n",
"title": "On the exponential large sieve inequality for sparse sequences modulo primes"
}
| null | null | null | null | true | null |
11073
| null |
Default
| null | null |
null |
{
"abstract": " $YBaCuO-Ag$ pressure point contacts with direct conduction are investigated.\nThe excess (relative to the normal state) conductivity mainly caused by\nfluctuational pairing of electrons above $T_c$ is measured in the temperature\ninterval 100-200~$K$. The superconductivity above 120~$K$ is found to be of the\ntwo-dimensional type. The obtained preliminary results indicate the presence of\nsmall amount of an unknown phase with $T'_c\\gtrsim 200~K$ in $YBaCuO$.\n",
"title": "Excess conduction of YBaCuO point contacts between 100 and 200 K"
}
| null | null |
[
"Physics"
] | null | true | null |
11074
| null |
Validated
| null | null |
null |
{
"abstract": " We show that Zamolodchikov dynamics of a recurrent quiver has zero algebraic\nentropy only if the quiver has a weakly subadditive labeling, and conjecture\nthe converse. By assigning a pair of generalized Cartan matrices of affine type\nto each quiver with an additive labeling, we completely classify such quivers,\nobtaining $40$ infinite families and $13$ exceptional quivers. This completes\nthe program of classifying Zamolodchikov periodic and integrable quivers.\n",
"title": "Quivers with additive labelings: classification and algebraic entropy"
}
| null | null | null | null | true | null |
11075
| null |
Default
| null | null |
null |
{
"abstract": " In most macro-scale robotics systems , propulsion and controls are enabled\nthrough a physical tether or complex on-board electronics and batteries. A\ntether simplifies the design process but limits the range of motion of the\nrobot, while on-board controls and power supplies are heavy and complicate the\ndesign process. Here we present a simple design principle for an untethered,\nentirely soft, swimming robot with the ability to achieve preprogrammed,\ndirectional propulsion without a battery or on-board electronics. Locomotion is\nachieved by employing actuators that harness the large displacements of\nbistable elements, triggered by surrounding temperature changes. Powered by\nshape memory polymer (SMP) muscles, the bistable elements in turn actuates the\nrobot's fins. Our robots are fabricated entirely using a commercially available\n3D printer with no post-processing. As a proof-of-concept, we demonstrate the\nability to program a vessel, which can autonomously deliver a cargo and\nnavigate back to the deployment point.\n",
"title": "Harnessing bistability for directional propulsion of untethered, soft robots"
}
| null | null | null | null | true | null |
11076
| null |
Default
| null | null |
null |
{
"abstract": " Long-range macrodimers formed by D-state cesium Rydberg atoms are studied in\nexperiments and in calculations. Cesium 62DJ-62DJ Rydberg-atom macrodimers,\nbonded via long-range multipole interaction, are prepared by two-color\nphoto-association in a cesium atom trap. The first color (pulse A) resonantly\nexcites seed Rydberg atoms, while the second (pulse B, detuned by the molecular\nbinding energy) resonantly excites the Rydberg-atom macrodimer states below the\n62DJ pair asymptotes. The Rydberg-atom molecules are measured by extraction of\nauto-ionization products and Rydberg-atom electric-field ionization, and ion\ndetection. Molecular spectra are compared with calculations of adiabatic\nmolecular potentials. The lifetime of the molecules is obtained from\nexponential fits to the dependence of the molecular signal on the detection\ndelay time; lifetimes of about 6 us are found.\n",
"title": "Cs nDJ Rydberg-atom macrodimers formed by long-range multipole interaction"
}
| null | null | null | null | true | null |
11077
| null |
Default
| null | null |
null |
{
"abstract": " Galactic winds from star-forming galaxies play at key role in the evolution\nof galaxies and the inter-galactic medium. They transport metals out of\ngalaxies, chemically-enriching the inter-galactic medium and modifying the\nchemical evolution of galaxies. They affect the surrounding inter-stellar and\ncircum-galactic media, thereby influencing the growth of galaxies through gas\naccretion and star-formation. In this contribution we first summarize the\nphysical mechanisms by which the momentum and energy output from a population\nof massive stars and associated supernovae can drive galactic winds. We use the\nproto-typical example of M82 to illustrate the multiphase nature of galactic\nwinds. We then describe how the basic properties of galactic winds are derived\nfrom the data, and summarize how the properties of galactic winds vary\nsystematically with the properties of the galaxies that launch them. We\nconclude with a brief discussion of the broad implications of galactic winds.\n",
"title": "A Brief Review of Galactic Winds"
}
| null | null |
[
"Physics"
] | null | true | null |
11078
| null |
Validated
| null | null |
null |
{
"abstract": " End-to-end training of automated speech recognition (ASR) systems requires\nmassive data and compute resources. We explore transfer learning based on model\nadaptation as an approach for training ASR models under constrained GPU memory,\nthroughput and training data. We conduct several systematic experiments\nadapting a Wav2Letter convolutional neural network originally trained for\nEnglish ASR to the German language. We show that this technique allows faster\ntraining on consumer-grade resources while requiring less training data in\norder to achieve the same accuracy, thereby lowering the cost of training ASR\nmodels in other languages. Model introspection revealed that small adaptations\nto the network's weights were sufficient for good performance, especially for\ninner layers.\n",
"title": "Transfer Learning for Speech Recognition on a Budget"
}
| null | null | null | null | true | null |
11079
| null |
Default
| null | null |
null |
{
"abstract": " Even though sequence-to-sequence neural machine translation (NMT) model have\nachieved state-of-art performance in the recent fewer years, but it is widely\nconcerned that the recurrent neural network (RNN) units are very hard to\ncapture the long-distance state information, which means RNN can hardly find\nthe feature with long term dependency as the sequence becomes longer.\nSimilarly, convolutional neural network (CNN) is introduced into NMT for\nspeeding recently, however, CNN focus on capturing the local feature of the\nsequence; To relieve this issue, we incorporate a relation network into the\nstandard encoder-decoder framework to enhance information-propogation in neural\nnetwork, ensuring that the information of the source sentence can flow into the\ndecoder adequately. Experiments show that proposed framework outperforms the\nstatistical MT model and the state-of-art NMT model significantly on two data\nsets with different scales.\n",
"title": "Information-Propogation-Enhanced Neural Machine Translation by Relation Model"
}
| null | null | null | null | true | null |
11080
| null |
Default
| null | null |
null |
{
"abstract": " This tutorial introduces a new and powerful set of techniques variously\ncalled \"neural machine translation\" or \"neural sequence-to-sequence models\".\nThese techniques have been used in a number of tasks regarding the handling of\nhuman language, and can be a powerful tool in the toolbox of anyone who wants\nto model sequential data of some sort. The tutorial assumes that the reader\nknows the basics of math and programming, but does not assume any particular\nexperience with neural networks or natural language processing. It attempts to\nexplain the intuition behind the various methods covered, then delves into them\nwith enough mathematical detail to understand them concretely, and culiminates\nwith a suggestion for an implementation exercise, where readers can test that\nthey understood the content in practice.\n",
"title": "Neural Machine Translation and Sequence-to-sequence Models: A Tutorial"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
11081
| null |
Validated
| null | null |
null |
{
"abstract": " We investigate the minimum cases for realtime probabilistic machines that can\ndefine uncountably many languages with bounded error. We show that logarithmic\nspace is enough for realtime PTMs on unary languages. On binary case, we follow\nthe same result for double logarithmic space, which is tight. When replacing\nthe worktape with some limited memories, we can follow uncountable results on\nunary languages for two counters.\n",
"title": "Uncountable realtime probabilistic classes"
}
| null | null | null | null | true | null |
11082
| null |
Default
| null | null |
null |
{
"abstract": " We introduce the abstract notion of a necklical set in order to describe a\nfunctorial combinatorial model of the path fibration over the geometric\nrealization of a path connected simplicial set. In particular, to any path\nconnected simplicial set $X$ we associate a necklical set\n$\\widehat{\\mathbf{\\Omega}}X$ such that its geometric realization\n$|\\widehat{\\mathbf{\\Omega}}X|$, a space built out of gluing cubical cells, is\nhomotopy equivalent to the based loop space on $|X|$ and the differential\ngraded module of chains $C_*(\\widehat{\\mathbf{\\Omega}}X)$ is a differential\ngraded associative algebra generalizing Adams' cobar construction.\n",
"title": "A combinatorial model for the path fibration"
}
| null | null | null | null | true | null |
11083
| null |
Default
| null | null |
null |
{
"abstract": " Application Programming Interfaces (APIs) often have usage constraints, such\nas restrictions on call order or call conditions. API misuses, i.e., violations\nof these constraints, may lead to software crashes, bugs, and vulnerabilities.\nThough researchers developed many API-misuse detectors over the last two\ndecades, recent studies show that API misuses are still prevalent. Therefore,\nwe need to understand the capabilities and limitations of existing detectors in\norder to advance the state of the art. In this paper, we present the first-ever\nqualitative and quantitative evaluation that compares static API-misuse\ndetectors along the same dimensions, and with original author validation. To\naccomplish this, we develop MUC, a classification of API misuses, and\nMUBenchPipe, an automated benchmark for detector comparison, on top of our\nmisuse dataset, MUBench. Our results show that the capabilities of existing\ndetectors vary greatly and that existing detectors, though capable of detecting\nmisuses, suffer from extremely low precision and recall. A systematic\nroot-cause analysis reveals that, most importantly, detectors need to go beyond\nthe naive assumption that a deviation from the most-frequent usage corresponds\nto a misuse and need to obtain additional usage examples to train their models.\nWe present possible directions towards more-powerful API-misuse detectors.\n",
"title": "A Systematic Evaluation of Static API-Misuse Detectors"
}
| null | null | null | null | true | null |
11084
| null |
Default
| null | null |
null |
{
"abstract": " Does the human lifespan have an impenetrable biological upper limit which\nultimately will stop further increase in life lengths? This question is\nimportant for understanding aging, and for society, and has led to intense\ncontroversies. Demographic data for humans has been interpreted as showing\nexistence of a limit, or even as an indication of a decreasing limit, but also\nas evidence that a limit does not exist. This paper studies what can be\ninferred from data about human mortality at extreme age. We show that in\nwestern countries and Japan and after age 110 the probability of dying is about\n47% per year. Hence there is no finite upper limit to the human lifespan.\nStill, given the present stage of biotechnology, it is unlikely that during the\nnext 25 years anyone will live longer than 128 years in these countries. Data,\nremarkably, shows no difference in mortality after age 110 between sexes,\nbetween ages, or between different lifestyles or genetic backgrounds. These\nresults, and the analysis methods developed in this paper, can help testing\nbiological theories of ageing and aid confirmation of success of efforts to\nfind a cure for ageing.\n",
"title": "Human life is unlimited - but short"
}
| null | null | null | null | true | null |
11085
| null |
Default
| null | null |
null |
{
"abstract": " Object detection when provided image-level labels instead of instance-level\nlabels (i.e., bounding boxes) during training is an important problem in\ncomputer vision, since large scale image datasets with instance-level labels\nare extremely costly to obtain. In this paper, we address this challenging\nproblem by developing an Expectation-Maximization (EM) based object detection\nmethod using deep convolutional neural networks (CNNs). Our method is\napplicable to both the weakly-supervised and semi-supervised settings.\nExtensive experiments on PASCAL VOC 2007 benchmark show that (1) in the weakly\nsupervised setting, our method provides significant detection performance\nimprovement over current state-of-the-art methods, (2) having access to a small\nnumber of strongly (instance-level) annotated images, our method can almost\nmatch the performace of the fully supervised Fast RCNN. We share our source\ncode at this https URL.\n",
"title": "Weakly- and Semi-Supervised Object Detection with Expectation-Maximization Algorithm"
}
| null | null | null | null | true | null |
11086
| null |
Default
| null | null |
null |
{
"abstract": " The coverage probability of a user in a mmwave system depends on the\navailability of line-of-sight paths or reflected paths from any base station.\nMany prior works modelled blockages using random shape theory and analyzed the\nSIR distribution with and without interference. While, it is intuitive that the\nreflected paths do not significantly contribute to the coverage (because of\nlonger path lengths), there are no works which provide a model and study the\ncoverage with reflections. In this paper, we model and analyze the impact of\nreflectors using stochastic geometry. We observe that the reflectors have very\nlittle impact on the coverage probability.\n",
"title": "Coverage Analysis in Millimeter Wave Cellular Networks with Reflections"
}
| null | null | null | null | true | null |
11087
| null |
Default
| null | null |
null |
{
"abstract": " We study the application of polar codes in deletion channels by analyzing the\ncascade of a binary erasure channel (BEC) and a deletion channel. We show how\npolar codes can be used effectively on a BEC with a single deletion, and\npropose a list decoding algorithm with a cyclic redundancy check for this case.\nThe decoding complexity is $O(N^2\\log N)$, where $N$ is the blocklength of the\ncode. An important contribution is an optimization of the amount of redundancy\nadded to minimize the overall error probability. Our theoretical results are\ncorroborated by numerical simulations which show that the list size can be\nreduced to one and the original message can be recovered with high probability\nas the length of the code grows.\n",
"title": "Polar Coding for the Binary Erasure Channel with Deletions"
}
| null | null | null | null | true | null |
11088
| null |
Default
| null | null |
null |
{
"abstract": " In this dissertation, we focus on several important problems in structured\nprediction. In structured prediction, the label has a rich intrinsic\nsubstructure, and the loss varies with respect to the predicted label and the\ntrue label pair. Structured SVM is an extension of binary SVM to adapt to such\nstructured tasks.\nIn the first part of the dissertation, we study the surrogate losses and its\nefficient methods. To minimize the empirical risk, a surrogate loss which upper\nbounds the loss, is used as a proxy to minimize the actual loss. Since the\nobjective function is written in terms of the surrogate loss, the choice of the\nsurrogate loss is important, and the performance depends on it. Another issue\nregarding the surrogate loss is the efficiency of the argmax label inference\nfor the surrogate loss. Efficient inference is necessary for the optimization\nsince it is often the most time-consuming step. We present a new class of\nsurrogate losses named bi-criteria surrogate loss, which is a generalization of\nthe popular surrogate losses. We first investigate an efficient method for a\nslack rescaling formulation as a starting point utilizing decomposability of\nthe model. Then, we extend the algorithm to the bi-criteria surrogate loss,\nwhich is very efficient and also shows performance improvements.\nIn the second part of the dissertation, another important issue of\nregularization is studied. Specifically, we investigate a problem of\nregularization in hierarchical classification when a structural imbalance\nexists in the label structure. We present a method to normalize the structure,\nas well as a new norm, namely shared Frobenius norm. It is suitable for\nhierarchical classification that adapts to the data in addition to the label\nstructure.\n",
"title": "Efficient Structured Surrogate Loss and Regularization in Structured Prediction"
}
| null | null | null | null | true | null |
11089
| null |
Default
| null | null |
null |
{
"abstract": " We consider how to connect a set of disjoint networks to optimize the\nperformance of the resulting composite network. We quantify this performance by\nthe coherence of the composite network, which is defined by an $H_2$ norm of\nthe system. Two dynamics are considered: noisy consensus dynamics with and\nwithout stubborn agents. For noisy consensus dynamics without stubborn agents,\nwe derive analytical expressions for the coherence of composite networks in\nterms of the coherence of the individual networks and the structure of their\ninterconnections. We also identify optimal interconnection topologies and give\nbounds on coherence for general composite graphs. For noisy consensus dynamics\nwith stubborn agents, we develop a non-combinatorial algorithm that identifies\nconnecting edges such that the composite network coherence closely approximates\nthe performance of the optimal composite graph.\n",
"title": "Optimizing the Coherence of Composite Networks"
}
| null | null | null | null | true | null |
11090
| null |
Default
| null | null |
null |
{
"abstract": " In target tracking, the estimation of an unknown weaving target frequency is\ncrucial for improving the miss distance. The estimation process is commonly\ncarried out in a Kalman framework. The objective of this paper is to examine\nthe potential of using neural networks in target tracking applications. To that\nend, we propose estimating the weaving frequency using deep neural networks,\ninstead of classical Kalman framework based estimation. Particularly, we focus\non the case where a set of possible constant target frequencies is known.\nSeveral neural network architectures, requiring low computational resources\nwere designed to estimate the unknown frequency out of the known set of\nfrequencies. The proposed approach performance is compared with the multiple\nmodel adaptive estimation algorithm. Simulation results show that in the\nexamined scenarios, deep neural network outperforms multiple model adaptive\nestimation in terms of accuracy and the amount of required measurements to\nconvergence.\n",
"title": "Deep Learning based Estimation of Weaving Target Maneuvers"
}
| null | null | null | null | true | null |
11091
| null |
Default
| null | null |
null |
{
"abstract": " We give a detailed proof of some facts about the blow-up of horizontal curves\nin Carnot-Carathéodory spaces.\n",
"title": "On tangent cones to length minimizers in Carnot-Carathéodory spaces"
}
| null | null |
[
"Mathematics"
] | null | true | null |
11092
| null |
Validated
| null | null |
null |
{
"abstract": " In many machine learning applications, there are multiple decision-makers\ninvolved, both automated and human. The interaction between these agents often\ngoes unaddressed in algorithmic development. In this work, we explore a simple\nversion of this interaction with a two-stage framework containing an automated\nmodel and an external decision-maker. The model can choose to say \"Pass\", and\npass the decision downstream, as explored in rejection learning. We extend this\nconcept by proposing \"learning to defer\", which generalizes rejection learning\nby considering the effect of other agents in the decision-making process. We\npropose a learning algorithm which accounts for potential biases held by\nexternal decision-makers in a system. Experiments demonstrate that learning to\ndefer can make systems not only more accurate but also less biased. Even when\nworking with inconsistent or biased users, we show that deferring models still\ngreatly improve the accuracy and/or fairness of the entire system.\n",
"title": "Predict Responsibly: Improving Fairness and Accuracy by Learning to Defer"
}
| null | null | null | null | true | null |
11093
| null |
Default
| null | null |
null |
{
"abstract": " We consider two greedy algorithms for minimizing a convex function in a\nbounded convex set: an algorithm by Jones [1992] and the Frank-Wolfe (FW)\nalgorithm. We first consider approximate versions of these algorithms. For\nsmooth convex functions, we give sufficient conditions for convergence, a\nunified analysis for the well-known convergence rate of O(1/k) together with a\nresult showing that this rate is the best obtainable from the proof technique,\nand an equivalence result for the two algorithms. We also consider approximate\nstochastic greedy algorithms for minimizing expectations. We show that\nreplacing the full gradient by a single stochastic gradient can fail even on\nsmooth convex functions. We give a convergent approximate stochastic Jones\nalgorithm and a convergent approximate stochastic FW algorithm for smooth\nconvex functions. In addition, we give a convergent approximate stochastic FW\nalgorithm for nonsmooth convex functions. Convergence rates for these\nalgorithms are given and proved.\n",
"title": "Approximate and Stochastic Greedy Optimization"
}
| null | null | null | null | true | null |
11094
| null |
Default
| null | null |
null |
{
"abstract": " We study the phase diagram of the triangular-lattice $Q$-state Potts model in\nthe real $(Q,v)$-plane, where $v=e^J-1$ is the temperature variable. Our first\ngoal is to provide an obviously missing feature of this diagram: the position\nof the antiferromagnetic critical curve. This curve turns out to possess a\nbifurcation point with two branches emerging from it, entailing important\nconsequences for the global phase diagram. We have obtained accurate numerical\nestimates for the position of this curve by combining the transfer-matrix\napproach for strip graphs with toroidal boundary conditions and the recent\nmethod of critical polynomials. The second goal of this work is to study the\ncorresponding $A_{p-1}$ RSOS model on the torus, for integer $p=4,5,\\ldots,8$.\nWe clarify its relation to the corresponding Potts model, in particular\nconcerning the role of boundary conditions. For certain values of $p$, we\nidentify several new critical points and regimes for the RSOS model and we\ninitiate the study of the flows between the corresponding field theories.\n",
"title": "Phase diagram of the triangular-lattice Potts antiferromagnet"
}
| null | null | null | null | true | null |
11095
| null |
Default
| null | null |
null |
{
"abstract": " On June 24, 2018, Turkey held a historical election, transforming its\nparliamentary system to a presidential one. One of the main questions for\nTurkish voters was whether to start this new political era with reelecting its\nlong-time political leader Recep Tayyip Erdogan or not. In this paper, we\nanalyzed 108M tweets posted in the two months leading to the election to\nunderstand the groups that supported or opposed Erdogan's reelection. We\nexamined the most distinguishing hashtags and retweeted accounts for both\ngroups. Our findings indicate strong polarization between both groups as they\ndiffer in terms of ideology, news sources they follow, and preferred TV\nentertainment.\n",
"title": "Devam vs. Tamam: 2018 Turkish Elections"
}
| null | null |
[
"Computer Science"
] | null | true | null |
11096
| null |
Validated
| null | null |
null |
{
"abstract": " Insertion is a challenging haptic and visual control problem with significant\npractical value for manufacturing. Existing approaches in the model-based\nrobotics community can be highly effective when task geometry is known, but are\ncomplex and cumbersome to implement, and must be tailored to each individual\nproblem by a qualified engineer. Within the learning community there is a long\nhistory of insertion research, but existing approaches are typically either too\nsample-inefficient to run on real robots, or assume access to high-level object\nfeatures, e.g. socket pose. In this paper we show that relatively minor\nmodifications to an off-the-shelf Deep-RL algorithm (DDPG), combined with a\nsmall number of human demonstrations, allows the robot to quickly learn to\nsolve these tasks efficiently and robustly. Our approach requires no modeling\nor simulation, no parameterized search or alignment behaviors, no vision system\naside from raw images, and no reward shaping. We evaluate our approach on a\nnarrow-clearance peg-insertion task and a deformable clip-insertion task, both\nof which include variability in the socket position. Our results show that\nthese tasks can be solved reliably on the real robot in less than 10 minutes of\ninteraction time, and that the resulting policies are robust to variance in the\nsocket position and orientation.\n",
"title": "A Practical Approach to Insertion with Variable Socket Position Using Deep Reinforcement Learning"
}
| null | null |
[
"Computer Science"
] | null | true | null |
11097
| null |
Validated
| null | null |
null |
{
"abstract": " The Lorentz off-axis electron holography technique is applied to study the\nmagnetic nature of Mn rich nanoprecipitates in (Mn,Ga)As system. The\neffectiveness of this technique is demonstrated in detection of the magnetic\nfield even for small nanocrystals having an average size down to 20 nm.\n",
"title": "Off-axis electron holography of magnetic nanostructures: magnetic behavior of Mn rich nanoprecipitates in (Mn,Ga)As system"
}
| null | null | null | null | true | null |
11098
| null |
Default
| null | null |
null |
{
"abstract": " Capturing both the structural and temporal aspects of interactions is crucial\nfor many real world datasets like contact between individuals. Using the link\nstream formalism to capture the dynamic of the systems, we tackle the issue of\nactivity prediction in link streams, that is to say predicting the number of\nlinks occurring during a given period of time and we present a protocol that\ntakes advantage of the temporal and structural information contained in the\nlink stream. Using a supervised learning method, we are able to model the\ndynamic of our system to improve the prediction. We investigate the behavior of\nour algorithm and crucial elements affecting the prediction. By introducing\ndifferent categories of pair of nodes, we are able to improve the quality as\nwell as increase the diversity of our prediction.\n",
"title": "Predicting interactions between individuals with structural and dynamical information"
}
| null | null | null | null | true | null |
11099
| null |
Default
| null | null |
null |
{
"abstract": " In this article we develop algorithms for data assimilation based upon a\ncomputational time dependent stable/unstable splitting. Our particular method\nis based upon shadowing refinement and synchronization techniques and is\nmotivated by work on Assimilation in the Unstable Subspace (AUS) and\nPseudo-orbit Data Assimilation (PDA). The algorithm utilizes time dependent\nprojections onto the non-stable subspace determined by employing computational\ntechniques for Lyapunov exponents/vectors. The method is extended to parameter\nestimation without changing the problem dynamics and we address techniques for\nadapting the method when (as is commonly the case) observations are not\navailable in the full model state space. We use a combination of analysis and\nnumerical experiments (with the Lorenz 63 and Lorenz 96 models) to illustrate\nthe efficacy of the techniques and show that the results compare favorably with\nother variational techniques.\n",
"title": "Projected Shadowing-based Data Assimilation"
}
| null | null | null | null | true | null |
11100
| null |
Default
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.