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": " Many supervised learning tasks are emerged in dual forms, e.g.,\nEnglish-to-French translation vs. French-to-English translation, speech\nrecognition vs. text to speech, and image classification vs. image generation.\nTwo dual tasks have intrinsic connections with each other due to the\nprobabilistic correlation between their models. This connection is, however,\nnot effectively utilized today, since people usually train the models of two\ndual tasks separately and independently. In this work, we propose training the\nmodels of two dual tasks simultaneously, and explicitly exploiting the\nprobabilistic correlation between them to regularize the training process. For\nease of reference, we call the proposed approach \\emph{dual supervised\nlearning}. We demonstrate that dual supervised learning can improve the\npractical performances of both tasks, for various applications including\nmachine translation, image processing, and sentiment analysis.\n",
"title": "Dual Supervised Learning"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
4101
| null |
Validated
| null | null |
null |
{
"abstract": " A graph is said to be symmetric if its automorphism group is transitive on\nits arcs. Guo et al. (Electronic J. Combin. 18, \\#P233, 2011) and Pan et al.\n(Electronic J. Combin. 20, \\#P36, 2013) determined all pentavalent symmetric\ngraphs of order $4pq$. In this paper, we shall generalize this result by\ndetermining all connected pentavalent symmetric graphs of order four times an\nodd square-free integer. It is shown in this paper that, for each of such\ngraphs $\\it\\Gamma$, either the full automorphism group ${\\sf Aut}\\it\\Gamma$ is\nisomorphic to ${\\sf PSL}(2,p)$, ${\\sf PGL}(2,p)$, ${\\sf\nPSL}(2,p){\\times}\\mathbb{Z}_2$ or ${\\sf PGL}(2,p){\\times}\\mathbb{Z}_2$, or\n$\\it\\Gamma$ is isomorphic to one of 8 graphs.\n",
"title": "Pentavalent symmetric graphs of order four times an odd square-free integer"
}
| null | null | null | null | true | null |
4102
| null |
Default
| null | null |
null |
{
"abstract": " The run time of many scientific computation applications for numerical\nmethods is heavily dependent on just a few multi-dimensional loop nests. Since\nthese applications are often limited by memory bandwidth rather than\ncomputational resources they can benefit greatly from any optimizations which\ndecrease the run time of their loops by improving data reuse and thus reducing\nthe total memory traffic. Some of the most effective of these optimizations are\nnot suitable for development by hand or require advanced software engineering\nknowledge which is beyond the level of many researchers who are not specialists\nin code optimization. Several tools exist to automate the generation of\nhigh-performance code for numerical methods, such as Devito which produces code\nfor finite-difference approximations typically used in the seismic imaging\ndomain. We present a loop-tiling optimization which can be applied to\nDevito-generated loops and improves run time by up to 27.5%, and options for\nautomating this optimization in the Devito framework.\n",
"title": "Applying the Polyhedral Model to Tile Time Loops in Devito"
}
| null | null | null | null | true | null |
4103
| null |
Default
| null | null |
null |
{
"abstract": " In applications of deep reinforcement learning to robotics, it is often the\ncase that we want to learn pose invariant policies: policies that are invariant\nto changes in the position and orientation of objects in the world. For\nexample, consider a peg-in-hole insertion task. If the agent learns to insert a\npeg into one hole, we would like that policy to generalize to holes presented\nin different poses. Unfortunately, this is a challenge using conventional\nmethods. This paper proposes a novel state and action abstraction that is\ninvariant to pose shifts called \\textit{deictic image maps} that can be used\nwith deep reinforcement learning. We provide broad conditions under which\noptimal abstract policies are optimal for the underlying system. Finally, we\nshow that the method can help solve challenging robotic manipulation problems.\n",
"title": "Deictic Image Maps: An Abstraction For Learning Pose Invariant Manipulation Policies"
}
| null | null |
[
"Computer Science"
] | null | true | null |
4104
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper, we investigate the Hawking radiation process as a\nsemiclassical quantum tunneling phenomenon from black ring and Myers-Perry\nblack holes in 5-dimensional (5D) spaces. Using Lagrangian of\nGlashow-Weinberg-Salam model with background electromagnetic field (for charged\nW-bosons) and the WKB approximation, we will evaluate the tunneling\nrate/probability of charged vector particles through horizons by taking into\naccount the electromagnetic vector potential. Moreover, we investigate the\ncorresponding Hawking temperature values by considering Boltzmann factor for\nboth cases and analyze the whole spectrum generally.\n",
"title": "Charged Vector Particles Tunneling From 5D Black Hole and Black Ring"
}
| null | null | null | null | true | null |
4105
| null |
Default
| null | null |
null |
{
"abstract": " OpenMP is a shared memory programming model which supports the offloading of\ntarget regions to accelerators such as NVIDIA GPUs. The implementation in\nClang/LLVM aims to deliver a generic GPU compilation toolchain that supports\nboth the native CUDA C/C++ and the OpenMP device offloading models. There are\nsituations where the semantics of OpenMP and those of CUDA diverge. One such\nexample is the policy for implicitly handling local variables. In CUDA, local\nvariables are implicitly mapped to thread local memory and thus become private\nto a CUDA thread. In OpenMP, due to semantics that allow the nesting of regions\nexecuted by different numbers of threads, variables need to be implicitly\n\\emph{shared} among the threads of a contention group. In this paper we\nintroduce a re-design of the OpenMP device data sharing infrastructure that is\nresponsible for the implicit sharing of local variables in the Clang/LLVM\ntoolchain. We introduce a new data sharing infrastructure that lowers\nimplicitly shared variables to the shared memory of the GPU. We measure the\namount of shared memory used by our scheme in cases that involve scalar\nvariables and statically allocated arrays. The evaluation is carried out by\noffloading to K40 and P100 NVIDIA GPUs. For scalar variables the pressure on\nshared memory is relatively low, under 26\\% of shared memory utilization for\nthe K40, and does not negatively impact occupancy. The limiting occupancy\nfactor in that case is register pressure. The data sharing scheme offers the\nusers a simple memory model for controlling the implicit allocation of device\nshared memory.\n",
"title": "Implementing implicit OpenMP data sharing on GPUs"
}
| null | null |
[
"Computer Science"
] | null | true | null |
4106
| null |
Validated
| null | null |
null |
{
"abstract": " In the recent years image processing techniques are used as a tool to improve\ndetection and diagnostic capabilities in the medical applications. Medical\napplications have been so much affected by these techniques which some of them\nare embedded in medical instruments such as MRI, CT and other medical devices.\nAmong these techniques, medical image enhancement algorithms play an essential\nrole in removal of the noise which can be produced by medical instruments and\nduring image transfer. It has been proved that impulse noise is a major type of\nnoise, which is produced during medical operations, such as MRI, CT, and\nangiography, by their image capturing devices. An embeddable hardware module\nwhich is able to denoise medical images before and during surgical operations\ncould be very helpful. In this paper an accurate algorithm is proposed for\nreal-time removal of impulse noise in medical images. All image blocks are\ndivided into three categories of edge, smooth, and disordered areas. A\ndifferent reconstruction method is applied to each category of blocks for the\npurpose of noise removal. The proposed method is tested on MR images.\nSimulation results show acceptable denoising accuracy for various levels of\nnoise. Also an FPAG implementation of our denoising algorithm shows acceptable\nhardware resource utilization. Hence, the algorithm is suitable for embedding\nin medical hardware instruments such as radiosurgery devices.\n",
"title": "Real-Time Impulse Noise Removal from MR Images for Radiosurgery Applications"
}
| null | null | null | null | true | null |
4107
| null |
Default
| null | null |
null |
{
"abstract": " Semi-supervised node classification in graphs is a fundamental problem in\ngraph mining, and the recently proposed graph neural networks (GNNs) have\nachieved unparalleled results on this task. Due to their massive success, GNNs\nhave attracted a lot of attention, and many novel architectures have been put\nforward. In this paper we show that existing evaluation strategies for GNN\nmodels have serious shortcomings. We show that using the same\ntrain/validation/test splits of the same datasets, as well as making\nsignificant changes to the training procedure (e.g. early stopping criteria)\nprecludes a fair comparison of different architectures. We perform a thorough\nempirical evaluation of four prominent GNN models and show that considering\ndifferent splits of the data leads to dramatically different rankings of\nmodels. Even more importantly, our findings suggest that simpler GNN\narchitectures are able to outperform the more sophisticated ones if the\nhyperparameters and the training procedure are tuned fairly for all models.\n",
"title": "Pitfalls of Graph Neural Network Evaluation"
}
| null | null | null | null | true | null |
4108
| null |
Default
| null | null |
null |
{
"abstract": " The CANDECOMP/PARAFAC (CP) tensor decomposition is a popular\ndimensionality-reduction method for multiway data. Dimensionality reduction is\noften sought since many high-dimensional tensors have low intrinsic rank\nrelative to the dimension of the ambient measurement space. However, the\nemergence of `big data' poses significant computational challenges for\ncomputing this fundamental tensor decomposition. Leveraging modern randomized\nalgorithms, we demonstrate that the coherent structure can be learned from a\nsmaller representation of the tensor in a fraction of the time. Moreover, the\nhigh-dimensional signal can be faithfully approximated from the compressed\nmeasurements. Thus, this simple but powerful algorithm enables one to compute\nthe approximate CP decomposition even for massive tensors. The approximation\nerror can thereby be controlled via oversampling and the computation of power\niterations. In addition to theoretical results, several empirical results\ndemonstrate the performance of the proposed algorithm.\n",
"title": "Randomized CP Tensor Decomposition"
}
| null | null | null | null | true | null |
4109
| null |
Default
| null | null |
null |
{
"abstract": " Given the increasing competition in mobile app ecosystems, improving the\nexperience of users has become a major goal for app vendors. This article\nintroduces a visionary app store, called APP STORE 2.0, which exploits\ncrowdsourced information about apps, devices and users to increase the overall\nquality of the delivered mobile apps. We sketch a blueprint architecture of the\nenvisioned app stores and discuss the different kinds of actionable feedbacks\nthat app stores can generate using crowdsourced information.\n",
"title": "App Store 2.0: From Crowd Information to Actionable Feedback in Mobile Ecosystems"
}
| null | null |
[
"Computer Science"
] | null | true | null |
4110
| null |
Validated
| null | null |
null |
{
"abstract": " This paper studies the problem of remote state estimation in the presence of\na passive eavesdropper. A sensor measures a linear plant's state and transmits\nit to an authorized user over a packet-dropping channel, which is susceptible\nto eavesdropping. Our goal is to design a coding scheme such that the\neavesdropper cannot infer the plant's current state, while the user\nsuccessfully decodes the sent messages. We employ a novel class of codes,\ntermed State-Secrecy Codes, which are fast and efficient for dynamical systems.\nThey apply linear time-varying transformations to the current and past states\nreceived by the user. In this way, they force the eavesdropper's information\nmatrix to decrease with asymptotically the same rate as in the open-loop\nprediction case, i.e. when the eavesdropper misses all messages. As a result,\nthe eavesdropper's minimum mean square error (mmse) for the unstable states\ngrows unbounded, while the respective error for the stable states converges to\nthe open-loop prediction one. These secrecy guarantees are achieved under\nminimal conditions, which require that, at least once, the user receives the\ncorresponding packet while the eavesdropper fails to intercept it. Meanwhile,\nthe user's estimation performance remains optimal. The theoretical results are\nillustrated in simulations.\n",
"title": "An Information Matrix Approach for State Secrecy"
}
| null | null | null | null | true | null |
4111
| null |
Default
| null | null |
null |
{
"abstract": " The goal of network representation learning is to learn low-dimensional node\nembeddings that capture the graph structure and are useful for solving\ndownstream tasks. However, despite the proliferation of such methods there is\ncurrently no study of their robustness to adversarial attacks. We provide the\nfirst adversarial vulnerability analysis on the widely used family of methods\nbased on random walks. We derive efficient adversarial perturbations that\npoison the network structure and have a negative effect on both the quality of\nthe embeddings and the downstream tasks. We further show that our attacks are\ntransferable -- they generalize to many models -- and are successful even when\nthe attacker has restricted actions.\n",
"title": "Adversarial Attacks on Node Embeddings"
}
| null | null | null | null | true | null |
4112
| null |
Default
| null | null |
null |
{
"abstract": " Manipulating and focusing light deep inside biological tissue and tissue-like\ncomplex media has been desired for long yet considered challenging. One\nfeasible strategy is through optical wavefront engineering, where the optical\nscattering-induced phase distortions are time reversed or pre-compensated so\nthat photons travel along different optical paths interfere constructively at\nthe targeted position within a scattering medium. To define the targeted\nposition, an internal guidestar is needed to guide or provide a feedback for\nwavefront engineering. It could be injected or embedded probes such as\nfluorescence or nonlinear microspheres, ultrasonic modulation, as well as\nabsorption perturbation. Here we propose to use a magnetically controlled\noptical absorbing microsphere as the internal guidestar. Using a digital\noptical phase conjugation system, we obtained sharp optical focusing within\nscattering media through time-reversing the scattered light perturbed by the\nmagnetic microshpere. Since the object is magnetically controlled, dynamic\noptical focusing is allowed with a relatively large field-of-view by scanning\nthe magnetic field externally. Moreover, the magnetic microsphere can be\npackaged with an organic membrane, using biological or chemical means to serve\nas a carrier. Therefore the technique may find particular applications for\nenhanced targeted drug delivery, and imaging and photoablation of angiogenic\nvessels in tumours.\n",
"title": "Time-reversed magnetically controlled perturbation (TRMCP) optical focusing inside scattering media"
}
| null | null | null | null | true | null |
4113
| null |
Default
| null | null |
null |
{
"abstract": " The competition between spin-orbit coupling, bandwidth ($W$) and\nelectron-electron interaction ($U$) makes iridates highly susceptible to small\nexternal perturbations, which can trigger the onset of novel types of\nelectronic and magnetic states. Here we employ {\\em first principles}\ncalculations based on density functional theory and on the constrained random\nphase approximation to study how dimensionality and strain affect the strength\nof $U$ and $W$ in (SrIrO$_3$)$_m$/(SrTiO$_3$) superlattices. The result is a\nphase diagram explaining two different types of controllable magnetic and\nelectronic transitions, spin-flop and insulator-to-metal, connected with the\ndisruption of the $J_{eff}=1/2$ state which cannnot be understood within a\nsimplified local picture.\n",
"title": "Dimensionality-strain phase diagram of strontium iridates"
}
| null | null | null | null | true | null |
4114
| null |
Default
| null | null |
null |
{
"abstract": " Protoplanetary disks undergo substantial mass-loss by photoevaporation, a\nmechanism which is crucial to their dynamical evolution. However, the processes\nregulating the gas energetics have not been well constrained by observations so\nfar. We aim at studying the processes involved in disk photoevaporation when it\nis driven by far-UV photons. We present a unique Herschel survey and new ALMA\nobservations of four externally-illuminated photoevaporating disks (a.k.a.\nproplyds). For the analysis of these data, we developed a 1D model of the\nphotodissociation region (PDR) of a proplyd, based on the Meudon PDR code and\ncomputed the far infrared line emission. We successfully reproduce most of the\nobservations and derive key physical parameters, i.e. densities at the disk\nsurface of about $10^{6}$ cm$^{-3}$ and local gas temperatures of about 1000 K.\nOur modelling suggests that all studied disks are found in a transitional\nregime resulting from the interplay between several heating and cooling\nprocesses that we identify. These differ from those dominating in classical\nPDRs, i.e. grain photo-electric effect and cooling by [OI] and [CII] FIR lines.\nThis energetic regime is associated to an equilibrium dynamical point of the\nphotoevaporation flow: the mass-loss rate is self-regulated to set the envelope\ncolumn density at a value that maintains the temperature at the disk surface\naround 1000 K. From our best-fit models, we estimate mass-loss rates - of the\norder of $10^{-7}$ $\\mathrm{M}_\\odot$/yr - that are in agreement with earlier\nspectroscopic observation of ionised gas tracers. This holds only if we assume\nan evaporation flow launched from the disk surface at sound speed\n(supercritical regime). We have identified the energetic regime regulating\nFUV-photoevaporation in proplyds. This regime could be implemented into models\nof the dynamical evolution of protoplanetary disks.\n",
"title": "Herschel survey and modelling of externally-illuminated photoevaporating protoplanetary disks"
}
| null | null |
[
"Physics"
] | null | true | null |
4115
| null |
Validated
| null | null |
null |
{
"abstract": " We study the spatio-temporal instability generated by a universal unstable\nattractor in normal dispersion graded-index multimode fiber (GRIN MMF) for\nfemtosecond pulses. Our results present the generation of geometric parametric\ninstability (GPI) sidebands with ultrashort input pulse for the first time.\nObserved GPI sidebands are 91 THz detuned from the pump wavelength, 800 nm.\nDetailed analysis carried out numerically by employing coupled-mode pulse\npropagation model including optical shock and Raman nonlinearity terms. A\nsimplified theoretical model and numerically calculated spectra are\nwell-aligned with experimental results. For input pulses of 200-fs duration,\nformation and evolution of GPI are shown in both spatial and temporal domains.\nThe spatial intensity distribution of the total field and GPI sidebands are\ncalculated. Numerically and experimentally obtained beam shapes of first GPI\nfeatures a Gaussian-like beam profile. Our numerical results verify the unique\nfeature of GPI and generated sidebands preserve their inherited spatial\nintensity profile from the input pulse for different propagation distances\nparticularly for focused and spread the total field inside the GRIN MMF.\n",
"title": "Observation of Spatio-temporal Instability of Femtosecond Pulses in Normal Dispersion Multimode Graded-Index Fiber"
}
| null | null | null | null | true | null |
4116
| null |
Default
| null | null |
null |
{
"abstract": " Here we find the spectral curves, corresponding to the known rational or\nquasi-rational solutions of AKNS hierarchy equations, ultimately connected with\nthe modeling of the rogue waves events in the optical waveguides and in\nhydrodynamics. We also determine spectral curves for the multi-phase\ntrigonometric, hyperbolic and elliptic solutions for the same hierarchy. It\nseams that the nature of the related spectral curves was not sufficiently\ndiscussed in existing literature.\n",
"title": "Spectral curves for the rogue waves"
}
| null | null | null | null | true | null |
4117
| null |
Default
| null | null |
null |
{
"abstract": " We prove that a critical metric of the volume functional on a\nfour-dimensional compact manifold with boundary satisfying a second-order\nvanishing condition on the Weyl tensor must be isometric to a geodesic ball in\na simply connected space form $\\mathbb{R}^{4}$, $\\mathbb{H}^{4}$ or\n$\\mathbb{S}^{4}.$ Moreover, we provide an integral curvature estimate involving\nthe Yamabe constant for critical metrics of the volume functional, which allows\nus to get a rigidity result for such critical metrics on four-dimensional\nmanifolds.\n",
"title": "Volume functional of compact manifolds with a prescribed boundary metric"
}
| null | null | null | null | true | null |
4118
| null |
Default
| null | null |
null |
{
"abstract": " We present a deep neural architecture that parses sentences into three\nsemantic dependency graph formalisms. By using efficient, nearly arc-factored\ninference and a bidirectional-LSTM composed with a multi-layer perceptron, our\nbase system is able to significantly improve the state of the art for semantic\ndependency parsing, without using hand-engineered features or syntax. We then\nexplore two multitask learning approaches---one that shares parameters across\nformalisms, and one that uses higher-order structures to predict the graphs\njointly. We find that both approaches improve performance across formalisms on\naverage, achieving a new state of the art. Our code is open-source and\navailable at this https URL.\n",
"title": "Deep Multitask Learning for Semantic Dependency Parsing"
}
| null | null | null | null | true | null |
4119
| null |
Default
| null | null |
null |
{
"abstract": " Chentsov's theorem characterizes the Fisher information metric on statistical\nmodels as essentially the only Riemannian metric that is invariant under\nsufficient statistics. This implies that each statistical model is naturally\nequipped with a geometry, so Chentsov's theorem explains why many statistical\nproperties can be described in geometric terms. However, despite being one of\nthe foundational theorems of statistics, Chentsov's theorem has only been\nproved previously in very restricted settings or under relatively strong\nregularity and invariance assumptions. We therefore prove a version of this\ntheorem for the important case of exponential families. In particular, we\ncharacterise the Fisher information metric as the only Riemannian metric (up to\nrescaling) on an exponential family and its derived families that is invariant\nunder independent and identically distributed extensions and canonical\nsufficient statistics. Our approach is based on the central limit theorem, so\nit gives a unified proof for both discrete and continuous exponential families,\nand it is less technical than previous approaches.\n",
"title": "Chentsov's theorem for exponential families"
}
| null | null | null | null | true | null |
4120
| null |
Default
| null | null |
null |
{
"abstract": " The direct band gap character and large spin-orbit splitting of the valence\nband edges (at the K and K' valleys) in monolayer transition metal\ndichalcogenides have put these two-dimensional materials under the spot-light\nof intense experimental and theoretical studies. In particular, for Tungsten\ndichalcogenides it has been found that the sign of spin splitting of conduction\nband edges makes ground state excitons radiatively inactive (dark) due to spin\nand momentum mismatch between the constituent electron and hole. One might\nsimilarly assume that the ground states of charged excitons and biexcitons in\nthese monolayers are also dark. Here, we show that the intervalley\nK$\\leftrightarrows$K' electron-electron scattering mixes bright and dark states\nof these complexes, and estimate the radiative lifetimes in the ground states\nof these \"semi-dark\" trions and biexcitons to be ~ 10ps, and analyse how these\ncomplexes appear in the temperature-dependent photoluminescence spectra of WS2\nand WSe2 monolayers.\n",
"title": "Dark trions and biexcitons in WS2 and WSe2 made bright by e-e scattering"
}
| null | null | null | null | true | null |
4121
| null |
Default
| null | null |
null |
{
"abstract": " We consider the problem of finding the minimizer of a function $f:\n\\mathbb{R}^d \\rightarrow \\mathbb{R}$ of the finite-sum form $\\min f(w) =\n1/n\\sum_{i}^n f_i(w)$. This problem has been studied intensively in recent\nyears in the field of machine learning (ML). One promising approach for\nlarge-scale data is to use a stochastic optimization algorithm to solve the\nproblem. SGDLibrary is a readable, flexible and extensible pure-MATLAB library\nof a collection of stochastic optimization algorithms. The purpose of the\nlibrary is to provide researchers and implementers a comprehensive evaluation\nenvironment for the use of these algorithms on various ML problems.\n",
"title": "SGDLibrary: A MATLAB library for stochastic gradient descent algorithms"
}
| null | null | null | null | true | null |
4122
| null |
Default
| null | null |
null |
{
"abstract": " This paper is the first work to propose a network to predict a structured\nuncertainty distribution for a synthesized image. Previous approaches have been\nmostly limited to predicting diagonal covariance matrices. Our novel model\nlearns to predict a full Gaussian covariance matrix for each reconstruction,\nwhich permits efficient sampling and likelihood evaluation.\nWe demonstrate that our model can accurately reconstruct ground truth\ncorrelated residual distributions for synthetic datasets and generate plausible\nhigh frequency samples for real face images. We also illustrate the use of\nthese predicted covariances for structure preserving image denoising.\n",
"title": "Structured Uncertainty Prediction Networks"
}
| null | null | null | null | true | null |
4123
| null |
Default
| null | null |
null |
{
"abstract": " Starting from the pioneering works of Shannon and Weiner in 1948, a plethora\nof works have been reported on entropy in different directions. Entropy-related\nreview work in the direction of statistics, reliability and information\nscience, to the best of our knowledge, has not been reported so far. Here we\nhave tried to collect all possible works in this direction during the period\n1948-2018 so that people interested in entropy, specially the new researchers,\nget benefited.\n",
"title": "Shannon's entropy and its Generalizations towards Statistics, Reliability and Information Science during 1948-2018"
}
| null | null | null | null | true | null |
4124
| null |
Default
| null | null |
null |
{
"abstract": " A short overview demystifying the midi audio format is presented. The goal is\nto explain the file structure and how the instructions are used to produce a\nmusic signal, both in the case of monophonic signals as for polyphonic signals.\n",
"title": "Understanding MIDI: A Painless Tutorial on Midi Format"
}
| null | null | null | null | true | null |
4125
| null |
Default
| null | null |
null |
{
"abstract": " We develop the first Bayesian Optimization algorithm, BLOSSOM, which selects\nbetween multiple alternative acquisition functions and traditional local\noptimization at each step. This is combined with a novel stopping condition\nbased on expected regret. This pairing allows us to obtain the best\ncharacteristics of both local and Bayesian optimization, making efficient use\nof function evaluations while yielding superior convergence to the global\nminimum on a selection of optimization problems, and also halting optimization\nonce a principled and intuitive stopping condition has been fulfilled.\n",
"title": "Optimization, fast and slow: optimally switching between local and Bayesian optimization"
}
| null | null | null | null | true | null |
4126
| null |
Default
| null | null |
null |
{
"abstract": " The optical observations of wide fields of view encounter the problem of\nselection of best exposure time. As there are usually plenty of objects\nobserved simultaneously, the quality of photometry of the brightest ones is\nalways better than of the dimmer ones. Frequently all of them are equally\ninteresting for the astronomers and thus it is desired to have all of them\nmeasured with the highest possible accuracy.\nIn this paper we present a novel optimization algorithm dedicated for the\ndivision of exposure time into sub-exposures, which allows to perform\nphotometry with more balanced noise budget. Thanks to the proposed technique,\nthe photometric precision of dimmer objects is increased at the expense of the\nmeasurement fidelity of the brightest ones. We tested the method on real\nobservations using two telescope setups demonstrating its usefulness and good\nagreement with the theoretical expectations. The main application of our\napproach is a wide range of sky surveys, including the ones performed by the\nspace telescopes. The method can be applied for planning virtually any\nphotometric observations, in which the objects of interest show a wide range of\nmagnitudes.\n",
"title": "Optimization of exposure time division for wide field observations"
}
| null | null | null | null | true | null |
4127
| null |
Default
| null | null |
null |
{
"abstract": " Multi-Entity Dependence Learning (MEDL) explores conditional correlations\namong multiple entities. The availability of rich contextual information\nrequires a nimble learning scheme that tightly integrates with deep neural\nnetworks and has the ability to capture correlation structures among\nexponentially many outcomes. We propose MEDL_CVAE, which encodes a conditional\nmultivariate distribution as a generating process. As a result, the variational\nlower bound of the joint likelihood can be optimized via a conditional\nvariational auto-encoder and trained end-to-end on GPUs. Our MEDL_CVAE was\nmotivated by two real-world applications in computational sustainability: one\nstudies the spatial correlation among multiple bird species using the eBird\ndata and the other models multi-dimensional landscape composition and human\nfootprint in the Amazon rainforest with satellite images. We show that\nMEDL_CVAE captures rich dependency structures, scales better than previous\nmethods, and further improves on the joint likelihood taking advantage of very\nlarge datasets that are beyond the capacity of previous methods.\n",
"title": "Multi-Entity Dependence Learning with Rich Context via Conditional Variational Auto-encoder"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
4128
| null |
Validated
| null | null |
null |
{
"abstract": " A simple-triangle graph is the intersection graph of triangles that are\ndefined by a point on a horizontal line and an interval on another horizontal\nline. The time complexity of the recognition problem for simple-triangle graphs\nwas a longstanding open problem, which was recently settled. This paper\nprovides a new recognition algorithm for simple-triangle graphs to improve the\ntime bound from $O(n^2 \\overline{m})$ to $O(nm)$, where $n$, $m$, and\n$\\overline{m}$ are the number of vertices, edges, and non-edges of the graph,\nrespectively. The algorithm uses the vertex ordering characterization that a\ngraph is a simple-triangle graph if and only if there is a linear ordering of\nthe vertices containing both an alternating orientation of the graph and a\ntransitive orientation of the complement of the graph. We also show, as a\nbyproduct, that an alternating orientation can be obtained in $O(nm)$ time for\ncocomparability graphs, and it is NP-complete to decide whether a graph has an\norientation that is alternating and acyclic.\n",
"title": "A recognition algorithm for simple-triangle graphs"
}
| null | null | null | null | true | null |
4129
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we generalize the normalized gradient flow method to compute\nthe ground states of Bose-Einstein condensates (BEC) with higher order\ninteractions (HOI), which is modelled via the modified Gross-Pitaevskii\nequation (MGPE). Schemes constructed in naive ways suffer from severe stability\nproblems due to the high restrictions on time steps. To build an efficient and\nstable scheme, we split the HOI term into two parts with each part treated\nseparately. The part corresponding to a repulsive/positive energy is treated\nsemi-implicitly while the one corresponding to an attractive/negative energy is\ntreated fully explicitly. Based on the splitting, we construct the\nBEFD-splitting and BESP-splitting schemes. A variety of numerical experiments\nshows that the splitting will improve the stability of the schemes\nsignificantly. Besides, we will show that the methods can be applied to\nmultidimensional problems and to the computation of the first excited state as\nwell.\n",
"title": "A normalized gradient flow method with attractive-repulsive splitting for computing ground states of Bose-Einstein condensates with higher-order interaction"
}
| null | null | null | null | true | null |
4130
| null |
Default
| null | null |
null |
{
"abstract": " We show that in any nontrivial Hahn field with truncation as a primitive\noperation we can interpret the monadic second-order logic of the additive\nmonoid of natural numbers and are thus undecidable. We also specify a definable\nbinary relation on such a structure that has $\\SOP$ and $\\TP$.\n",
"title": "Truncation in Hahn Fields is Undecidable and Wild"
}
| null | null | null | null | true | null |
4131
| null |
Default
| null | null |
null |
{
"abstract": " Laser communication has advances in compared with radio frequency\ncommunication as result of much high carrier frequency from ultraviolet to near\ninfrared. Very narrow laser beam is possible to form with very high power\ndensity. But laser beam has high destruction and attenuation on clouds,\nturbulence, scattering on aerosols and molecules of the atmosphere. Low Earth\norbits (LEO), Middling Earth orbits (MEO) and partly Geosynchronous Earth orbit\n(GSO) satellites moving on the sky and laser light from satellites moves across\ndifferent turbulence conditions of the atmosphere, clouds, molecules of the\natmosphere H2O, O2, N2, CO, O3 and other. We performed unique experiments with\npropagation of laser beams from beacon of OPALE terminal of ARTEMIS satellite\nthrough thin clouds. We have found that small part of laser radiation is\nreceived from ahead point there the satellite will be after time of propagation\nof laser radiation from the satellite to telescope. It is in accordance with\ntheory of relativity for aberration of light during transition from moving to\nnot moving coordinate systems. It is positive effect for laser communication\nthrough the atmosphere and clouds because will be possible to develop a system\nfor reduce of the atmosphere turbulence during of laser communication from\nground to the satellites. The interest is what will be during propagation of\nlaser radiation from the satellite through strong clouds. The detail\ndescriptions of laser experiment with ARTEMIS GSO satellite through strong\nclouds and estimations of the laser power through strong clouds are presented\nin this paper. Accordingly we must search the optimal wave lengths and power of\nlasers for performs laser communication in different cloudy conditions.\n",
"title": "Direct measurement of laser aberration and ahead point from ARTEMIS satellite through strong clouds"
}
| null | null | null | null | true | null |
4132
| null |
Default
| null | null |
null |
{
"abstract": " This paper introduces the first deep neural network-based estimation metric\nfor the jigsaw puzzle problem. Given two puzzle piece edges, the neural network\npredicts whether or not they should be adjacent in the correct assembly of the\npuzzle, using nothing but the pixels of each piece. The proposed metric\nexhibits an extremely high precision even though no manual feature extraction\nis performed. When incorporated into an existing puzzle solver, the solution's\naccuracy increases significantly, achieving thereby a new state-of-the-art\nstandard.\n",
"title": "DNN-Buddies: A Deep Neural Network-Based Estimation Metric for the Jigsaw Puzzle Problem"
}
| null | null | null | null | true | null |
4133
| null |
Default
| null | null |
null |
{
"abstract": " Although adverse effects of attacks have been acknowledged in many\ncyber-physical systems, there is no system-theoretic comprehension of how a\ncompromised agent can leverage communication capabilities to maximize the\ndamage in distributed multi-agent systems. A rigorous analysis of\ncyber-physical attacks enables us to increase the system awareness against\nattacks and design more resilient control protocols. To this end, we will take\nthe role of the attacker to identify the worst effects of attacks on root nodes\nand non-root nodes in a distributed control system. More specifically, we show\nthat a stealthy attack on root nodes can mislead the entire network to a wrong\nunderstanding of the situation and even destabilize the synchronization\nprocess. This will be called the internal model principle for the attacker and\nwill intensify the urgency of designing novel control protocols to mitigate\nthese types of attacks.\n",
"title": "Attack Analysis for Distributed Control Systems: An Internal Model Principle Approach"
}
| null | null | null | null | true | null |
4134
| null |
Default
| null | null |
null |
{
"abstract": " A conservative scheme has been formulated and verified for gyrokinetic\nparticle simulations of electromagnetic waves and instabilities in magnetized\nplasmas. An electron continuity equation derived from drift kinetic equation is\nused to time advance electron density perturbation by using the perturbed\nmechanical flow calculated from the parallel vector potential, and the parallel\nvector potential is solved by using the perturbed canonical flow from the\nperturbed distribution function. In gyrokinetic particle simulations using this\nnew scheme, shear Alfvén wave dispersion relation in shearless slab and\ncontinuum damping in sheared cylinder have been recovered. The new scheme\novercomes the stringent requirement in conventional perturbative simulation\nmethod that perpendicular grid size needs to be as small as electron\ncollisionless skin depth even for the long wavelength Alfvén waves. The new\nscheme also avoids the problem in conventional method that an unphysically\nlarge parallel electric field arises due to the inconsistency between\nelectrostatic potential calculated from the perturbed density and vector\npotential calculated from the perturbed canonical flow. Finally, the\ngyrokinetic particle simulations of the Alfvén waves in sheared cylinder have\nsuperior numerical properties compared with the fluid simulations, which suffer\nfrom numerical difficulties associated with singular mode structures.\n",
"title": "A conservative scheme for electromagnetic simulation of magnetized plasmas with kinetic electrons"
}
| null | null | null | null | true | null |
4135
| null |
Default
| null | null |
null |
{
"abstract": " We study a variant of the stochastic multi-armed bandit (MAB) problem in\nwhich the rewards are corrupted. In this framework, motivated by privacy\npreservation in online recommender systems, the goal is to maximize the sum of\nthe (unobserved) rewards, based on the observation of transformation of these\nrewards through a stochastic corruption process with known parameters. We\nprovide a lower bound on the expected regret of any bandit algorithm in this\ncorrupted setting. We devise a frequentist algorithm, KLUCB-CF, and a Bayesian\nalgorithm, TS-CF and give upper bounds on their regret. We also provide the\nappropriate corruption parameters to guarantee a desired level of local privacy\nand analyze how this impacts the regret. Finally, we present some experimental\nresults that confirm our analysis.\n",
"title": "Corrupt Bandits for Preserving Local Privacy"
}
| null | null | null | null | true | null |
4136
| null |
Default
| null | null |
null |
{
"abstract": " Energy consumption for hot water production is a major draw in high\nefficiency buildings. Optimizing this has typically been approached from a\nthermodynamics perspective, decoupled from occupant influence. Furthermore,\noptimization usually presupposes existence of a detailed dynamics model for the\nhot water system. These assumptions lead to suboptimal energy efficiency in the\nreal world. In this paper, we present a novel reinforcement learning based\nmethodology which optimizes hot water production. The proposed methodology is\ncompletely generalizable, and does not require an offline step or human domain\nknowledge to build a model for the hot water vessel or the heating element.\nOccupant preferences too are learnt on the fly. The proposed system is applied\nto a set of 32 houses in the Netherlands where it reduces energy consumption\nfor hot water production by roughly 20% with no loss of occupant comfort.\nExtrapolating, this translates to absolute savings of roughly 200 kWh for a\nsingle household on an annual basis. This performance can be replicated to any\ndomestic hot water system and optimization objective, given that the fairly\nminimal requirements on sensor data are met. With millions of hot water systems\noperational worldwide, the proposed framework has the potential to reduce\nenergy consumption in existing and new systems on a multi Gigawatt-hour scale\nin the years to come.\n",
"title": "Deep Reinforcement Learning based Optimal Control of Hot Water Systems"
}
| null | null | null | null | true | null |
4137
| null |
Default
| null | null |
null |
{
"abstract": " Self-adaptive system (SAS) is capable of adjusting its behavior in response\nto meaningful changes in the operational context and itself. Due to the\ninherent volatility of the open and changeable environment in which SAS is\nembedded, the ability of adaptation is highly demanded by many\nsoftware-intensive systems. Two concerns, i.e., the requirements uncertainty\nand the context uncertainty are most important among others. An essential issue\nto be addressed is how to dynamically adapt non-functional requirements (NFRs)\nand task configurations of SASs with context uncertainty. In this paper, we\npropose a model-based fuzzy control approach that is underpinned by the\nfeedforward-feedback control mechanism. This approach identifies and represents\nNFR uncertainties, task uncertainties and context uncertainties with linguistic\nvariables, and then designs an inference structure and rules for the fuzzy\ncontroller based on the relations between the requirements model and the\ncontext model. The adaptation of NFRs and task configurations is achieved\nthrough fuzzification, inference, defuzzification and readaptation. Our\napproach is demonstrated with a mobile computing application and is evaluated\nthrough a series of simulation experiments.\n",
"title": "A Model-Based Fuzzy Control Approach to Achieving Adaptation with Contextual Uncertainties"
}
| null | null | null | null | true | null |
4138
| null |
Default
| null | null |
null |
{
"abstract": " The Intelligent Transportation System (ITS) targets to a coordinated traffic\nsystem by applying the advanced wireless communication technologies for road\ntraffic scheduling. Towards an accurate road traffic control, the short-term\ntraffic forecasting to predict the road traffic at the particular site in a\nshort period is often useful and important. In existing works, Seasonal\nAutoregressive Integrated Moving Average (SARIMA) model is a popular approach.\nThe scheme however encounters two challenges: 1) the analysis on related data\nis insufficient whereas some important features of data may be neglected; and\n2) with data presenting different features, it is unlikely to have one\npredictive model that can fit all situations. To tackle above issues, in this\nwork, we develop a hybrid model to improve accuracy of SARIMA. In specific, we\nfirst explore the autocorrelation and distribution features existed in traffic\nflow to revise structure of the time series model. Based on the Gaussian\ndistribution of traffic flow, a hybrid model with a Bayesian learning algorithm\nis developed which can effectively expand the application scenarios of SARIMA.\nWe show the efficiency and accuracy of our proposal using both analysis and\nexperimental studies. Using the real-world trace data, we show that the\nproposed predicting approach can achieve satisfactory performance in practice.\n",
"title": "See the Near Future: A Short-Term Predictive Methodology to Traffic Load in ITS"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
4139
| null |
Validated
| null | null |
null |
{
"abstract": " N-polar GaN p-n diodes are realized on single-crystal N-polar GaN bulk wafers\nby plasma-assisted molecular beam epitaxy growth. The current-voltage\ncharacteristics show high-quality rectification and electroluminescence\ncharacteristics with a high on/off current ratio and interband photon emission.\nThe measured electroluminescence spectrum is dominated by strong near-band edge\nemission, while deep level luminescence is greatly suppressed. A very low\ndislocation density leads to a high reverse breakdown electric field. The low\nleakage current N-polar diodes open up several potential applications in\npolarization-engineered photonic and electronic devices.\n",
"title": "Single-Crystal N-polar GaN p-n Diodes by Plasma-Assisted Molecular Beam Epitaxy"
}
| null | null |
[
"Physics"
] | null | true | null |
4140
| null |
Validated
| null | null |
null |
{
"abstract": " Particle identification at the Belle II experiment will be provided by two\nring imaging Cherenkov devices, the time of propagation counters in the central\nregion and the proximity focusing RICH with aerogel radiator in the forward\nend-cap region. The key features of these two detectors, the performance\nstudies, and the construction progress is presented.\n",
"title": "Particle Identification with the TOP and ARICH detectors at Belle II"
}
| null | null | null | null | true | null |
4141
| null |
Default
| null | null |
null |
{
"abstract": " This paper presents an easy and efficient face detection and face recognition\napproach using free software components from the internet. Face detection and\nface recognition problems have wide applications in home and office security.\nTherefore this work will helpful for those searching for a free face\noff-the-shelf face detection system. Using this system, faces can be detected\nin uncontrolled environments. In the detection phase, every individual face is\ndetected and in the recognition phase the detected faces are compared with the\nfaces in a given data set and recognized.\n",
"title": "Face Detection and Face Recognition In the Wild Using Off-the-Shelf Freely Available Components"
}
| null | null | null | null | true | null |
4142
| null |
Default
| null | null |
null |
{
"abstract": " Conventional decision trees have a number of favorable properties, including\ninterpretability, a small computational footprint and the ability to learn from\nlittle training data. However, they lack a key quality that has helped fuel the\ndeep learning revolution: that of being end-to-end trainable, and to learn from\nscratch those features that best allow to solve a given supervised learning\nproblem. Recent work (Kontschieder 2015) has addressed this deficit, but at the\ncost of losing a main attractive trait of decision trees: the fact that each\nsample is routed along a small subset of tree nodes only. We here propose a\nmodel and Expectation-Maximization training scheme for decision trees that are\nfully probabilistic at train time, but after a deterministic annealing process\nbecome deterministic at test time. We also analyze the learned oblique split\nparameters on image datasets and show that Neural Networks can be trained at\neach split node. In summary, we present the first end-to-end learning scheme\nfor deterministic decision trees and present results on par with or superior to\npublished standard oblique decision tree algorithms.\n",
"title": "End-to-end Learning of Deterministic Decision Trees"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
4143
| null |
Validated
| null | null |
null |
{
"abstract": " A topological shape analysis is proposed and utilized to learn concepts that\nreflect shape commonalities. Our approach is two-fold: i) a spatial topology\nanalysis of point cloud segment constellations within objects. Therein\nconstellations are decomposed and described in an hierarchical manner - from\nsingle segments to segment groups until a single group reflects an entire\nobject. ii) a topology analysis of the description space in which segment\ndecompositions are exposed in. Inspired by Persistent Homology, hidden groups\nof shape commonalities are revealed from object segment decompositions.\nExperiments show that extracted persistent groups of commonalities can\nrepresent semantically meaningful shape concepts. We also show the\ngeneralization capability of the proposed approach considering samples of\nexternal datasets.\n",
"title": "Conceptualization of Object Compositions Using Persistent Homology"
}
| null | null | null | null | true | null |
4144
| null |
Default
| null | null |
null |
{
"abstract": " Existing methods for arterial blood pressure (BP) estimation directly map the\ninput physiological signals to output BP values without explicitly modeling the\nunderlying temporal dependencies in BP dynamics. As a result, these models\nsuffer from accuracy decay over a long time and thus require frequent\ncalibration. In this work, we address this issue by formulating BP estimation\nas a sequence prediction problem in which both the input and target are\ntemporal sequences. We propose a novel deep recurrent neural network (RNN)\nconsisting of multilayered Long Short-Term Memory (LSTM) networks, which are\nincorporated with (1) a bidirectional structure to access larger-scale context\ninformation of input sequence, and (2) residual connections to allow gradients\nin deep RNN to propagate more effectively. The proposed deep RNN model was\ntested on a static BP dataset, and it achieved root mean square error (RMSE) of\n3.90 and 2.66 mmHg for systolic BP (SBP) and diastolic BP (DBP) prediction\nrespectively, surpassing the accuracy of traditional BP prediction models. On a\nmulti-day BP dataset, the deep RNN achieved RMSE of 3.84, 5.25, 5.80 and 5.81\nmmHg for the 1st day, 2nd day, 4th day and 6th month after the 1st day SBP\nprediction, and 1.80, 4.78, 5.0, 5.21 mmHg for corresponding DBP prediction,\nrespectively, which outperforms all previous models with notable improvement.\nThe experimental results suggest that modeling the temporal dependencies in BP\ndynamics significantly improves the long-term BP prediction accuracy.\n",
"title": "Long-term Blood Pressure Prediction with Deep Recurrent Neural Networks"
}
| null | null | null | null | true | null |
4145
| null |
Default
| null | null |
null |
{
"abstract": " We explore some of the ramifications arising from superflares on the\nevolutionary history of Earth, other planets in the Solar system, and\nexoplanets. We propose that the most powerful superflares can serve as\nplausible drivers of extinction events, and that their periodicity could\ncorrespond to certain patterns in the terrestrial fossil diversity record. On\nthe other hand, weaker superflares may play a positive role in enabling the\norigin of life through the formation of key organic compounds. Superflares\ncould also prove to be quite detrimental to the evolution of complex life on\npresent-day Mars and exoplanets in the habitable zone of M- and K-dwarfs. We\nconclude that the risk posed by superflares has not been sufficiently\nappreciated, and that humanity might potentially witness a superflare event in\nthe next $\\sim 10^3$ years leading to devastating economic and technological\nlosses. In light of the many uncertainties and assumptions associated with our\nanalysis, we recommend that these results should be viewed with due caution.\n",
"title": "Risks for life on habitable planets from superflares of their host stars"
}
| null | null | null | null | true | null |
4146
| null |
Default
| null | null |
null |
{
"abstract": " Quantum fluctuations from frustration can trigger quantum spin liquids (QSLs)\nat zero temperature. However, it is unclear how thermal fluctuations affect a\nQSL. We employ state-of-the-art tensor network-based methods to explore the\nground state and thermodynamic properties of the spin-1/2 kagome Heisenberg\nantiferromagnet (KHA). Its ground state is shown to be consistent with a\ngapless QSL by observing the absence of zero-magnetization plateau as well as\nthe algebraic behaviors of susceptibility and specific heat at low\ntemperatures, respectively. We show that there exists an \\textit{algebraic\nparamagnetic liquid} (APL) that possesses both the paramagnetic properties and\nthe algebraic behaviors inherited from the QSL. The APL is induced under the\ninterplay between quantum fluctuations from geometrical frustration and thermal\nfluctuations. By studying the temperature-dependent behaviors of specific heat\nand magnetic susceptibility, a finite-temperature phase diagram in a magnetic\nfield is suggested, where various phases are identified. This present study\ngains useful insight into the thermodynamic properties of the spin-1/2 KHA with\nor without a magnetic field and is helpful for relevant experimental studies.\n",
"title": "Thermodynamics of Spin-1/2 Kagomé Heisenberg Antiferromagnet: Algebraic Paramagnetic Liquid and Finite-Temperature Phase Diagram"
}
| null | null | null | null | true | null |
4147
| null |
Default
| null | null |
null |
{
"abstract": " A novel solution is obtained to solve the rigid 3D registration problem,\nmotivated by previous eigen-decomposition approaches. Different from existing\nsolvers, the proposed algorithm does not require sophisticated matrix\noperations e.g. singular value decomposition or eigenvalue decomposition.\nInstead, the optimal eigenvector of the point cross-covariance matrix can be\ncomputed within several iterations. It is also proven that the optimal rotation\nmatrix can be directly computed for cases without need of quaternion. The\nsimple framework provides very easy approach of integer-implementation on\nembedded platforms. Simulations on noise-corrupted point clouds have verified\nthe robustness and computation speed of the proposed method. The final results\nindicate that the proposed algorithm is accurate, robust and owns over $60\\%\n\\sim 80\\%$ less computation time than representatives. It has also been applied\nto real-world applications for faster relative robotic navigation.\n",
"title": "Fast Rigid 3D Registration Solution: A Simple Method Free of SVD and Eigen-Decomposition"
}
| null | null |
[
"Computer Science"
] | null | true | null |
4148
| null |
Validated
| null | null |
null |
{
"abstract": " Enhanced mobile broadband (eMBB) is one of the key use-cases for the\ndevelopment of the new standard 5G New Radio for the next generation of mobile\nwireless networks. Large-scale antenna arrays, a.k.a. Massive MIMO, the usage\nof carrier frequencies in the range 10-100 GHz, the so-called millimeter wave\n(mm-wave) band, and the network densification with the introduction of\nsmall-sized cells are the three technologies that will permit implementing eMBB\nservices and realizing the Gbit/s mobile wireless experience. This paper is\nfocused on the massive MIMO technology; initially conceived for conventional\ncellular frequencies in the sub-6 GHz range (\\mu-wave), the massive MIMO\nconcept has been then progressively extended to the case in which mm-wave\nfrequencies are used. However, due to different propagation mechanisms in urban\nscenarios, the resulting MIMO channel models at \\mu-wave and mm-wave are\nradically different. Six key basic differences are pinpointed in this paper,\nalong with the implications that they have on the architecture and algorithms\nof the communication transceivers and on the attainable performance in terms of\nreliability and multiplexing capabilities.\n",
"title": "Massive MIMO 5G Cellular Networks: mm-wave vs. μ-wave Frequencies"
}
| null | null |
[
"Computer Science"
] | null | true | null |
4149
| null |
Validated
| null | null |
null |
{
"abstract": " We study the problem of estimating the size of independent sets in a graph\n$G$ defined by a stream of edges. Our approach relies on the Caro-Wei bound,\nwhich expresses the desired quantity in terms of a sum over nodes of the\nreciprocal of their degrees, denoted by $\\beta(G)$. Our results show that\n$\\beta(G)$ can be approximated accurately, based on a provided lower bound on\n$\\beta$. Stronger results are possible when the edges are promised to arrive\ngrouped by an incident node. In this setting, we obtain a value that is at most\na logarithmic factor below the true value of $\\beta$ and no more than the true\nindependent set size. To justify the form of this bound, we also show an\n$\\Omega(n/\\beta)$ lower bound on any algorithm that approximates $\\beta$ up to\na constant factor.\n",
"title": "Independent Set Size Approximation in Graph Streams"
}
| null | null | null | null | true | null |
4150
| null |
Default
| null | null |
null |
{
"abstract": " This paper provides sufficient conditions for the existence of values and\nsolutions for two-person zero-sum one-step games with possibly noncompact\naction sets for both players and possibly unbounded payoff functions, which may\nbe neither convex nor concave. For such games payoffs may not be defined for\nsome pairs of strategies. In addition to the existence of values and solutions,\nthis paper investigates continuity properties of the value functions and\nsolution multifunctions for families of games with possibly noncompact action\nsets and unbounded payoff functions, when action sets and payoffs depend on a\nparameter.\n",
"title": "Two-Person Zero-Sum Games with Unbounded Payoff Functions and Uncertain Expected Payoffs"
}
| null | null | null | null | true | null |
4151
| null |
Default
| null | null |
null |
{
"abstract": " This article is the second in a series of two presenting the Scale\nRelativistic approach to non-differentiability in mechanics and its relation to\nquantum mechanics. Here, we show Schroedinger's equation to be a reformulation\nof Newton's fundamental relation of dynamics as generalized to\nnon-differentiable geometries in the first paper \\cite{paper1}. It motivates an\nalternative interpretation of the other axioms of standard quantum mechanics in\na coherent picture. This exercise validates the Scale Relativistic approach\nand, at the same time, it allows to identify macroscopic chaotic systems\nconsidered at time scales exceeding their horizon of predictability as\ncandidates in which to search for quantum-like structuring or behavior.\n",
"title": "Scale relativistic formulation of non-differentiable mechanics II: The Schroedinger picture"
}
| null | null | null | null | true | null |
4152
| null |
Default
| null | null |
null |
{
"abstract": " We present a simple apparatus for femtosecond laser induced generation of\nX-rays. The apparatus consists of a vacuum chamber containing an off-axis\nparabolic focusing mirror, a reel system, a debris protection setup, a quartz\nwindow for the incoming laser beam, and an X-ray window. Before entering the\nvacuum chamber, the femtosecond laser is expanded with an all reflective\ntelescope design to minimize laser intensity losses and pulse broadening while\nallowing for focusing as well as peak intensity optimization. The laser pulse\nduration was characterized by second-harmonic generation frequency resolved\noptical gating. A high spatial resolution knife-edge technique was implemented\nto characterize the beam size at the focus of the X-ray generation apparatus.\nWe have characterized x-ray spectra obtained with three different samples:\ntitanium, iron:chromium alloy, and copper. In all three cases, the femtosecond\nlaser generated X-rays give spectral lines consistent with literature reports.\nWe present a rms amplitude analysis of the generated X-ray pulses, and provide\nan upper bound for the duration of the X-ray pulses.\n",
"title": "Compact arrangement for femtosecond laser induced generation of broadband hard x-ray pulses"
}
| null | null |
[
"Physics"
] | null | true | null |
4153
| null |
Validated
| null | null |
null |
{
"abstract": " We study changes in metrics that are defined on a cartesian product of trees.\nSuch metrics occur naturally in many practical applications, where a global\nmetric (such as revenue) can be broken down along several hierarchical\ndimensions (such as location, gender, etc).\nGiven a change in such a metric, our goal is to identify a small set of\nnon-overlapping data segments that account for the change. An organization\ninterested in improving the metric can then focus their attention on these data\nsegments.\nOur key contribution is an algorithm that mimics the operation of a\nhierarchical organization of analysts. The algorithm has been successfully\napplied, for example within Google Adwords to help advertisers triage the\nperformance of their advertising campaigns.\nWe show that the algorithm is optimal for two dimensions, and has an\napproximation ratio $\\log^{d-2}(n+1)$ for $d \\geq 3$ dimensions, where $n$ is\nthe number of input data segments. For the Adwords application, we can show\nthat our algorithm is in fact a $2$-approximation.\nMathematically, we identify a certain data pattern called a \\emph{conflict}\nthat both guides the design of the algorithm, and plays a central role in the\nhardness results. We use these conflicts to both derive a lower bound of\n$1.144^{d-2}$ (again $d\\geq3$) for our algorithm, and to show that the problem\nis NP-hard, justifying the focus on approximation.\n",
"title": "Hierarchical Summarization of Metric Changes"
}
| null | null | null | null | true | null |
4154
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we answer the following question: what is the infinitesimal\ngenerator of the diffusion process defined by a kernel that is normalized such\nthat it is bi-stochastic with respect to a specified measure? More precisely,\nunder the assumption that data is sampled from a Riemannian manifold we\ndetermine how the resulting infinitesimal generator depends on the potentially\nnonuniform distribution of the sample points, and the specified measure for the\nbi-stochastic normalization. In a special case, we demonstrate a connection to\nthe heat kernel. We consider both the case where only a single data set is\ngiven, and the case where a data set and a reference set are given. The\nspectral theory of the constructed operators is studied, and Nyström\nextension formulas for the gradients of the eigenfunctions are computed.\nApplications to discrete point sets and manifold learning are discussed.\n",
"title": "Manifold learning with bi-stochastic kernels"
}
| null | null | null | null | true | null |
4155
| null |
Default
| null | null |
null |
{
"abstract": " The modification of geometry and interactions in two-dimensional magnetic\nnanosystems has enabled a range of studies addressing the magnetic order,\ncollective low-energy dynamics, and emergent magnetic properties, in e.g.\nartificial spin ice structures. The common denominator of all these\ninvestigations is the use of Ising-like mesospins as building blocks, in the\nform of elongated magnetic islands. Here we introduce a new approach: single\ninteraction modifiers, using slave-mesospins in the form of discs, within which\nthe mesospin is free to rotate in the disc plane. We show that by placing these\non the vertices of square artificial spin ice arrays and varying their\ndiameter, it is possible to tailor the strength and the ratio of the\ninteraction energies. We demonstrate the existence of degenerate ice-rule\nobeying states in square artificial spin ice structures, enabling the\nexploration of thermal dynamics in a spin liquid manifold. Furthermore, we even\nobserve the emergence of flux lattices on larger length-scales, when the energy\nlandscape of the vertices is reversed. The work highlights the potential of a\ndesign strategy for two-dimensional magnetic nano-architectures, through which\nmixed dimensionality of mesospins can be used to promote thermally emergent\nmesoscale magnetic states.\n",
"title": "The importance of the weak: Interaction modifiers in artificial spin ices"
}
| null | null |
[
"Physics"
] | null | true | null |
4156
| null |
Validated
| null | null |
null |
{
"abstract": " With the proliferation of social media, fashion inspired from celebrities,\nreputed designers as well as fashion influencers has shortened the cycle of\nfashion design and manufacturing. However, with the explosion of fashion\nrelated content and large number of user generated fashion photos, it is an\narduous task for fashion designers to wade through social media photos and\ncreate a digest of trending fashion. This necessitates deep parsing of fashion\nphotos on social media to localize and classify multiple fashion items from a\ngiven fashion photo. While object detection competitions such as MSCOCO have\nthousands of samples for each of the object categories, it is quite difficult\nto get large labeled datasets for fast fashion items. Moreover,\nstate-of-the-art object detectors do not have any functionality to ingest large\namount of unlabeled data available on social media in order to fine tune object\ndetectors with labeled datasets. In this work, we show application of a generic\nobject detector, that can be pretrained in an unsupervised manner, on 24\ncategories from recently released Open Images V4 dataset. We first train the\nbase architecture of the object detector using unsupervisd learning on 60K\nunlabeled photos from 24 categories gathered from social media, and then\nsubsequently fine tune it on 8.2K labeled photos from Open Images V4 dataset.\nOn 300 X 300 image inputs, we achieve 72.7% mAP on a test dataset of 2.4K\nphotos while performing 11% to 17% better as compared to the state-of-the-art\nobject detectors. We show that this improvement is due to our choice of\narchitecture that lets us do unsupervised learning and that performs\nsignificantly better in identifying small objects.\n",
"title": "How To Extract Fashion Trends From Social Media? A Robust Object Detector With Support For Unsupervised Learning"
}
| null | null | null | null | true | null |
4157
| null |
Default
| null | null |
null |
{
"abstract": " This paper introduces a new and effective algorithm for learning kernels in a\nMulti-Task Learning (MTL) setting. Although, we consider a MTL scenario here,\nour approach can be easily applied to standard single task learning, as well.\nAs shown by our empirical results, our algorithm consistently outperforms the\ntraditional kernel learning algorithms such as uniform combination solution,\nconvex combinations of base kernels as well as some kernel alignment-based\nmodels, which have been proven to give promising results in the past. We\npresent a Rademacher complexity bound based on which a new Multi-Task Multiple\nKernel Learning (MT-MKL) model is derived. In particular, we propose a Support\nVector Machine-regularized model in which, for each task, an optimal kernel is\nlearned based on a neighborhood-defining kernel that is not restricted to be\npositive semi-definite. Comparative experimental results are showcased that\nunderline the merits of our neighborhood-defining framework in both\nclassification and regression problems.\n",
"title": "Multi-Task Learning Using Neighborhood Kernels"
}
| null | null | null | null | true | null |
4158
| null |
Default
| null | null |
null |
{
"abstract": " Current spacecraft need to launch with all of their required fuel for travel.\nThis limits the system performance, payload capacity, and mission flexibility.\nOne compelling alternative is to perform In-Situ Resource Utilization (ISRU) by\nextracting fuel from small bodies in local space such as asteroids or small\nsatellites. Compared to the Moon or Mars, the microgravity on an asteroid\ndemands a fraction of the energy for digging and accessing hydrated regolith\njust below the surface. Previous asteroid excavation efforts have focused on\ndiscrete capture events (an extension of sampling technology) or whole-asteroid\ncapture and processing. This paper proposes an optimized bucket wheel design\nfor surface excavation of an asteroid or small-body. Asteroid regolith is\nexcavated and water extracted for use as rocket propellant. Our initial study\nfocuses on system design, bucket wheel mechanisms, and capture dynamics applied\nto ponded materials known to exist on asteroids like Itokawa and Eros and small\nsatellites like Phobos and Deimos. For initial evaluation of\nmaterial-spacecraft dynamics and mechanics, we assume lunar-like regolith for\nbulk density, particle size and cohesion. We shall present our estimates for\nthe energy balance of excavation and processing versus fuel gained.\nConventional electrolysis of water is used to produce hydrogen and oxygen. It\nis compared with steam for propulsion and both show significant delta-v. We\nshow that a return trip from Deimos to Earth is possible for a 12 kg craft\nusing ISRU processed fuel.\n",
"title": "Optimized Bucket Wheel Design for Asteroid Excavation"
}
| null | null |
[
"Computer Science",
"Physics"
] | null | true | null |
4159
| null |
Validated
| null | null |
null |
{
"abstract": " Compared with numerous X-ray dominant active galactic nuclei (AGNs) without\nemission-line signatures in their optical spectra, the X-ray selected AGNs with\noptical emission lines are probably still in the high-accretion phase of black\nhole growth. This paper presents an investigation on the fraction of these\nX-ray detected AGNs with optical emission-line spectra in 198 galaxy groups at\n$z<1$ in a rest frame 0.1-2.4 keV luminosity range 41.3 <log(L_X/erg s-1) <\n44.1 within the COSMOS field, as well as its variations with redshift and group\nrichness. For various selection criteria of member galaxies, the numbers of\ngalaxies and the AGNs with optical emission lines in each galaxy group are\nobtained. It is found that, in total 198 X-ray groups, there are 27 AGNs\ndetected in 26 groups. AGN fraction is on everage less than $4.6 (\\pm 1.2)\\%$\nfor individual groups hosting at least one AGN. The corrected overall AGN\nfraction for whole group sample is less than $0.98 (\\pm 0.11) \\%$. The\nnormalized locations of group AGNs show that 15 AGNs are found to be located in\ngroup centers, including all 6 low-luminosity group AGNs. A week rising\ntendency with $z$ are found: overall AGN fraction is 0.30-0.43% for the groups\nat $z<0.5$, and 0.55-0.64% at 0.5 < z < 1.0. For the X-ray groups at $z>0.5$,\nmost member AGNs are X-ray bright, optically dull, which results in a lower AGN\nfractions at higher redshifts. The AGN fraction in isolated fields also\nexhibits a rising trend with redshift, and the slope is consistent with that in\ngroups. The environment of galaxy groups seems to make no difference in\ndetection probability of the AGNs with emission lines. Additionally, a larger\nAGN fractions are found in poorer groups, which implies that the AGNs in poorer\ngroups might still be in the high-accretion phase, whereas the AGN population\nin rich clusters is mostly in the low-accretion, X-ray dominant phase.\n",
"title": "Fraction of the X-ray selected AGNs with optical emission lines in galaxy groups"
}
| null | null |
[
"Physics"
] | null | true | null |
4160
| null |
Validated
| null | null |
null |
{
"abstract": " We study the {\\em maximum duo-preservation string mapping} ({\\sc Max-Duo})\nproblem, which is the complement of the well studied {\\em minimum common string\npartition} ({\\sc MCSP}) problem. Both problems have applications in many fields\nincluding text compression and bioinformatics. Motivated by an earlier local\nsearch algorithm, we present an improved approximation and show that its\nperformance ratio is no greater than ${35}/{12} < 2.917$. This beats the\ncurrent best $3.25$-approximation for {\\sc Max-Duo}. The performance analysis\nof our algorithm is done through a complex yet interesting amortization. Two\nlower bounds on the locality gap of our algorithm are also provided.\n",
"title": "A local search 2.917-approximation algorithm for duo-preservation string mapping"
}
| null | null | null | null | true | null |
4161
| null |
Default
| null | null |
null |
{
"abstract": " The current trends in next-generation exascale systems go towards integrating\na wide range of specialized (co-)processors into traditional supercomputers.\nDue to the efficiency of heterogeneous systems in terms of Watts and FLOPS per\nsurface unit, opening the access of heterogeneous platforms to a wider range of\nusers is an important problem to be tackled. However, heterogeneous platforms\nlimit the portability of the applications and increase development complexity\ndue to the programming skills required. Program transformation can help make\nprogramming heterogeneous systems easier by defining a step-wise transformation\nprocess that translates a given initial code into a semantically equivalent\nfinal code, but adapted to a specific platform. Program transformation systems\nrequire the definition of efficient transformation strategies to tackle the\ncombinatorial problem that emerges due to the large set of transformations\napplicable at each step of the process. In this paper we propose a machine\nlearning-based approach to learn heuristics to define program transformation\nstrategies. Our approach proposes a novel combination of reinforcement learning\nand classification methods to efficiently tackle the problems inherent to this\ntype of systems. Preliminary results demonstrate the suitability of this\napproach.\n",
"title": "Towards Automatic Learning of Heuristics for Mechanical Transformations of Procedural Code"
}
| null | null | null | null | true | null |
4162
| null |
Default
| null | null |
null |
{
"abstract": " We propose a general approach to modeling semi-supervised learning (SSL)\nalgorithms. Specifically, we present a declarative language for modeling both\ntraditional supervised classification tasks and many SSL heuristics, including\nboth well-known heuristics such as co-training and novel domain-specific\nheuristics. In addition to representing individual SSL heuristics, we show that\nmultiple heuristics can be automatically combined using Bayesian optimization\nmethods. We experiment with two classes of tasks, link-based text\nclassification and relation extraction. We show modest improvements on\nwell-studied link-based classification benchmarks, and state-of-the-art results\non relation-extraction tasks for two realistic domains.\n",
"title": "Using Graphs of Classifiers to Impose Declarative Constraints on Semi-supervised Learning"
}
| null | null | null | null | true | null |
4163
| null |
Default
| null | null |
null |
{
"abstract": " We report the fabrication of a 1.2 cm long cavity directly on a nanofiber\nusing femtosecond laser ablation. The cavity modes with finesse value in the\nrange 200-400 can still maintain the transmission between 40-60%, which can\nenable \"strong-coupling\" regime of cavity QED for a single atom trapped 200 nm\naway from the fiber surface. For such cavity modes, we estimate the one-pass\nintra-cavity transmission to be 99.53%. Other cavity modes, which can enable\nhigh cooperativity in the range 3-10, show transmission over 60-85% and are\nsuitable for fiber-based single photon sources and quantum nonlinear optics in\nthe \"Purcell\" regime.\n",
"title": "Fabrication of a centimeter-long cavity on a nanofiber for cavity QED"
}
| null | null | null | null | true | null |
4164
| null |
Default
| null | null |
null |
{
"abstract": " Spectral clustering is one of the most popular, yet still incompletely\nunderstood, methods for community detection on graphs. In this article we study\nspectral clustering based on the deformed Laplacian matrix $D-rA$, for sparse\nheterogeneous graphs (following a two-class degree-corrected stochastic block\nmodel). For a specific value $r = \\zeta$, we show that, unlike competing\nmethods such as the Bethe Hessian or non-backtracking operator approaches,\nclustering is insensitive to the graph heterogeneity. Based on heuristic\narguments, we study the behavior of the informative eigenvector of $D-\\zeta A$\nand, as a result, we accurately predict the clustering accuracy. Via extensive\nsimulations and application to real networks, the resulting clustering\nalgorithm is validated and observed to systematically outperform\nstate-of-the-art competing methods.\n",
"title": "Optimized Deformed Laplacian for Spectrum-based Community Detection in Sparse Heterogeneous Graphs"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
4165
| null |
Validated
| null | null |
null |
{
"abstract": " We consider the recovery of a low rank $M \\times N$ matrix $S$ from its noisy\nobservation $\\tilde{S}$ in two different regimes. Under the assumption that $M$\nis comparable to $N$, we propose two consistent estimators for $S$. Our\nanalysis relies on the local behavior of the large dimensional rectangular\nmatrices with finite rank perturbation. We also derive the convergent limits\nand rates for the singular values and vectors of such matrices.\n",
"title": "High dimensional deformed rectangular matrices with applications in matrix denoising"
}
| null | null |
[
"Mathematics",
"Statistics"
] | null | true | null |
4166
| null |
Validated
| null | null |
null |
{
"abstract": " The aim of this work is to propose a first coarse-grained model of Bacillus\nsubtilis cell wall, handling explicitly the existence of multiple layers of\npeptidoglycans. In this first work, we aim at the validation of the recently\nproposed \"three under two\" principle.\n",
"title": "Validation of the 3-under-2 principle of cell wall growth in Gram-positive bacteria by simulation of a simple coarse-grained model"
}
| null | null | null | null | true | null |
4167
| null |
Default
| null | null |
null |
{
"abstract": " Self-bound quantum droplets are a newly discovered phase in the context of\nultracold atoms. In this work we report their experimental realization\nfollowing the original proposal by Petrov [Phys. Rev. Lett. 115, 155302\n(2015)], using an attractive bosonic mixture. In this system spherical droplets\nform due to the balance of competing attractive and repulsive forces, provided\nby the mean-field energy close to the collapse threshold and the first-order\ncorrection due to quantum fluctuations. Thanks to an optical levitating\npotential with negligible residual confinement we observe self-bound droplets\nin free space and we characterize the conditions for their formation as well as\ntheir equilibrium properties. This work sets the stage for future studies on\nquantum droplets, from the measurement of their peculiar excitation spectrum,\nto the exploration of their superfluid nature.\n",
"title": "Self-bound quantum droplets in atomic mixtures"
}
| null | null |
[
"Physics"
] | null | true | null |
4168
| null |
Validated
| null | null |
null |
{
"abstract": " This paper is concerned with the problem of stochastic control of gene\nregulatory networks (GRNs) observed indirectly through noisy measurements and\nwith uncertainty in the intervention inputs. The partial observability of the\ngene states and uncertainty in the intervention process are accounted for by\nmodeling GRNs using the partially-observed Boolean dynamical system (POBDS)\nsignal model with noisy gene expression measurements. Obtaining the optimal\ninfinite-horizon control strategy for this problem is not attainable in\ngeneral, and we apply reinforcement learning and Gaussian process techniques to\nfind a near-optimal solution. The POBDS is first transformed to a\ndirectly-observed Markov Decision Process in a continuous belief space, and the\nGaussian process is used for modeling the cost function over the belief and\nintervention spaces. Reinforcement learning then is used to learn the cost\nfunction from the available gene expression data. In addition, we employ\nsparsification, which enables the control of large partially-observed GRNs. The\nperformance of the resulting algorithm is studied through a comprehensive set\nof numerical experiments using synthetic gene expression data generated from a\nmelanoma gene regulatory network.\n",
"title": "Control of Gene Regulatory Networks with Noisy Measurements and Uncertain Inputs"
}
| null | null | null | null | true | null |
4169
| null |
Default
| null | null |
null |
{
"abstract": " Growing uncertainty in design parameters (and therefore, in design\nfunctionality) renders stochastic computing particularly promising, which\nrepresents and processes data as quantized probabilities. However, due to the\ndifference in data representation, integrating conventional memory (designed\nand optimized for non-stochastic computing) in stochastic computing systems\ninevitably incurs a significant data conversion overhead. Barely any stochastic\ncomputing proposal to-date covers the memory impact. In this paper, as the\nfirst study of its kind to the best of our knowledge, we rethink the memory\nsystem design for stochastic computing. The result is a seamless stochastic\nsystem, StochMem, which features analog memory to trade the energy and area\noverhead of data conversion for computation accuracy. In this manner StochMem\ncan reduce the energy (area) overhead by up-to 52.8% (93.7%) at the cost of at\nmost 0.7% loss in computation accuracy.\n",
"title": "On Memory System Design for Stochastic Computing"
}
| null | null |
[
"Computer Science"
] | null | true | null |
4170
| null |
Validated
| null | null |
null |
{
"abstract": " In this paper we propose a new approach to obtain mixing least square\nregression estimate by means of stochastic online mirror descent in\nnon-euclidian set-up.\n",
"title": "Strongly convex stochastic online optimization on a unit simplex with application to the mixing least square regression"
}
| null | null | null | null | true | null |
4171
| null |
Default
| null | null |
null |
{
"abstract": " Despite rapid advances in face recognition, there remains a clear gap between\nthe performance of still image-based face recognition and video-based face\nrecognition, due to the vast difference in visual quality between the domains\nand the difficulty of curating diverse large-scale video datasets. This paper\naddresses both of those challenges, through an image to video feature-level\ndomain adaptation approach, to learn discriminative video frame\nrepresentations. The framework utilizes large-scale unlabeled video data to\nreduce the gap between different domains while transferring discriminative\nknowledge from large-scale labeled still images. Given a face recognition\nnetwork that is pretrained in the image domain, the adaptation is achieved by\n(i) distilling knowledge from the network to a video adaptation network through\nfeature matching, (ii) performing feature restoration through synthetic data\naugmentation and (iii) learning a domain-invariant feature through a domain\nadversarial discriminator. We further improve performance through a\ndiscriminator-guided feature fusion that boosts high-quality frames while\neliminating those degraded by video domain-specific factors. Experiments on the\nYouTube Faces and IJB-A datasets demonstrate that each module contributes to\nour feature-level domain adaptation framework and substantially improves video\nface recognition performance to achieve state-of-the-art accuracy. We\ndemonstrate qualitatively that the network learns to suppress diverse artifacts\nin videos such as pose, illumination or occlusion without being explicitly\ntrained for them.\n",
"title": "Unsupervised Domain Adaptation for Face Recognition in Unlabeled Videos"
}
| null | null | null | null | true | null |
4172
| null |
Default
| null | null |
null |
{
"abstract": " We introduce a new class of graphical models that generalizes\nLauritzen-Wermuth-Frydenberg chain graphs by relaxing the semi-directed\nacyclity constraint so that only directed cycles are forbidden. Moreover, up to\ntwo edges are allowed between any pair of nodes. Specifically, we present\nlocal, pairwise and global Markov properties for the new graphical models and\nprove their equivalence. We also present an equivalent factorization property.\nFinally, we present a causal interpretation of the new models.\n",
"title": "Unifying DAGs and UGs"
}
| null | null | null | null | true | null |
4173
| null |
Default
| null | null |
null |
{
"abstract": " We argue that turning a logic program into a set of completed definitions can\nbe sometimes thought of as the \"reverse engineering\" process of generating a\nset of conditions that could serve as a specification for it. Accordingly, it\nmay be useful to define completion for a large class of ASP programs and to\nautomate the process of generating and simplifying completion formulas.\nExamining the output produced by this kind of software may help programmers to\nsee more clearly what their program does, and to what degree its behavior\nconforms with their expectations. As a step toward this goal, we propose here a\ndefinition of program completion for a large class of programs in the input\nlanguage of the ASP grounder GRINGO, and study its properties. This note is\nunder consideration for publication in Theory and Practice of Logic\nProgramming.\n",
"title": "Program Completionin the Input Language of GRINGO"
}
| null | null | null | null | true | null |
4174
| null |
Default
| null | null |
null |
{
"abstract": " In the present manuscript, we consider the practical problem of wave\ninteraction with a vertical wall. However, the novelty here consists in the\nfact that the wall can move horizontally due to a system of springs. The water\nwave evolution is described with the free surface potential flow model. Then, a\nsemi-analytical numerical method is presented. It is based on a mapping\ntechnique and a finite difference scheme in the transformed domain. The idea is\nto pose the equations on a fixed domain. This method is thoroughly tested and\nvalidated in our study. By choosing specific values of spring parameters, this\nsystem can be used to damp (or in other words to extract the energy of)\nincident water waves.\n",
"title": "Numerical modelling of surface water wave interaction with a moving wall"
}
| null | null | null | null | true | null |
4175
| null |
Default
| null | null |
null |
{
"abstract": " Human attribute analysis is a challenging task in the field of computer\nvision, since the data is largely imbalance-distributed. Common techniques such\nas re-sampling and cost-sensitive learning require prior-knowledge to train the\nsystem. To address this problem, we propose a unified framework called Dynamic\nCurriculum Learning (DCL) to online adaptively adjust the sampling strategy and\nloss learning in single batch, which resulting in better generalization and\ndiscrimination. Inspired by the curriculum learning, DCL consists of two level\ncurriculum schedulers: (1) sampling scheduler not only manages the data\ndistribution from imbalanced to balanced but also from easy to hard; (2) loss\nscheduler controls the learning importance between classification and metric\nlearning loss. Learning from these two schedulers, we demonstrate our DCL\nframework with the new state-of-the-art performance on the widely used face\nattribute dataset CelebA and pedestrian attribute dataset RAP.\n",
"title": "Dynamic Curriculum Learning for Imbalanced Data Classification"
}
| null | null | null | null | true | null |
4176
| null |
Default
| null | null |
null |
{
"abstract": " Measurement error in the observed values of the variables can greatly change\nthe output of various causal discovery methods. This problem has received much\nattention in multiple fields, but it is not clear to what extent the causal\nmodel for the measurement-error-free variables can be identified in the\npresence of measurement error with unknown variance. In this paper, we study\nprecise sufficient identifiability conditions for the measurement-error-free\ncausal model and show what information of the causal model can be recovered\nfrom observed data. In particular, we present two different sets of\nidentifiability conditions, based on the second-order statistics and\nhigher-order statistics of the data, respectively. The former was inspired by\nthe relationship between the generating model of the\nmeasurement-error-contaminated data and the factor analysis model, and the\nlatter makes use of the identifiability result of the over-complete independent\ncomponent analysis problem.\n",
"title": "Causal Discovery in the Presence of Measurement Error: Identifiability Conditions"
}
| null | null | null | null | true | null |
4177
| null |
Default
| null | null |
null |
{
"abstract": " We consider the K3 surfaces that arise as double covers of the elliptic\nmodular surface of level 5, $R_{5,5}$. Such surfaces have a natural elliptic\nfibration induced by the fibration on $R_{5,5}$. Moreover, they admit several\nother elliptic fibrations. We describe such fibrations in terms of linear\nsystems of curves on $R_{5,5}$. This has a major advantage over other methods\nof classification of elliptic fibrations, namely, a simple algorithm that has\nas input equations of linear systems of curves in the projective plane yields a\nWeierstrass equation for each elliptic fibration. We deal in detail with the\ncases for which the double cover is branched over the two reducible fibers of\ntype $I_5$ and for which it is branched over two smooth fibers, giving a\ncomplete list of elliptic fibrations for these two scenarios.\n",
"title": "Elliptic fibrations on covers of the elliptic modular surface of level 5"
}
| null | null |
[
"Mathematics"
] | null | true | null |
4178
| null |
Validated
| null | null |
null |
{
"abstract": " A side-fed crossed Dragone telescope provides a wide field-of-view. This type\nof a telescope is commonly employed in the measurement of cosmic microwave\nbackground (CMB) polarization, which requires an image-space telecentric\ntelescope with a large focal plane over broadband coverage. We report the\ndesign of the wide field-of-view crossed Dragone optical system using the\nanamorphic aspherical surfaces with correction terms up to the 10th order. We\nachieved the Strehl ratio larger than 0.95 over 32 by 18 square degrees at 150\nGHz. This design is an image-space telecentric and fully diffraction-limited\nsystem below 400 GHz. We discuss the optical performance in the uniformity of\nthe axially symmetric point spread function and telecentricity over the\nfield-of-view. We also address the analysis to evaluate the polarization\nproperties, including the instrumental polarization, extinction rate, and\npolarization angle rotation. This work is a part of programs to design a\ncompact multi-color wide field-of-view telescope for LiteBIRD, which is a next\ngeneration CMB polarization satellite.\n",
"title": "A wide field-of-view crossed Dragone optical system using the anamorphic aspherical surfaces"
}
| null | null | null | null | true | null |
4179
| null |
Default
| null | null |
null |
{
"abstract": " Individual Neurons in the nervous systems exploit various dynamics. To\ncapture these dynamics for single neurons, we tune the parameters of an\nelectrophysiological model of nerve cells, to fit experimental data obtained by\ncalcium imaging. A search for the biophysical parameters of this model is\nperformed by means of a genetic algorithm, where the model neuron is exposed to\na predefined input current representing overall inputs from other parts of the\nnervous system. The algorithm is then constrained for keeping the ion-channel\ncurrents within reasonable ranges, while producing the best fit to a calcium\nimaging time series of the AVA interneuron, from the brain of the soil-worm, C.\nelegans. Our settings enable us to project a set of biophysical parameters to\nthe the neuron kinetics observed in neuronal imaging.\n",
"title": "Searching for Biophysically Realistic Parameters for Dynamic Neuron Models by Genetic Algorithms from Calcium Imaging Recording"
}
| null | null | null | null | true | null |
4180
| null |
Default
| null | null |
null |
{
"abstract": " In Part 1 we study the spherical functions on compact symmetric pairs of\narbitrary rank under a suitable multiplicity freeness assumption and additional\nconditions on the branching rules. The spherical functions are taking values in\nthe spaces of linear operators of a finite dimensional representation of the\nsubgroup, so the spherical functions are matrix-valued. Under these assumptions\nthese functions can be described in terms of matrix-valued orthogonal\npolynomials in several variables, where the number of variables is the rank of\nthe compact symmetric pair. Moreover, these polynomials are uniquely determined\nas simultaneous eigenfunctions of a commutative algebra of differential\noperators.\nIn Part 2 we verify that the group case $\\mathrm{SU}(n+1)$ meets all the\nconditions that we impose in Part 1. For any $k\\in\\mathbb{N}_{0}$ we obtain\nfamilies of orthogonal polynomials in $n$ variables with values in the $N\\times\nN$-matrices, where $N=\\binom{n+k}{k}$. The case $k=0$ leads to the classical\nHeckman-Opdam polynomials of type $A_{n}$ with geometric parameter. For $k=1$\nwe obtain the most complete results. In this case we give an explicit\nexpression of the matrix weight, which we show to be irreducible whenever\n$n\\ge2$. We also give explicit expressions of the spherical functions that\ndetermine the matrix weight for $k=1$. These expressions are used to calculate\nthe spherical functions that determine the matrix weight for general $k$ up to\ninvertible upper-triangular matrices. This generalizes and gives a new proof of\na formula originally obtained by Koornwinder for the case $n=1$. The commuting\nfamily of differential operators that have the matrix-valued polynomials as\nsimultaneous eigenfunctions contains an element of order one. We give explicit\nformulas for differential operators of order one and two for $(n,k)$ equal to\n$(2,1)$ and $(3,1)$.\n",
"title": "Matrix elements of irreducible representations of $\\mathrm{SU}(n+1)\\times\\mathrm{SU}(n+1)$ and multivariable matrix-valued orthogonal polynomials"
}
| null | null | null | null | true | null |
4181
| null |
Default
| null | null |
null |
{
"abstract": " We consider the problem of estimating species trees from unrooted gene tree\ntopologies in the presence of incomplete lineage sorting, a common phenomenon\nthat creates gene tree heterogeneity in multilocus datasets. One popular class\nof reconstruction methods in this setting is based on internode distances, i.e.\nthe average graph distance between pairs of species across gene trees. While\nstatistical consistency in the limit of large numbers of loci has been\nestablished in some cases, little is known about the sample complexity of such\nmethods. Here we make progress on this question by deriving a lower bound on\nthe worst-case variance of internode distance which depends linearly on the\ncorresponding graph distance in the species tree. We also discuss some\nalgorithmic implications.\n",
"title": "On the variance of internode distance under the multispecies coalescent"
}
| null | null | null | null | true | null |
4182
| null |
Default
| null | null |
null |
{
"abstract": " Recently, the introduction of the generative adversarial network (GAN) and\nits variants has enabled the generation of realistic synthetic samples, which\nhas been used for enlarging training sets. Previous work primarily focused on\ndata augmentation for semi-supervised and supervised tasks. In this paper, we\ninstead focus on unsupervised anomaly detection and propose a novel generative\ndata augmentation framework optimized for this task. In particular, we propose\nto oversample infrequent normal samples - normal samples that occur with small\nprobability, e.g., rare normal events. We show that these samples are\nresponsible for false positives in anomaly detection. However, oversampling of\ninfrequent normal samples is challenging for real-world high-dimensional data\nwith multimodal distributions. To address this challenge, we propose to use a\nGAN variant known as the adversarial autoencoder (AAE) to transform the\nhigh-dimensional multimodal data distributions into low-dimensional unimodal\nlatent distributions with well-defined tail probability. Then, we\nsystematically oversample at the `edge' of the latent distributions to increase\nthe density of infrequent normal samples. We show that our oversampling\npipeline is a unified one: it is generally applicable to datasets with\ndifferent complex data distributions. To the best of our knowledge, our method\nis the first data augmentation technique focused on improving performance in\nunsupervised anomaly detection. We validate our method by demonstrating\nconsistent improvements across several real-world datasets.\n",
"title": "DOPING: Generative Data Augmentation for Unsupervised Anomaly Detection with GAN"
}
| null | null | null | null | true | null |
4183
| null |
Default
| null | null |
null |
{
"abstract": " Gestures are a natural communication modality for humans. The ability to\ninterpret gestures is fundamental for robots aiming to naturally interact with\nhumans. Wearable sensors are promising to monitor human activity, in particular\nthe usage of triaxial accelerometers for gesture recognition have been\nexplored. Despite this, the state of the art presents lack of systems for\nreliable online gesture recognition using accelerometer data. The article\nproposes SLOTH, an architecture for online gesture recognition, based on a\nwearable triaxial accelerometer, a Recurrent Neural Network (RNN) probabilistic\nclassifier and a procedure for continuous gesture detection, relying on\nmodelling gesture probabilities, that guarantees (i) good recognition results\nin terms of precision and recall, (ii) immediate system reactivity.\n",
"title": "Online Human Gesture Recognition using Recurrent Neural Networks and Wearable Sensors"
}
| null | null | null | null | true | null |
4184
| null |
Default
| null | null |
null |
{
"abstract": " When training a deep network for image classification, one can broadly\ndistinguish between two types of latent features of images that will drive the\nclassification. Following the notation of Gong et al. (2016), we can divide\nlatent features into (i) \"core\" features $X^\\text{core}$ whose distribution\n$X^\\text{core}\\vert Y$ does not change substantially across domains and (ii)\n\"style\" features $X^{\\text{style}}$ whose distribution $X^{\\text{style}}\\vert\nY$ can change substantially across domains. These latter orthogonal features\nwould generally include features such as rotation, image quality or brightness\nbut also more complex ones like hair color or posture for images of persons.\nGuarding against future adversarial domain shifts implies that the influence of\nthe second type of style features in the prediction has to be limited. We\nassume that the domain itself is not observed and hence a latent variable. We\ndo assume, however, that we can sometimes observe a typically discrete\nidentifier or $\\mathrm{ID}$ variable. We know in some applications, for\nexample, that two images show the same person, and $\\mathrm{ID}$ then refers to\nthe identity of the person. The method requires only a small fraction of images\nto have an $\\mathrm{ID}$ variable. We group data samples if they share the same\nclass and identifier $(Y,\\mathrm{ID})=(y,\\mathrm{id})$ and penalize the\nconditional variance of the prediction if we condition on $(Y,\\mathrm{ID})$.\nUsing this approach is shown to protect against shifts in the distribution of\nthe style variables for both regression and classification models.\nSpecifically, the conditional variance penalty CoRe is shown to be equivalent\nto minimizing the risk under noise interventions in a regression setting and is\nshown to lead to adversarial risk consistency in a partially linear\nclassification setting.\n",
"title": "Conditional Variance Penalties and Domain Shift Robustness"
}
| null | null | null | null | true | null |
4185
| null |
Default
| null | null |
null |
{
"abstract": " Deep narrow-band HST imaging of the iconic spiral galaxy M101 has revealed\nover a thousand new Wolf Rayet (WR) candidates. We report spectrographic\nconfirmation of 10 HeII emission line sources hosting 15 WR stars. We find WR\nstars present at both sub- and super-solar metalicities with WC stars favouring\nmore metal-rich regions compared to WN stars. We investigate the association of\nWR stars with HII regions using archival HST imaging and conclude that the\nmajority of WR stars are in or associated with HII regions. Of the 10 emission\nlines sources, only one appears to be unassociated with a star-forming region.\nOur spectroscopic survey provides confidence that our narrow-band photometric\ncandidates are in fact bonafide WR stars, which will allow us to characterise\nthe progenitors of any core-collapse supernovae that erupt in the future in\nM101.\n",
"title": "The First Optical Spectra of Wolf Rayet Stars in M101 Revealed with Gemini/GMOS"
}
| null | null |
[
"Physics"
] | null | true | null |
4186
| null |
Validated
| null | null |
null |
{
"abstract": " In this Letter we supervisedly train neural networks to distinguish different\ntopological phases in the context of topological band insulators. After\ntraining with Hamiltonians of one-dimensional insulators with chiral symmetry,\nthe neural network can predict their topological winding numbers with nearly\n100% accuracy, even for Hamiltonians with larger winding numbers that are not\nincluded in the training data. These results show a remarkable success that the\nneural network can capture the global and nonlinear topological features of\nquantum phases from local inputs. By opening up the neural network, we confirm\nthat the network does learn the discrete version of the winding number formula.\nWe also make a couple of remarks regarding the role of the symmetry and the\nopposite effect of regularization techniques when applying machine learning to\nphysical systems.\n",
"title": "Machine Learning Topological Invariants with Neural Networks"
}
| null | null | null | null | true | null |
4187
| null |
Default
| null | null |
null |
{
"abstract": " We propose a Label Propagation based algorithm for weakly supervised text\nclassification. We construct a graph where each document is represented by a\nnode and edge weights represent similarities among the documents. Additionally,\nwe discover underlying topics using Latent Dirichlet Allocation (LDA) and\nenrich the document graph by including the topics in the form of additional\nnodes. The edge weights between a topic and a text document represent level of\n\"affinity\" between them. Our approach does not require document level\nlabelling, instead it expects manual labels only for topic nodes. This\nsignificantly minimizes the level of supervision needed as only a few topics\nare observed to be enough for achieving sufficiently high accuracy. The Label\nPropagation Algorithm is employed on this enriched graph to propagate labels\namong the nodes. Our approach combines the advantages of Label Propagation\n(through document-document similarities) and Topic Modelling (for minimal but\nsmart supervision). We demonstrate the effectiveness of our approach on various\ndatasets and compare with state-of-the-art weakly supervised text\nclassification approaches.\n",
"title": "Topics and Label Propagation: Best of Both Worlds for Weakly Supervised Text Classification"
}
| null | null | null | null | true | null |
4188
| null |
Default
| null | null |
null |
{
"abstract": " Indian Buffet Process based models are an elegant way for discovering\nunderlying features within a data set, but inference in such models can be\nslow. Inferring underlying features using Markov chain Monte Carlo either\nrelies on an uncollapsed representation, which leads to poor mixing, or on a\ncollapsed representation, which leads to a quadratic increase in computational\ncomplexity. Existing attempts at distributing inference have introduced\nadditional approximation within the inference procedure. In this paper we\npresent a novel algorithm to perform asymptotically exact parallel Markov chain\nMonte Carlo inference for Indian Buffet Process models. We take advantage of\nthe fact that the features are conditionally independent under the\nbeta-Bernoulli process. Because of this conditional independence, we can\npartition the features into two parts: one part containing only the finitely\nmany instantiated features and the other part containing the infinite tail of\nuninstantiated features. For the finite partition, parallel inference is simple\ngiven the instantiation of features. But for the infinite tail, performing\nuncollapsed MCMC leads to poor mixing and hence we collapse out the features.\nThe resulting hybrid sampler, while being parallel, produces samples\nasymptotically from the true posterior.\n",
"title": "Parallel Markov Chain Monte Carlo for the Indian Buffet Process"
}
| null | null |
[
"Statistics"
] | null | true | null |
4189
| null |
Validated
| null | null |
null |
{
"abstract": " We propose a new model for formalizing reward collection problems on graphs\nwith dynamically generated rewards which may appear and disappear based on a\nstochastic model. The *robot routing problem* is modeled as a graph whose nodes\nare stochastic processes generating potential rewards over discrete time. The\nrewards are generated according to the stochastic process, but at each step, an\nexisting reward disappears with a given probability. The edges in the graph\nencode the (unit-distance) paths between the rewards' locations. On visiting a\nnode, the robot collects the accumulated reward at the node at that time, but\ntraveling between the nodes takes time. The optimization question asks to\ncompute an optimal (or epsilon-optimal) path that maximizes the expected\ncollected rewards.\nWe consider the finite and infinite-horizon robot routing problems. For\nfinite-horizon, the goal is to maximize the total expected reward, while for\ninfinite horizon we consider limit-average objectives. We study the\ncomputational and strategy complexity of these problems, establish NP-lower\nbounds and show that optimal strategies require memory in general. We also\nprovide an algorithm for computing epsilon-optimal infinite paths for arbitrary\nepsilon > 0.\n",
"title": "The Robot Routing Problem for Collecting Aggregate Stochastic Rewards"
}
| null | null | null | null | true | null |
4190
| null |
Default
| null | null |
null |
{
"abstract": " We study the estimation of integral type functionals $\\int_{0}^{t}f(X_{r})dr$\nfor a function $f$ and a $d$-dimensional càdlàg process $X$ with respect to\ndiscrete observations by a Riemann-sum estimator. Based on novel semimartingale\napproximations in the Fourier domain, central limit theorems are proved for\n$L^{2}$-Sobolev functions $f$ with fractional smoothness and continuous Itô\nsemimartingales $X$. General $L^{2}(\\mathbb{P})$-upper bounds on the error for\ncàdlàg processes are given under weak assumptions. These bounds combine and\ngeneralize all previously obtained results in the literature and apply also to\nnon-Markovian processes. Several detailed examples are discussed. As\napplication the approximation of local times for fractional Brownian motion is\nstudied. The optimality of the $L^{2}(\\mathbb{P})$-upper bounds is shown by\nproving the corresponding lower bounds in case of Brownian motion.\n",
"title": "Estimating occupation time functionals"
}
| null | null |
[
"Mathematics",
"Statistics"
] | null | true | null |
4191
| null |
Validated
| null | null |
null |
{
"abstract": " Emotional arousal increases activation and performance but may also lead to\nburnout in software development. We present the first version of a Software\nEngineering Arousal lexicon (SEA) that is specifically designed to address the\nproblem of emotional arousal in the software developer ecosystem. SEA is built\nusing a bootstrapping approach that combines word embedding model trained on\nissue-tracking data and manual scoring of items in the lexicon. We show that\nour lexicon is able to differentiate between issue priorities, which are a\nsource of emotional activation and then act as a proxy for arousal. The best\nperformance is obtained by combining SEA (428 words) with a previously created\ngeneral purpose lexicon by Warriner et al. (13,915 words) and it achieves\nCohen's d effect sizes up to 0.5.\n",
"title": "Bootstrapping a Lexicon for Emotional Arousal in Software Engineering"
}
| null | null |
[
"Computer Science"
] | null | true | null |
4192
| null |
Validated
| null | null |
null |
{
"abstract": " X-ray emission in young stellar objects (YSOs) is orders of magnitude more\nintense than in main sequence stars1,2, suggestive of cosmic ray irradiation of\nsurrounding accretion disks. Protoplanetary disk irradiation has been detected\naround YSOs by HERSCHEL3. In our solar system, short-lived 10Be (half-life =\n1.39 My4), which cannot be produced by stellar nucleosynthesis, was discovered\nin the oldest solar system solids, the calcium-aluminium-rich inclusions\n(CAIs)5. The high 10Be abundance, as well as detection of other irradiation\ntracers6,7, suggest 10Be likely originates from cosmic ray irradiation caused\nby solar flares8. Nevertheless, the nature of these flares (gradual or\nimpulsive), the target (gas or dust), and the duration and location of\nirradiation remain unknown. Here we use the vanadium isotopic composition,\ntogether with initial 10Be abundance to quantify irradiation conditions in the\nearly Solar System9. For the initial 10Be abundances recorded in CAIs, 50V\nexcesses of a few per mil relative to chondrites have been predicted10,11. We\nreport 50V excesses in CAIs up to 4.4 per mil that co-vary with 10Be abundance.\nTheir co-variation dictates that excess 50V and 10Be were synthesised through\nirradiation of refractory dust. Modelling of the production rate of 50V and\n10Be demonstrates that the dust was exposed to solar cosmic rays produced by\ngradual flares for less than 300 years at about 0.1 au from the protoSun.\n",
"title": "Early Solar System irradiation quantified by linked vanadium and beryllium isotope variations in meteorites"
}
| null | null |
[
"Physics"
] | null | true | null |
4193
| null |
Validated
| null | null |
null |
{
"abstract": " We consider reaction-diffusion equations and Korteweg-de Vries-Burgers (KdVB)\nequations, i.e. scalar conservation laws with diffusive-dispersive\nregularization. We review the existence of traveling wave solutions for these\ntwo classes of evolution equations. For classical equations the traveling wave\nproblem (TWP) for a local KdVB equation can be identified with the TWP for a\nreaction-diffusion equation. In this article we study this relationship for\nthese two classes of evolution equations with nonlocal diffusion/dispersion.\nThis connection is especially useful, if the TW equation is not studied\ndirectly, but the existence of a TWS is proven using one of the evolution\nequations instead. Finally, we present three models from fluid dynamics and\ndiscuss the TWP via its link to associated reaction-diffusion equations.\n",
"title": "Two classes of nonlocal Evolution Equations related by a shared Traveling Wave Problem"
}
| null | null |
[
"Mathematics"
] | null | true | null |
4194
| null |
Validated
| null | null |
null |
{
"abstract": " We present the first systematic analysis of read, write, and space\namplification in Linux file systems. While many researchers are tackling write\namplification in key-value stores, IO amplification in file systems has been\nlargely unexplored. We analyze data and metadata operations on five widely-used\nLinux file systems: ext2, ext4, XFS, btrfs, and F2FS. We find that data\noperations result in significant write amplification (2-32X) and that metadata\noperations have a large IO cost. For example, a single rename requires 648 KB\nwrite IO in btrfs. We also find that small random reads result in read\namplification of 2-13X. Based on these observations, we present the CReWS\nconjecture about the relationship between IO amplification, consistency, and\nstorage space utilization. We hope this paper spurs people to design future\nfile systems with less IO amplification, especially for non-volatile memory\ntechnologies.\n",
"title": "Analyzing IO Amplification in Linux File Systems"
}
| null | null | null | null | true | null |
4195
| null |
Default
| null | null |
null |
{
"abstract": " The features of collaboration patterns are often considered to be different\nfrom discipline to discipline. Meanwhile, collaborating among disciplines is an\nobvious feature emerged in modern scientific research, which incubates several\ninterdisciplines. The features of collaborations in and among the disciplines\nof biological, physical and social sciences are analyzed based on 52,803 papers\npublished in a multidisciplinary journal PNAS during 1999 to 2013. From those\ndata, we found similar transitivity and assortativity of collaboration patterns\nas well as the identical distribution type of collaborators per author and that\nof papers per author, namely a mixture of generalized Poisson and power-law\ndistributions. In addition, we found that interdisciplinary research is\nundertaken by a considerable fraction of authors, not just those with many\ncollaborators or those with many papers. This case study provides a window for\nunderstanding aspects of multidisciplinary and interdisciplinary collaboration\npatterns.\n",
"title": "Feature analysis of multidisciplinary scientific collaboration patterns based on PNAS"
}
| null | null | null | null | true | null |
4196
| null |
Default
| null | null |
null |
{
"abstract": " We study the generation of magnetic fields during inflation making use of a\ncoupling of the inflaton and moduli fields to electromagnetism via the photon\nkinetic term, and assuming that the coupling is an increasing function of time.\nWe demonstrate that the strong coupling problem of inflationary magnetogenesis\ncan be avoided by incorporating the destabilization of moduli fields after\ninflation. The magnetic field always dominates over the electric one, and thus\nthe severe constraints on the latter from backreaction, which are the demanding\nobstacles in the case of a decreasing coupling function, do not apply to the\ncurrent scenario. However, we show that this loophole to the strong coupling\nproblem comes at a price: the normalization of the amplitude of magnetic fields\nis determined by this coupling term and is therefore suppressed by a large\nfactor after the moduli destabilization completes. From this we conclude that\nthere is no self-consistent and generic realization of primordial\nmagnetogenesis producing scale-invariant fields in the case of an increasing\nkinetic coupling.\n",
"title": "Inflationary magneto-(non)genesis, increasing kinetic couplings, and the strong coupling problem"
}
| null | null |
[
"Physics"
] | null | true | null |
4197
| null |
Validated
| null | null |
null |
{
"abstract": " We present HornDroid, a new tool for the static analysis of information flow\nproperties in Android applications. The core idea underlying HornDroid is to\nuse Horn clauses for soundly abstracting the semantics of Android applications\nand to express security properties as a set of proof obligations that are\nautomatically discharged by an off-the-shelf SMT solver. This approach makes it\npossible to fine-tune the analysis in order to achieve a high degree of\nprecision while still using off-the-shelf verification tools, thereby\nleveraging the recent advances in this field. As a matter of fact, HornDroid\noutperforms state-of-the-art Android static analysis tools on benchmarks\nproposed by the community. Moreover, HornDroid is the first static analysis\ntool for Android to come with a formal proof of soundness, which covers the\ncore of the analysis technique: besides yielding correctness assurances, this\nproof allowed us to identify some critical corner-cases that affect the\nsoundness guarantees provided by some of the previous static analysis tools for\nAndroid.\n",
"title": "HornDroid: Practical and Sound Static Analysis of Android Applications by SMT Solving"
}
| null | null | null | null | true | null |
4198
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we study the robustness of network topologies. We use the\nconcept of percolation as measuring tool to assess the reliability polynomial\nof those systems which can be modeled as a general inhomogeneous random graph\nas well as scale-free random graph.\n",
"title": "Assessing the reliability polynomial based on percolation theory"
}
| null | null |
[
"Mathematics",
"Statistics"
] | null | true | null |
4199
| null |
Validated
| null | null |
null |
{
"abstract": " In [Mas82] and [Vee78] it was proved independently that almost every interval\nexchange transformation is uniquely ergodic. The Birkhoff ergodic theorem\nimplies that these maps mainly have uniformly distributed orbits. This raises\nthe question under which conditions the orbits yield low-discrepancy sequences.\nThe case of $n=2$ intervals corresponds to circle rotation, where conditions\nfor low-discrepancy are well-known. In this paper, we give corresponding\nconditions in the case $n=3$. Furthermore, we construct infinitely many\ninterval exchange transformations with low-discrepancy orbits for $n \\geq 4$.\nWe also show that these examples do not coincide with $LS$-sequences if $S \\geq\n2$.\n",
"title": "Interval Exchange Transformations and Low-Discrepancy"
}
| null | null | null | null | true | null |
4200
| null |
Default
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.