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": " Planetesimals may form from the gravitational collapse of dense particle\nclumps initiated by the streaming instability. We use simulations of\naerodynamically coupled gas-particle mixtures to investigate whether the\nproperties of planetesimals formed in this way depend upon the sizes of the\nparticles that participate in the instability. Based on three high resolution\nsimulations that span a range of dimensionless stopping time $6 \\times 10^{-3}\n\\leq \\tau \\leq 2$ no statistically significant differences in the initial\nplanetesimal mass function are found. The mass functions are fit by a\npower-law, ${\\rm d}N / {\\rm d}M_p \\propto M_p^{-p}$, with $p=1.5-1.7$ and\nerrors of $\\Delta p \\approx 0.1$. Comparing the particle density fields prior\nto collapse, we find that the high wavenumber power spectra are similarly\nindistinguishable, though the large-scale geometry of structures induced via\nthe streaming instability is significantly different between all three cases.\nWe interpret the results as evidence for a near-universal slope to the mass\nfunction, arising from the small-scale structure of streaming-induced\nturbulence.\n",
"title": "Evidence for universality in the initial planetesimal mass function"
} | null | null | null | null | true | null | 20701 | null | Default | null | null |
null | {
"abstract": " A covering system of the integers is a finite collection of modular residue\nclasses $\\{a_m \\bmod{m}\\}_{m \\in S}$ whose union is all integers. Given a\nfinite set $S$ of moduli, it is often difficult to tell whether there is a\nchoice of residues modulo elements of $S$ covering the integers. Hough has\nshown that if the smallest modulus in $S$ is at least $10^{16}$, then there is\nnone. However, the question of whether there is a covering of the integers with\nall odd moduli remains open. We consider multiplicative restrictions on the set\nof moduli to generalize Hough's negative solution to the minimum modulus\nproblem. In particular, we find that every covering system of the integers has\na modulus divisible by a prime number less than or equal to $19$. Hough and\nNielsen have shown that every covering system has a modulus divisible by either\n$2$ or $3$.\n",
"title": "On covering systems of integers"
} | null | null | null | null | true | null | 20702 | null | Default | null | null |
null | {
"abstract": " Asymptotic Safety, based on a non-Gaussian fixed point of the gravitational\nrenormalization group flow, provides an elegant mechanism for completing the\ngravitational force at sub-Planckian scales. At high energies the fixed point\ncontrols the scaling of couplings such that unphysical divergences are absent\nwhile the emergence of classical low-energy physics is linked to a crossover\nbetween two renormalization group fixed points. These features make Asymptotic\nSafety an attractive framework for cosmological model building. The resulting\nscenarios may naturally give rise to a quantum gravity driven inflationary\nphase in the very early universe and an almost scale-free fluctuation spectrum.\nMoreover, effective descriptions arising from an renormalization group\nimprovement permit a direct comparison to cosmological observations as, e.g.\nPlanck data.\n",
"title": "Asymptotically safe cosmology - a status report"
} | null | null | null | null | true | null | 20703 | null | Default | null | null |
null | {
"abstract": " Future predictions on sequence data (e.g., videos or audios) require the\nalgorithms to capture non-Markovian and compositional properties of high-level\nsemantics. Context-free grammars are natural choices to capture such\nproperties, but traditional grammar parsers (e.g., Earley parser) only take\nsymbolic sentences as inputs. In this paper, we generalize the Earley parser to\nparse sequence data which is neither segmented nor labeled. This generalized\nEarley parser integrates a grammar parser with a classifier to find the optimal\nsegmentation and labels, and makes top-down future predictions. Experiments\nshow that our method significantly outperforms other approaches for future\nhuman activity prediction.\n",
"title": "Generalized Earley Parser: Bridging Symbolic Grammars and Sequence Data for Future Prediction"
} | null | null | null | null | true | null | 20704 | null | Default | null | null |
null | {
"abstract": " This paper presents a modular in-pipeline climbing robot with a novel\ncompliant foldable OmniCrawler mechanism. The circular cross-section of the\nOmniCrawler module enables a holonomic motion to facilitate the alignment of\nthe robot in the direction of bends. Additionally, the crawler mechanism\nprovides a fair amount of traction, even on slippery surfaces. These advantages\nof crawler modules have been further supplemented by incorporating active\ncompliance in the module itself which helps to negotiate sharp bends in small\ndiameter pipes. The robot has a series of 3 such compliant foldable modules\ninterconnected by the links via passive joints. For the desirable pipe diameter\nand curvature of the bends, the spring stiffness value for each passive joint\nis determined by formulating a constrained optimization problem using the\nquasi-static model of the robot. Moreover, a minimum friction coefficient value\nbetween the module-pipe surface which can be vertically climbed by the robot\nwithout slipping is estimated. The numerical simulation results have further\nbeen validated by experiments on real robot prototype.\n",
"title": "COCrIP: Compliant OmniCrawler In-pipeline Robot"
} | null | null | null | null | true | null | 20705 | null | Default | null | null |
null | {
"abstract": " This paper presents a learning-based approach for impromptu trajectory\ntracking for non-minimum phase systems, i.e., systems with unstable inverse\ndynamics. Inversion-based feedforward approaches are commonly used for\nimproving tracking performance; however, these approaches are not directly\napplicable to non-minimum phase systems due to their inherent instability. In\norder to resolve the instability issue, existing methods have assumed that the\nsystem model is known and used pre-actuation or inverse approximation\ntechniques. In this work, we propose an approach for learning a stable,\napproximate inverse of a non-minimum phase baseline system directly from its\ninput-output data. Through theoretical discussions, simulations, and\nexperiments on two different platforms, we show the stability of our proposed\napproach and its effectiveness for high-accuracy, impromptu tracking. Our\napproach also shows that including more information in the training, as is\ncommonly assumed to be useful, does not lead to better performance but may\ntrigger instability and impact the effectiveness of the overall approach.\n",
"title": "An Inversion-Based Learning Approach for Improving Impromptu Trajectory Tracking of Robots with Non-Minimum Phase Dynamics"
} | null | null | null | null | true | null | 20706 | null | Default | null | null |
null | {
"abstract": " In this work decay estimates are derived for the solutions of 1-D linear\nparabolic PDEs with disturbances at both boundaries and distributed\ndisturbances. The decay estimates are given in the L2 and H1 norms of the\nsolution and discontinuous disturbances are allowed. Although an eigenfunction\nexpansion for the solution is exploited for the proof of the decay estimates,\nthe estimates do not require knowledge of the eigenvalues and the\neigenfunctions of the corresponding Sturm-Liouville operator. Examples show\nthat the obtained results can be applied for the stability analysis of\nparabolic PDEs with nonlocal terms.\n",
"title": "Decay Estimates for 1-D Parabolic PDEs with Boundary Disturbances"
} | null | null | null | null | true | null | 20707 | null | Default | null | null |
null | {
"abstract": " A simple and self-consistent approach has been proposed for simulation of the\nproton-induced soft error rate based on the heavy ion induced single event\nupset cross-section data and vice versa. The approach relies on the GEANT4\nassisted Monte Carlo simulation of the secondary particle LET spectra produced\nby nuclear interactions. The method has been validated with the relevant\nin-flight soft error rate data for space protons and heavy ions. An approximate\nanalytical relation is proposed and validated for a fast recalculation between\nthe two types of experimental data.\n",
"title": "GEANT4 Simulation of Nuclear Interaction Induced Soft Errors in Digital Nanoscale Electronics: Interrelation Between Proton and Heavy Ion Impacts"
} | null | null | null | null | true | null | 20708 | null | Default | null | null |
null | {
"abstract": " Superconducting linacs are capable of producing intense, stable, high-quality\nelectron beams that have found widespread applications in science and industry.\nThe 9-cell 1.3-GHz superconducting standing-wave accelerating RF cavity\noriginally developed for $e^+/e^-$ linear-collider applications [B. Aunes, {\\em\net al.} Phys. Rev. ST Accel. Beams {\\bf 3}, 092001 (2000)] has been broadly\nemployed in various superconducting-linac designs. In this paper we discuss the\ntransfer matrix of such a cavity and present its measurement performed at the\nFermilab Accelerator Science and Technology (FAST) facility. The experimental\nresults are found to be in agreement with analytical calculations and numerical\nsimulations.\n",
"title": "Analysis and Measurement of the Transfer Matrix of a 9-cell 1.3-GHz Superconducting Cavity"
} | null | null | null | null | true | null | 20709 | null | Default | null | null |
null | {
"abstract": " This paper introduces the concept of size-aware sharding to improve tail\nlatencies for in-memory key-value stores, and describes its implementation in\nthe Minos key-value store. Tail latencies are crucial in distributed\napplications with high fan-out ratios, because overall response time is\ndetermined by the slowest response. Size-aware sharding distributes requests\nfor keys to cores according to the size of the item associated with the key. In\nparticular, requests for small and large items are sent to disjoint subsets of\ncores. Size-aware sharding improves tail latencies by avoiding head-of-line\nblocking, in which a request for a small item gets queued behind a request for\na large item. Alternative size-unaware approaches to sharding, such as\nkeyhash-based sharding, request dispatching and stealing do not avoid\nhead-of-line blocking, and therefore exhibit worse tail latencies. The\nchallenge in implementing size-aware sharding is to maintain high throughput by\navoiding the cost of software dispatching and by achieving load balancing\nbetween different cores. Minos uses hardware dispatch for all requests for\nsmall items, which form the very large majority of all requests. It achieves\nload balancing by adapting the number of cores handling requests for small and\nlarge items to their relative presence in the workload. We compare Minos to\nthree state-of-the-art designs of in-memory KV stores. Compared to its closest\ncompetitor, Minos achieves a 99th percentile latency that is up to two orders\nof magnitude lower. Put differently, for a given value for the 99th percentile\nlatency equal to 10 times the mean service time, Minos achieves a throughput\nthat is up to 7.4 times higher.\n",
"title": "Size-aware Sharding For Improving Tail Latencies in In-memory Key-value Stores"
} | null | null | null | null | true | null | 20710 | null | Default | null | null |
null | {
"abstract": " To explain the unusual richness and compactness of the Abell 2744, we propose\na hypothesis that it may be a rich supercluster aligned along the sightline,\nand present a supporting evidence obtained numerically from the MultiDark\nPlanck 2 simulations with a linear box size of $1\\,h^{-1}$Gpc. Applying the\nfriends-of-friends (FoF) algorithm with a linkage length of $0.33$ to a sample\nof the cluster-size halos from the simulations, we identify the superclusters\nand investigate how many superclusters have filamentary branches that would\nappear to be similar to the Abell 2744 if the filamentary axis is aligned with\nthe sightline. Generating randomly a unit vector as a sightline at the position\nof the core member of each supercluster and projecting the positions of the\nmembers onto the plane perpendicular to the direction of the sightline, we\nmeasure two dimensional distances ($R_{2d}$) of the member halos from the core\nfor each supercluster. Defining a Abell 2744-like spuercluster as the one\nhaving a filamentary branch composed of eight or more members with $R_{2d}\\le\n1\\,$Mpc and masses comparable to those of the observed Abell 2744\nsubstructures, we find one Abell 2744-like supercluster at $z=0.3$ and two at\n$z=0$. Repeating the same analysis but with the data from the Big MultiDark\nPlanck simulations performed on a larger box of linear size of\n$2.5\\,h^{-1}$Mpc, we find that the number of the Abell 2744-like superclusters\nat $z=0$ increases up to eighteen, among which three are found more massive\nthan $5\\times 10^{15}\\,M_{\\odot}$.\n",
"title": "Abell 2744 may be a supercluster aligned along the sightline"
} | null | null | null | null | true | null | 20711 | null | Default | null | null |
null | {
"abstract": " The current work is done to see which artery has more chance of having\ncardiovascular diseases by measuring value of pressure gradient in the common\ncarotid artery (CCA) and ascending aorta according to age and gender. Pressure\ngradient is determined in the CCA and ascending aorta of presumed healthy\nvolunteers, having age between 10 and 60 years. A real 2D model of both aorta\nand common carotid artery is constructed for different age groups using\ncomputational fluid dynamics (CFD). Pressure gradient of both the arteries are\ncalculated and compared for different age groups and gender. It is found that\nwith increase in diameter of common carotid artery and ascending aorta with\nadvancing age pressure gradient decreases. The value of pressure gradient of\naorta is found less than common carotid artery in both cases of age and gender.\n",
"title": "Classification of Pressure Gradient of Human Common Carotid Artery and Ascending Aorta on the Basis of Age and Gender"
} | null | null | null | null | true | null | 20712 | null | Default | null | null |
null | {
"abstract": " The last decade has witnessed an increase of interest in the spatial analysis\nof structured point patterns over networks whose analysis is challenging\nbecause of geometrical complexities and unique methodological problems. In this\ncontext, it is essential to incorporate the network specificity into the\nanalysis as the locations of events are restricted to areas covered by line\nsegments. Relying on concepts originating from graph theory, we extend the\nnotions of first-order network intensity functions to second-order and local\nnetwork intensity functions. We consider two types of local indicators of\nnetwork association functions which can be understood as adaptations of the\nprimary ideas of local analysis on the plane. We develop the node-wise and\ncross-hierarchical type of local functions. A real dataset on urban\ndisturbances is also presented.\n",
"title": "Second-order and local characteristics of network intensity functions"
} | null | null | null | null | true | null | 20713 | null | Default | null | null |
null | {
"abstract": " In order to understand the physical hysteresis loops clearly, we constructed\na novel model, which is combined with the electric field, the temperature, and\nthe stress as one synthetically parameter. This model revealed the shape of\nhysteresis loop was determined by few variables in ferroelectric materials: the\nsaturation of polarization, the coercive field, the electric susceptibility and\nthe equivalent field. Comparison with experimental results revealed the model\ncan retrace polarization versus electric field and temperature. As a\napplications of this model, the calculate formula of energy storage efficiency,\nthe electrocaloric effect, and the P(E,T) function have also been included in\nthis article.\n",
"title": "Modeling of hysteresis loop and its applications in ferroelectric materials"
} | null | null | null | null | true | null | 20714 | null | Default | null | null |
null | {
"abstract": " Finding the reduced-dimensional structure is critical to understanding\ncomplex networks. Existing approaches such as spectral clustering are\napplicable only when the full network is explicitly observed. In this paper, we\nfocus on the online factorization and partition of implicit large-scale\nnetworks based on observations from an associated random walk. We formulate\nthis into a nonconvex stochastic factorization problem and propose an efficient\nand scalable stochastic generalized Hebbian algorithm. The algorithm is able to\nprocess dependent state-transition data dynamically generated by the underlying\nnetwork and learn a low-dimensional representation for each vertex. By applying\na diffusion approximation analysis, we show that the continuous-time limiting\nprocess of the stochastic algorithm converges globally to the \"principal\ncomponents\" of the Markov chain and achieves a nearly optimal sample\ncomplexity. Once given the learned low-dimensional representations, we further\napply clustering techniques to recover the network partition. We show that when\nthe associated Markov process is lumpable, one can recover the partition\nexactly with high probability. We apply the proposed approach to model the\ntraffic flow of Manhattan as city-wide random walks. By using our algorithm to\nanalyze the taxi trip data, we discover a latent partition of the Manhattan\ncity that closely matches the traffic dynamics.\n",
"title": "Online Factorization and Partition of Complex Networks From Random Walks"
} | null | null | null | null | true | null | 20715 | null | Default | null | null |
null | {
"abstract": " Two-sample summary-data Mendelian randomization (MR) has become a popular\nresearch design to estimate the causal effect of risk exposures. With the\nsample size of GWAS continuing to increase, it is now possible to utilize\ngenetic instruments that are only weakly associated with the exposure. To\nmaximize the statistical power of MR, we propose a genome-wide design where\nmore than a thousand genetic instruments are used. For the statistical\nanalysis, we use an empirical partially Bayes approach where instruments are\nweighted according to their strength, thus weak instruments bring less\nvariation to the estimator. The estimator is highly efficient with many weak\ngenetic instruments and is robust to balanced and/or sparse pleiotropy. We\napply our method to estimate the causal effect of body mass index (BMI) and\nmajor blood lipids on cardiovascular disease outcomes and obtain substantially\nshorter confidence intervals. Some new and statistically significant findings\nare: the estimated causal odds ratio of BMI on ischemic stroke is 1.19 (95% CI:\n1.07--1.32, p-value < 0.001); the estimated causal odds ratio of high-density\nlipoprotein cholesterol (HDL-C) on coronary artery disease (CAD) is 0.78 (95%\nCI 0.73--0.84, p-value < 0.001). However, the estimated effect of HDL-C becomes\nsubstantially smaller and statistically non-significant when we only use the\nstrong instruments. By employing a genome-wide design and robust statistical\nmethods, the statistical power of MR studies can be greatly improved. Our\nempirical results suggest that, even though the relationship between HDL-C and\nCAD appears to be highly heterogeneous, it may be too soon to completely\ndismiss the HDL hypothesis.\n",
"title": "Powerful genome-wide design and robust statistical inference in two-sample summary-data Mendelian randomization"
} | null | null | [
"Statistics"
]
| null | true | null | 20716 | null | Validated | null | null |
null | {
"abstract": " This paper contains two parts: the description of a real electrical system,\nwith many redundancies, reconfigurations and repairs, then the description of a\nreliability model of this system, based on the BDMP (Boolean logic Driven\nMarkov Processes) formalism and partial results of a reliability and\navailability calculation made from this model.\n",
"title": "A Benchmark on Reliability of Complex Discrete Systems: Emergency Power Supply of a Nuclear Power Plant"
} | null | null | null | null | true | null | 20717 | null | Default | null | null |
null | {
"abstract": " Recently, Riemannian Gaussian distributions were defined on spaces of\npositive-definite real and complex matrices. The present paper extends this\ndefinition to the space of positive-definite quaternion matrices. In order to\ndo so, it develops the Riemannian geometry of the space of positive-definite\nquaternion matrices, which is shown to be a Riemannian symmetric space of\nnon-positive curvature. The paper gives original formulae for the Riemannian\nmetric of this space, its geodesics, and distance function. Then, it develops\nthe theory of Riemannian Gaussian distributions, including the exact expression\nof their probability density, their sampling algorithm and statistical\ninference.\n",
"title": "Riemannian Gaussian distributions on the space of positive-definite quaternion matrices"
} | null | null | [
"Mathematics",
"Statistics"
]
| null | true | null | 20718 | null | Validated | null | null |
null | {
"abstract": " Power flow in a low voltage direct current grid (LVDC) is a non-linear\nproblem just as its counterpart ac. This paper demonstrates that, unlike in ac\ngrids, convergence and uniqueness of the solution can be guaranteed in this\ntype of grids. The result is not a linearization nor an approximation, but an\nanalysis of the set of non-linear algebraic equations, which is valid for any\nLVDC grid regardless its size, topology or load condition. Computer simulation\ncorroborate the theoretical analysis.\n",
"title": "Analisis of the power flow in Low Voltage DC grids"
} | null | null | [
"Mathematics"
]
| null | true | null | 20719 | null | Validated | null | null |
null | {
"abstract": " We prove (adjoint) bilinear restriction estimates for general phases at\ndifferent scales in the full non-endpoint mixed norm range, and give bounds\nwith a sharp and explicit dependence on the phases. These estimates have\napplications to high-low frequency interactions for solutions to partial\ndifferential equations, as well as to the linear restriction problem for\nsurfaces with degenerate curvature. As a consequence, we obtain new bilinear\nrestriction estimates for elliptic phases and wave/Klein-Gordon interactions in\nthe full bilinear range, and give a refined Strichartz inequality for the\nKlein-Gordon equation. In addition, we extend these bilinear estimates to hold\nin adapted function spaces by using a transference type principle which holds\nfor vector valued waves.\n",
"title": "Multi-scale bilinear restriction estimates for general phases"
} | null | null | null | null | true | null | 20720 | null | Default | null | null |
null | {
"abstract": " This note presents an algebraic theory of instruction sequences with\ninstructions for Turing tapes as basic instructions, the behaviours produced by\nthe instruction sequences concerned under execution, and the interaction\nbetween such behaviours and the Turing tapes provided by an execution\nenvironment. This theory provides a setting for investigating issues relating\nto computability and computational complexity that is more general than the\nclosely related Turing-machine models of computation. The theory is essentially\nan instantiation of a parameterized algebraic theory which is the basis of a\nline of research in which issues relating to a wide variety of subjects from\ncomputer science have been rigorously investigated thinking in terms of\ninstruction sequences.\n",
"title": "Program algebra for Turing-machine programs"
} | null | null | null | null | true | null | 20721 | null | Default | null | null |
null | {
"abstract": " How can we find patterns and anomalies in a tensor, or multi-dimensional\narray, in an efficient and directly interpretable way? How can we do this in an\nonline environment, where a new tensor arrives each time step? Finding patterns\nand anomalies in a tensor is a crucial problem with many applications,\nincluding building safety monitoring, patient health monitoring, cyber\nsecurity, terrorist detection, and fake user detection in social networks.\nStandard PARAFAC and Tucker decomposition results are not directly\ninterpretable. Although a few sampling-based methods have previously been\nproposed towards better interpretability, they need to be made faster, more\nmemory efficient, and more accurate.\nIn this paper, we propose CTD, a fast, accurate, and directly interpretable\ntensor decomposition method based on sampling. CTD-S, the static version of\nCTD, provably guarantees a high accuracy that is 17 ~ 83x more accurate than\nthat of the state-of-the-art method. Also, CTD-S is made 5 ~ 86x faster, and 7\n~ 12x more memory-efficient than the state-of-the-art method by removing\nredundancy. CTD-D, the dynamic version of CTD, is the first interpretable\ndynamic tensor decomposition method ever proposed. Also, it is made 2 ~ 3x\nfaster than already fast CTD-S by exploiting factors at previous time step and\nby reordering operations. With CTD, we demonstrate how the results can be\neffectively interpreted in the online distributed denial of service (DDoS)\nattack detection.\n",
"title": "CTD: Fast, Accurate, and Interpretable Method for Static and Dynamic Tensor Decompositions"
} | null | null | null | null | true | null | 20722 | null | Default | null | null |
null | {
"abstract": " Planning problems in partially observable environments cannot be solved\ndirectly with convolutional networks and require some form of memory. But, even\nmemory networks with sophisticated addressing schemes are unable to learn\nintelligent reasoning satisfactorily due to the complexity of simultaneously\nlearning to access memory and plan. To mitigate these challenges we introduce\nthe Memory Augmented Control Network (MACN). The proposed network architecture\nconsists of three main parts. The first part uses convolutions to extract\nfeatures and the second part uses a neural network-based planning module to\npre-plan in the environment. The third part uses a network controller that\nlearns to store those specific instances of past information that are necessary\nfor planning. The performance of the network is evaluated in discrete grid\nworld environments for path planning in the presence of simple and complex\nobstacles. We show that our network learns to plan and can generalize to new\nenvironments.\n",
"title": "Memory Augmented Control Networks"
} | null | null | [
"Computer Science"
]
| null | true | null | 20723 | null | Validated | null | null |
null | {
"abstract": " A message passing algorithm is derived for recovering communities within a\ngraph generated by a variation of the Barabási-Albert preferential\nattachment model. The estimator is assumed to know the arrival times, or order\nof attachment, of the vertices. The derivation of the algorithm is based on\nbelief propagation under an independence assumption. Two precursors to the\nmessage passing algorithm are analyzed: the first is a degree thresholding (DT)\nalgorithm and the second is an algorithm based on the arrival times of the\nchildren (C) of a given vertex, where the children of a given vertex are the\nvertices that attached to it. Comparison of the performance of the algorithms\nshows it is beneficial to know the arrival times, not just the number, of the\nchildren. The probability of correct classification of a vertex is\nasymptotically determined by the fraction of vertices arriving before it. Two\nextensions of Algorithm C are given: the first is based on joint likelihood of\nthe children of a fixed set of vertices; it can sometimes be used to seed the\nmessage passing algorithm. The second is the message passing algorithm.\nSimulation results are given.\n",
"title": "Community Recovery in a Preferential Attachment Graph"
} | null | null | null | null | true | null | 20724 | null | Default | null | null |
null | {
"abstract": " We explore the Borel complexity of some basic families of subsets of a\ncountable group (large, small, thin, sparse and other) defined by the size of\ntheir elements. Applying the obtained results to the Stone-Čech\ncompactification $\\beta G$ of $G$, we prove, in particular, that the closure of\nthe minimal ideal of $\\beta G$ is of type $F_{\\sigma\\delta}$.\n",
"title": "The descriptive look at the size of subsets of groups"
} | null | null | null | null | true | null | 20725 | null | Default | null | null |
null | {
"abstract": " Capable of significantly reducing cell size and enhancing spatial reuse,\nnetwork densification is shown to be one of the most dominant approaches to\nexpand network capacity. Due to the scarcity of available spectrum resources,\nnevertheless, the over-deployment of network infrastructures, e.g., cellular\nbase stations (BSs), would strengthen the inter-cell interference as well, thus\nin turn deteriorating the system performance. On this account, we investigate\nthe performance of downlink cellular networks in terms of user coverage\nprobability (CP) and network spatial throughput (ST), aiming to shed light on\nthe limitation of network densification. Notably, it is shown that both CP and\nST would be degraded and even diminish to be zero when BS density is\nsufficiently large, provided that practical antenna height difference (AHD)\nbetween BSs and users is involved to characterize pathloss. Moreover, the\nresults also reveal that the increase of network ST is at the expense of the\ndegradation of CP. Therefore, to balance the tradeoff between user and network\nperformance, we further study the critical density, under which ST could be\nmaximized under the CP constraint. Through a special case study, it follows\nthat the critical density is inversely proportional to the square of AHD. The\nresults in this work could provide helpful guideline towards the application of\nnetwork densification in the next-generation wireless networks.\n",
"title": "The Impact of Antenna Height Difference on the Performance of Downlink Cellular Networks"
} | null | null | null | null | true | null | 20726 | null | Default | null | null |
null | {
"abstract": " The information carrier of modern technologies is the electron charge whose\ntransport inevitably generates Joule heating. Spin-waves, the collective\nprecessional motion of electron spins, do not involve moving charges and thus\navoid Joule heating. In this respect, magnonic devices in which the information\nis carried by spin-waves attract interest for low-power computing. However\nimplementation of magnonic devices for practical use suffers from low spin-wave\nsignal and on/off ratio. Here we demonstrate that cubic anisotropic materials\ncan enhance spin-wave signals by improving spin-wave amplitude as well as group\nvelocity and attenuation length. Furthermore, cubic anisotropic material shows\nan enhanced on/off ratio through a laterally localized edge mode, which closely\nmimics the gate-controlled conducting channel in traditional field-effect\ntransistors. These attractive features of cubic anisotropic materials will\ninvigorate magnonics research towards wave-based functional devices.\n",
"title": "Spin-wave propagation in cubic anisotropic materials"
} | null | null | null | null | true | null | 20727 | null | Default | null | null |
null | {
"abstract": " In the typical framework for boolean games (BG) each player can change the\ntruth value of some propositional atoms, while attempting to make her goal\ntrue. In standard BG goals are propositional formulas, whereas in iterated BG\ngoals are formulas of Linear Temporal Logic. Both notions of BG are\ncharacterised by the fact that agents have exclusive control over their set of\natoms, meaning that no two agents can control the same atom. In the present\ncontribution we drop the exclusivity assumption and explore structures where an\natom can be controlled by multiple agents. We introduce Concurrent Game\nStructures with Shared Propositional Control (CGS-SPC) and show that they ac-\ncount for several classes of repeated games, including iterated boolean games,\ninfluence games, and aggregation games. Our main result shows that, as far as\nverification is concerned, CGS-SPC can be reduced to concurrent game structures\nwith exclusive control. This result provides a polynomial reduction for the\nmodel checking problem of specifications in Alternating-time Temporal Logic on\nCGS-SPC.\n",
"title": "Relaxing Exclusive Control in Boolean Games"
} | null | null | null | null | true | null | 20728 | null | Default | null | null |
null | {
"abstract": " Let $X$ be a connected open Riemann surface. Let $Y$ be an Oka domain in the\nsmooth locus of an analytic subvariety of $\\mathbb C^n$, $n\\geq 1$, such that\nthe convex hull of $Y$ is all of $\\mathbb C^n$. Let $\\mathscr O_*(X, Y)$ be the\nspace of nondegenerate holomorphic maps $X\\to Y$. Take a holomorphic $1$-form\n$\\theta$ on $X$, not identically zero, and let $\\pi:\\mathscr O_*(X,Y) \\to\nH^1(X,\\mathbb C^n)$ send a map $g$ to the cohomology class of $g\\theta$. Our\nmain theorem states that $\\pi$ is a Serre fibration. This result subsumes the\n1971 theorem of Kusunoki and Sainouchi that both the periods and the divisor of\na holomorphic form on $X$ can be prescribed arbitrarily. It also subsumes two\nparametric h-principles in minimal surface theory proved by Forstneric and\nLarusson in 2016.\n",
"title": "Representing de Rham cohomology classes on an open Riemann surface by holomorphic forms"
} | null | null | null | null | true | null | 20729 | null | Default | null | null |
null | {
"abstract": " Unidirectional control of optically induced spin waves in a rare-earth iron\ngarnet crystal is demonstrated. We observed the interference of two spin-wave\npackets with different initial phases generated by circularly polarized light\npulses. This interference results in unidirectional propagation if the\nspin-wave sources are spaced apart at 1/4 of the wavelength of the spin waves\nand the initial phase difference is set to pi/2. The propagating direction of\nthe spin wave is switched by the polarization helicity of the light pulses.\nMoreover, in a numerical simulation, applying more than two spin-wave sources\nwith a suitable polarization and spot shape, arbitrary manipulation of the spin\nwave by the phased array method was replicated.\n",
"title": "Unidirectional control of optically induced spin waves"
} | null | null | null | null | true | null | 20730 | null | Default | null | null |
null | {
"abstract": " This paper proposes a novel deep reinforcement learning (RL) architecture,\ncalled Value Prediction Network (VPN), which integrates model-free and\nmodel-based RL methods into a single neural network. In contrast to typical\nmodel-based RL methods, VPN learns a dynamics model whose abstract states are\ntrained to make option-conditional predictions of future values (discounted sum\nof rewards) rather than of future observations. Our experimental results show\nthat VPN has several advantages over both model-free and model-based baselines\nin a stochastic environment where careful planning is required but building an\naccurate observation-prediction model is difficult. Furthermore, VPN\noutperforms Deep Q-Network (DQN) on several Atari games even with\nshort-lookahead planning, demonstrating its potential as a new way of learning\na good state representation.\n",
"title": "Value Prediction Network"
} | null | null | null | null | true | null | 20731 | null | Default | null | null |
null | {
"abstract": " Generalized Linear Bandits (GLBs), a natural extension of the stochastic\nlinear bandits, has been popular and successful in recent years. However,\nexisting GLBs scale poorly with the number of rounds and the number of arms,\nlimiting their utility in practice. This paper proposes new, scalable solutions\nto the GLB problem in two respects. First, unlike existing GLBs, whose\nper-time-step space and time complexity grow at least linearly with time $t$,\nwe propose a new algorithm that performs online computations to enjoy a\nconstant space and time complexity. At its heart is a novel Generalized Linear\nextension of the Online-to-confidence-set Conversion (GLOC method) that takes\n\\emph{any} online learning algorithm and turns it into a GLB algorithm. As a\nspecial case, we apply GLOC to the online Newton step algorithm, which results\nin a low-regret GLB algorithm with much lower time and memory complexity than\nprior work. Second, for the case where the number $N$ of arms is very large, we\npropose new algorithms in which each next arm is selected via an inner product\nsearch. Such methods can be implemented via hashing algorithms (i.e.,\n\"hash-amenable\") and result in a time complexity sublinear in $N$. While a\nThompson sampling extension of GLOC is hash-amenable, its regret bound for\n$d$-dimensional arm sets scales with $d^{3/2}$, whereas GLOC's regret bound\nscales with $d$. Towards closing this gap, we propose a new hash-amenable\nalgorithm whose regret bound scales with $d^{5/4}$. Finally, we propose a fast\napproximate hash-key computation (inner product) with a better accuracy than\nthe state-of-the-art, which can be of independent interest. We conclude the\npaper with preliminary experimental results confirming the merits of our\nmethods.\n",
"title": "Scalable Generalized Linear Bandits: Online Computation and Hashing"
} | null | null | null | null | true | null | 20732 | null | Default | null | null |
null | {
"abstract": " The pressure dependence of the structural, magnetic and superconducting\ntransitions and of the superconducting upper critical field were studied in\nsulfur-substituted Fe(Se$_{1-x}$S$_{x}$). Resistance measurements were\nperformed on single crystals with three substitution levels ($x$=0.043, 0.096,\n0.12) under hydrostatic pressures up to 1.8 GPa and in magnetic fields up to 9\nT, and compared to data on pure FeSe. Our results illustrate the effects of\nchemical and physical pressure on Fe(Se$_{1-x}$S$_{x}$). On increasing sulfur\ncontent, magnetic order in the low-pressure range is strongly suppressed to a\nsmall dome-like region in the phase diagrams. However, $T_s$ is much less\nsuppressed by sulfur substitution and $T_c$ of Fe(Se$_{1-x}$S$_{x}$) exhibits\nsimilar non-monotonic pressure dependence with a local maximum and a local\nminimum present in the low pressure range for all $x$. The local maximum in\n$T_c$ coincides with the emergence of the magnetic order above $T_c$. At this\npressure the slope of the upper critical field decreases abruptly. The minimum\nof $T_c$ correlates with a broad maximum of the upper critical field slope\nnormalized by $T_c$.\n",
"title": "Dome of magnetic order inside the nematic phase of sulfur-substituted FeSe under pressure"
} | null | null | null | null | true | null | 20733 | null | Default | null | null |
null | {
"abstract": " This paper describes Luminoso's participation in SemEval 2017 Task 2,\n\"Multilingual and Cross-lingual Semantic Word Similarity\", with a system based\non ConceptNet. ConceptNet is an open, multilingual knowledge graph that focuses\non general knowledge that relates the meanings of words and phrases. Our\nsubmission to SemEval was an update of previous work that builds high-quality,\nmultilingual word embeddings from a combination of ConceptNet and\ndistributional semantics. Our system took first place in both subtasks. It\nranked first in 4 out of 5 of the separate languages, and also ranked first in\nall 10 of the cross-lingual language pairs.\n",
"title": "ConceptNet at SemEval-2017 Task 2: Extending Word Embeddings with Multilingual Relational Knowledge"
} | null | null | null | null | true | null | 20734 | null | Default | null | null |
null | {
"abstract": " We determine the symmetrized topological complexity of the circle, using\nprimarily just general topology.\n",
"title": "The symmetrized topological complexity of the circle"
} | null | null | [
"Mathematics"
]
| null | true | null | 20735 | null | Validated | null | null |
null | {
"abstract": " We introduce intertwining operators among twisted modules or twisted\nintertwining operators associated to not-necessarily-commuting automorphisms of\na vertex operator algebra. Let $V$ be a vertex operator algebra and let\n$g_{1}$, $g_{2}$ and $g_{3}$ be automorphisms of $V$. We prove that for\n$g_{1}$-, $g_{2}$- and $g_{3}$-twisted $V$-modules $W_{1}$, $W_{2}$ and\n$W_{3}$, respectively, such that the vertex operator map for $W_{3}$ is\ninjective, if there exists a twisted intertwining operator of type\n${W_{3}\\choose W_{1}W_{2}}$ such that the images of its component operators\nspan $W_{3}$, then $g_{3}=g_{1}g_{2}$. We also construct what we call the\nskew-symmetry and contragredient isomorphisms between spaces of twisted\nintertwining operators among twisted modules of suitable types. The proofs of\nthese results involve careful analysis of the analytic extensions corresponding\nto the actions of the not-necessarily-commuting automorphisms of the vertex\noperator algebra.\n",
"title": "Intertwining operators among twisted modules associated to not-necessarily-commuting automorphisms"
} | null | null | null | null | true | null | 20736 | null | Default | null | null |
null | {
"abstract": " The ever-increasing architectural complexity in contemporary ASIC projects\nturns Design Verification (DV) into a highly advanced endeavor. Pressing needs\nfor short time-to-market has made automation a key solution in DV. However,\nrecurring execution of large regression suites inevitably leads to challenging\namounts of test results. Following the design science paradigm, we present an\naction research study to introduce visual analytics in a commercial ASIC\nproject. We develop a cityscape visualization tool using the game engine Unity.\nInitial evaluations are promising, suggesting that the tool offers a novel\napproach to identify error-prone parts of the design, as well as coverage\nholes.\n",
"title": "Enabling Visual Design Verification Analytics - From Prototype Visualizations to an Analytics Tool using the Unity Game Engine"
} | null | null | null | null | true | null | 20737 | null | Default | null | null |
null | {
"abstract": " Cooperative behavior in real social dilemmas is often perceived as a\nphenomenon emerging from norms and punishment. To overcome this paradigm, we\nhighlight the interplay between the influence of social networks on\nindividuals, and the activation of spontaneous self-regulating mechanisms,\nwhich may lead them to behave cooperatively, while interacting with others and\ntaking conflicting decisions over time. By extending Evolutionary game theory\nover networks, we prove that cooperation partially or fully emerges whether\nself-regulating mechanisms are sufficiently stronger than social pressure.\nInterestingly, even few cooperative individuals act as catalyzing agents for\nthe cooperation of others, thus activating a recruiting mechanism, eventually\ndriving the whole population to cooperate.\n",
"title": "Self-regulation promotes cooperation in social networks"
} | null | null | null | null | true | null | 20738 | null | Default | null | null |
null | {
"abstract": " Binary random compacts with different proportions of small (volume V) and\nlarge (volume 2V) bare maghemite nanoparticles (NPs) are used to investigate\nthe effect of controllably broadening the particle size distribution on the\nmagnetic properties of magnetic NP assemblies with strong dipolar interaction.\nA series of eight random mixtures of highly uniform 9.0 and 11.5 nm diameter\nmaghemite particles prepared by thermal decomposition are studied. In spite of\nseverely broadened size distributions in the mixed samples, well defined\nsuperspin glass transition temperatures are observed across the series, their\nvalues increasing linearly with the weight fraction of large particles.\n",
"title": "Magnetic properties of nanoparticles compacts with controlled broadening of the particle size distribution"
} | null | null | null | null | true | null | 20739 | null | Default | null | null |
null | {
"abstract": " Many real-world networks known as attributed networks contain two types of\ninformation: topology information and node attributes. It is a challenging task\non how to use these two types of information to explore structural\nregularities. In this paper, by characterizing potential relationship between\nlink communities and node attributes, a principled statistical model named\nPSB_PG that generates link topology and node attributes is proposed. This model\nfor generating links is based on the stochastic blockmodels following a Poisson\ndistribution. Therefore, it is capable of detecting a wide range of network\nstructures including community structures, bipartite structures and other\nmixture structures. The model for generating node attributes assumes that node\nattributes are high dimensional and sparse and also follow a Poisson\ndistribution. This makes the model be uniform and the model parameters can be\ndirectly estimated by expectation-maximization (EM) algorithm. Experimental\nresults on artificial networks and real networks containing various structures\nhave shown that the proposed model PSB_PG is not only competitive with the\nstate-of-the-art models, but also provides good semantic interpretation for\neach community via the learned relationship between the community and its\nrelated attributes.\n",
"title": "A Generative Model for Exploring Structure Regularities in Attributed Networks"
} | null | null | null | null | true | null | 20740 | null | Default | null | null |
null | {
"abstract": " We compute the maximal halfspace depth for a class of permutation-invariant\ndistributions on the probability simplex. The derivations are based on\nstochastic ordering results that so far were only showed to be relevant for the\nBehrens-Fisher problem.\n",
"title": "On the maximal halfspace depth of permutation-invariant distributions on the simplex"
} | null | null | null | null | true | null | 20741 | null | Default | null | null |
null | {
"abstract": " This letter is about a principal weakness of the published article by Li et\nal. in 2014. It seems that the mentioned work has a terrible conceptual mistake\nwhile presenting its theoretical approach. In fact, the work has tried to\ndesign a new attack and its effective solution for a basic watermarking\nalgorithm by Zhu et al. published in 2013, however in practice, we show the Li\net al.'s approach is not correct to obtain the aim. For disproof of the\nincorrect approach, we only apply a numerical example as the counterexample of\nthe Li et al.'s approach.\n",
"title": "Theoretical Evaluation of Li et al.'s Approach for Improving a Binary Watermark-Based Scheme in Remote Sensing Data Communications"
} | null | null | null | null | true | null | 20742 | null | Default | null | null |
null | {
"abstract": " Many of the current scientific advances in the life sciences have their\norigin in the intensive use of data for knowledge discovery. In no area this is\nso clear as in bioinformatics, led by technological breakthroughs in data\nacquisition technologies. It has been argued that bioinformatics could quickly\nbecome the field of research generating the largest data repositories, beating\nother data-intensive areas such as high-energy physics or astroinformatics.\nOver the last decade, deep learning has become a disruptive advance in machine\nlearning, giving new live to the long-standing connectionist paradigm in\nartificial intelligence. Deep learning methods are ideally suited to\nlarge-scale data and, therefore, they should be ideally suited to knowledge\ndiscovery in bioinformatics and biomedicine at large. In this brief paper, we\nreview key aspects of the application of deep learning in bioinformatics and\nmedicine, drawing from the themes covered by the contributions to an ESANN 2018\nspecial session devoted to this topic.\n",
"title": "Bioinformatics and Medicine in the Era of Deep Learning"
} | null | null | null | null | true | null | 20743 | null | Default | null | null |
null | {
"abstract": " Capabilities of detecting temporal relations between two events can benefit\nmany applications. Most of existing temporal relation classifiers were trained\nin a supervised manner. Instead, we explore the observation that regular event\npairs show a consistent temporal relation despite of their various contexts,\nand these rich contexts can be used to train a contextual temporal relation\nclassifier, which can further recognize new temporal relation contexts and\nidentify new regular event pairs. We focus on detecting after and before\ntemporal relations and design a weakly supervised learning approach that\nextracts thousands of regular event pairs and learns a contextual temporal\nrelation classifier simultaneously. Evaluation shows that the acquired regular\nevent pairs are of high quality and contain rich commonsense knowledge and\ndomain specific knowledge. In addition, the weakly supervised trained temporal\nrelation classifier achieves comparable performance with the state-of-the-art\nsupervised systems.\n",
"title": "A Weakly Supervised Approach to Train Temporal Relation Classifiers and Acquire Regular Event Pairs Simultaneously"
} | null | null | null | null | true | null | 20744 | null | Default | null | null |
null | {
"abstract": " We study the central problem in data privacy: how to share data with an\nanalyst while providing both privacy and utility guarantees to the user that\nowns the data. In this setting, we present an estimation-theoretic analysis of\nthe privacy-utility trade-off (PUT). Here, an analyst is allowed to reconstruct\n(in a mean-squared error sense) certain functions of the data (utility), while\nother private functions should not be reconstructed with distortion below a\ncertain threshold (privacy). We demonstrate how $\\chi^2$-information captures\nthe fundamental PUT in this case and provide bounds for the best PUT. We\npropose a convex program to compute privacy-assuring mappings when the\nfunctions to be disclosed and hidden are known a priori and the data\ndistribution is known. We derive lower bounds on the minimum mean-squared error\nof estimating a target function from the disclosed data and evaluate the\nrobustness of our approach when an empirical distribution is used to compute\nthe privacy-assuring mappings instead of the true data distribution. We\nillustrate the proposed approach through two numerical experiments.\n",
"title": "Privacy with Estimation Guarantees"
} | null | null | null | null | true | null | 20745 | null | Default | null | null |
null | {
"abstract": " We point out that there is a simple variant of the SYK model, which we call\ncSYK, that is $SL(2,R)$ invariant for all values of the coupling. The\nmodification consists of replacing the UV part of the SYK action with a\nquadratic bilocal term. The corresponding bulk dual is a non-gravitational\ntheory in a rigid AdS$_2$ background. At weak coupling cSYK is a generalized\nfree field theory; at strong coupling, it approaches the infrared of SYK. The\nexistence of this line of fixed points explains the previously found connection\nbetween the three-point function of bilinears in these two theories at large\n$q$.\n",
"title": "A line of CFTs: from generalized free fields to SYK"
} | null | null | null | null | true | null | 20746 | null | Default | null | null |
null | {
"abstract": " We derive integrable equations starting from autonomous mappings with a\ngeneral form inspired by the additive systems associated to the affine Weyl\ngroup E$_8^{(1)}$. By deautonomisation we obtain two hitherto unknown systems,\none of which turns out to be a linearisable one, and we show that both these\nsystems arise from the deautonomisation of a non-QRT mapping. In order to\nunambiguously prove the integrability of these nonautonomous systems, we\nintroduce a series of Miura transformations which allows us to prove that one\nof these systems is indeed a discrete Painlevé equation, related to the\naffine Weyl group E$_7^{(1)}$, and to cast it in canonical form. A similar\nsequence of Miura transformations allows us to effectively linearise the second\nsystem we obtain. An interesting off-shoot of our calculations is that the\nseries of Miura transformations, when applied at the autonomous limit, allows\none to transform a non-QRT invariant into a QRT one.\n",
"title": "Miura transformations for discrete Painlevé equations coming from the affine E$_8$ Weyl group"
} | null | null | null | null | true | null | 20747 | null | Default | null | null |
null | {
"abstract": " Molecular dynamics is based on solving Newton's equations for many-particle\nsystems that evolve along complex, highly fluctuating trajectories. The orbital\ninstability and short-time complexity of Newtonian orbits is in sharp contrast\nto the more coherent behavior of collective modes such as density profiles. The\nnotion of virtual molecular dynamics is introduced here based on temporal\ncoarse-graining via Pade approximants and the Ito formula for stochastic\nprocesses. It is demonstrated that this framework leads to significant\nefficiency over traditional molecular dynamics and avoids the need to introduce\ncoarse-grained variables and phenomenological equations for their evolution. In\nthis framework, an all-atom trajectory is represented by a Markov chain of\nvirtual atomic states at a discrete sequence of timesteps, transitions between\nwhich are determined by an integration of conventional molecular dynamics with\nPade approximants and a microstate energy annealing methodology. The latter is\nachieved by a conventional and an MD NVE energy minimization schemes. This\nmultiscale framework is demonstrated for a pertussis toxin subunit undergoing a\nstructural transition, a T=1 capsid-like structure of HPV16 L1 protein, and two\ncoalescing argon droplets.\n",
"title": "Virtual Molecular Dynamics"
} | null | null | null | null | true | null | 20748 | null | Default | null | null |
null | {
"abstract": " Contamination of covariates by measurement error is a classical problem in\nmultivariate regression, where it is well known that failing to account for\nthis contamination can result in substantial bias in the parameter estimators.\nThe nature and degree of this effect on statistical inference is also\nunderstood to crucially depend on the specific distributional properties of the\nmeasurement error in question. When dealing with functional covariates,\nmeasurement error has thus far been modelled as additive white noise over the\nobservation grid. Such a setting implicitly assumes that the error arises\npurely at the discrete sampling stage, otherwise the model can only be viewed\nin a weak (stochastic differential equation) sense, white noise not being a\nsecond-order process. Departing from this simple distributional setting can\nhave serious consequences for inference, similar to the multivariate case, and\ncurrent methodology will break down. In this paper, we consider the case when\nthe additive measurement error is allowed to be a valid stochastic process. We\npropose a novel estimator of the slope parameter in a functional linear model,\nfor scalar as well as functional responses, in the presence of this general\nmeasurement error specification. The proposed estimator is inspired by the\nmultivariate regression calibration approach, but hinges on recent advances on\nmatrix completion methods for functional data in order to handle the nontrivial\n(and unknown) error covariance structure. The asymptotic properties of the\nproposed estimators are derived. We probe the performance of the proposed\nestimator of slope using simulations and observe that it substantially improves\nupon the spectral truncation estimator based on the erroneous observations,\ni.e., ignoring measurement error. We also investigate the behaviour of the\nestimators on a real dataset on hip and knee angle curves during a gait cycle.\n",
"title": "Regression with genuinely functional errors-in-covariates"
} | null | null | null | null | true | null | 20749 | null | Default | null | null |
null | {
"abstract": " We consider the process $\\widehat\\Lambda_n-\\Lambda_n$, where $\\Lambda_n$ is a\ncadlag step estimator for the primitive $\\Lambda$ of a nonincreasing function\n$\\lambda$ on $[0,1]$, and $\\widehat\\Lambda_n$ is the least concave majorant of\n$\\Lambda_n$. We extend the results in Kulikov and Lopuhaä (2006, 2008) to the\ngeneral setting considered in Durot (2007). Under this setting we prove that a\nsuitably scaled version of $\\widehat\\Lambda_n-\\Lambda_n$ converges in\ndistribution to the corresponding process for two-sided Brownian motion with\nparabolic drift and we establish a central limit theorem for the $L_p$-distance\nbetween $\\widehat\\Lambda_n$ and $\\Lambda_n$.\n",
"title": "The distance between a naive cumulative estimator and its least concave majorant"
} | null | null | null | null | true | null | 20750 | null | Default | null | null |
null | {
"abstract": " This paper studies the approximate and null controllability for impulse\ncontrolled systems of heat equations coupled by a pair (A,B) of constant\nmatrices. We present a necessary and sufficient condition for the approximate\ncontrollability, which is exactly Kalman's controllability rank condition of\n(A,B). We prove that when such a system is approximately controllable, the\napproximate controllability over an interval [0,T] can be realized by adding\ncontrols at arbitrary n different control instants\n0<\\tau_1<\\tau_2<\\cdots<\\tau_n<T, provided that \\tau_n-\\tau_1<d_A, where\nd_A=\\min\\{\\pi/|Im \\lambda| : \\lambda\\in \\sigma(A)\\}. We also show that in\ngeneral, such systems are not null controllable.\n",
"title": "Controllability of impulse controlled systems of heat equations coupled by constant matrices"
} | null | null | null | null | true | null | 20751 | null | Default | null | null |
null | {
"abstract": " In this work we study two families of codes with availability, namely private\ninformation retrieval (PIR) codes and batch codes. While the former requires\nthat every information symbol has $k$ mutually disjoint recovering sets, the\nlatter asks this property for every multiset request of $k$ information\nsymbols. The main problem under this paradigm is to minimize the number of\nredundancy symbols. We denote this value by $r_P(n,k), r_B(n,k)$, for PIR,\nbatch codes, respectively, where $n$ is the number of information symbols.\nPrevious results showed that for any constant $k$, $r_P(n,k) =\n\\Theta(\\sqrt{n})$ and $r_B(n,k)=O(\\sqrt{n}\\log(n)$. In this work we study the\nasymptotic behavior of these codes for non-constant $k$ and specifically for\n$k=\\Theta(n^\\epsilon)$. We also study the largest value of $k$ such that the\nrate of the codes approaches 1, and show that for all $\\epsilon<1$,\n$r_P(n,n^\\epsilon) = o(n)$, while for batch codes, this property holds for all\n$\\epsilon< 0.5$.\n",
"title": "Nearly Optimal Constructions of PIR and Batch Codes"
} | null | null | null | null | true | null | 20752 | null | Default | null | null |
null | {
"abstract": " We present a CO(2-1) mosaic map of the spiral galaxy NGC 6946 by combining\ndata from the Submillimeter Array and the IRAM 30 m telescope. We identify 390\ngiant molecular clouds (GMCs) from the nucleus to 4.5 kpc in the disk. GMCs in\nthe inner 1 kpc are generally more luminous and turbulent, some of which have\nluminosities >10^6 K km/s pc^2 and velocity dispersions >10 km/s. Large-scale\nbar-driven dynamics likely regulate GMC properties in the nuclear region.\nSimilar to the Milky Way and other disk galaxies, GMC mass function of NGC 6946\nhas a shallower slope (index>-2) in the inner region, and a steeper slope\n(index<-2) in the outer region. This difference in mass spectra may be\nindicative of different cloud formation pathways: gravitational instabilities\nmight play a major role in the nuclear region, while cloud coalescence might be\ndominant in the outer disk. Finally, the NGC 6946 clouds are similar to those\nin M33 in terms of statistical properties, but they are generally less luminous\nand turbulent than the M51 clouds.\n",
"title": "Submillimeter Array CO(2-1) Imaging of the NGC 6946 Giant Molecular Clouds"
} | null | null | null | null | true | null | 20753 | null | Default | null | null |
null | {
"abstract": " We have obtained low-resolution optical (0.7-0.98 micron) and near-infrared\n(1.11-1.34 micron and 0.8-2.5 micron) spectra of twelve isolated planetary-mass\ncandidates (J = 18.2-19.9 mag) of the 3-Myr sigma Orionis star cluster with a\nview to determining the spectroscopic properties of very young, substellar\ndwarfs and assembling a complete cluster mass function. We have classified our\ntargets by visual comparison with high- and low-gravity standards and by\nmeasuring newly defined spectroscopic indices. We derived L0-L4.5 and M9-L2.5\nusing high- and low-gravity standards, respectively. Our targets reveal clear\nsignposts of youth, thus corroborating their cluster membership and planetary\nmasses (6-13 Mjup). These observations complete the sigma Orionis mass function\nby spectroscopically confirming the planetary-mass domain to a confidence level\nof $\\sim$75 percent. The comparison of our spectra with BT-Settl solar\nmetallicity model atmospheres yields a temperature scale of 2350-1800 K and a\nlow surface gravity of log g ~ 4.0 [cm/s2], as would be expected for young\nplanetary-mass objects. We discuss the properties of the cluster least-massive\npopulation as a function of spectral type. We have also obtained the first\noptical spectrum of S Ori 70, a T dwarf in the direction of sigma Orionis. Our\ndata provide reference optical and near-infrared spectra of very young L dwarfs\nand a mass function that may be used as templates for future studies of\nlow-mass substellar objects and exoplanets. The extrapolation of the sigma\nOrionis mass function to the solar neighborhood may indicate that isolated\nplanetary-mass objects with temperatures of 200-300 K and masses in the\ninterval 6-13-Mjup may be as numerous as very low-mass stars.\n",
"title": "Optical and Near-Infrared Spectra of sigma Orionis Isolated Planetary-mass Objects"
} | null | null | null | null | true | null | 20754 | null | Default | null | null |
null | {
"abstract": " The Blackbird unmanned aerial vehicle (UAV) dataset is a large-scale,\naggressive indoor flight dataset collected using a custom-built quadrotor\nplatform for use in evaluation of agile perception.Inspired by the potential of\nfuture high-speed fully-autonomous drone racing, the Blackbird dataset contains\nover 10 hours of flight data from 168 flights over 17 flight trajectories and 5\nenvironments at velocities up to $7.0ms^-1$. Each flight includes sensor data\nfrom 120Hz stereo and downward-facing photorealistic virtual cameras, 100Hz\nIMU, $\\sim190Hz$ motor speed sensors, and 360Hz millimeter-accurate motion\ncapture ground truth. Camera images for each flight were photorealistically\nrendered using FlightGoggles across a variety of environments to facilitate\neasy experimentation of high performance perception algorithms. The dataset is\navailable for download at this http URL\n",
"title": "The Blackbird Dataset: A large-scale dataset for UAV perception in aggressive flight"
} | null | null | null | null | true | null | 20755 | null | Default | null | null |
null | {
"abstract": " Motivation: Cellular Electron CryoTomography (CECT) is an emerging 3D imaging\ntechnique that visualizes subcellular organization of single cells at\nsubmolecular resolution and in near-native state. CECT captures large numbers\nof macromolecular complexes of highly diverse structures and abundances.\nHowever, the structural complexity and imaging limits complicate the systematic\nde novo structural recovery and recognition of these macromolecular complexes.\nEfficient and accurate reference-free subtomogram averaging and classification\nrepresent the most critical tasks for such analysis. Existing subtomogram\nalignment based methods are prone to the missing wedge effects and low\nsignal-to-noise ratio (SNR). Moreover, existing maximum-likelihood based\nmethods rely on integration operations, which are in principle computationally\ninfeasible for accurate calculation.\nResults: Built on existing works, we propose an integrated method, Fast\nAlignment Maximum Likelihood method (FAML), which uses fast subtomogram\nalignment to sample sub-optimal rigid transformations. The transformations are\nthen used to approximate integrals for maximum-likelihood update of subtomogram\naverages through expectation-maximization algorithm. Our tests on simulated and\nexperimental subtomograms showed that, compared to our previously developed\nfast alignment method (FA), FAML is significantly more robust to noise and\nmissing wedge effects with moderate increases of computation cost.Besides, FAML\nperforms well with significantly fewer input subtomograms when the FA method\nfails. Therefore, FAML can serve as a key component for improved construction\nof initial structural models from macromolecules captured by CECT.\n",
"title": "An integration of fast alignment and maximum-likelihood methods for electron subtomogram averaging and classification"
} | null | null | null | null | true | null | 20756 | null | Default | null | null |
null | {
"abstract": " This article focuses on a quasilinear wave equation of $p$-Laplacian type: $$\nu_{tt} - \\Delta_p u - \\Delta u_t=0$$ in a bounded domain\n$\\Omega\\subset\\mathbb{R}^3$ with a sufficiently smooth boundary\n$\\Gamma=\\partial\\Omega$ subject to a generalized Robin boundary condition\nfeaturing boundary damping and a nonlinear source term. The operator\n$\\Delta_p$, $2 < p < 3$, denotes the classical $p$-Laplacian. The nonlinear\nboundary term $f (u)$ is a source feedback that is allowed to have a\nsupercritical exponent, in the sense that the associated Nemytskii operator is\nnot locally Lipschitz from $W^{1,p}(\\Omega)$ into $L^2(\\Gamma)$. Under suitable\nassumptions on the parameters we provide a rigorous proof of existence of a\nlocal weak solution which can be extended globally in time provided the source\nterm satisfies an appropriate growth condition.\n",
"title": "Local and global existence of solutions to a strongly damped wave equation of the $p$-Laplacian type"
} | null | null | null | null | true | null | 20757 | null | Default | null | null |
null | {
"abstract": " Given a discrete group $\\Gamma=<g_1,\\ldots,g_M>$ and a number $K\\in\\mathbb\nN$, a unitary representation $\\rho:\\Gamma\\to U_K$ is called quasi-flat when the\neigenvalues of each $\\rho(g_i)\\in U_K$ are uniformly distributed among the\n$K$-th roots of unity. The quasi-flat representations of $\\Gamma$ form\naltogether a parametric matrix model $\\pi:\\Gamma\\to C(X,U_K)$.\nWe compute here the universal model space $X$ for various classes of discrete\ngroups, notably with results in the case where $\\Gamma$ is metabelian. We are\nparticularly interested in the case where $X$ is a union of compact homogeneous\nspaces, and where the induced representation $\\tilde{\\pi}:C^*(\\Gamma)\\to\nC(X,U_K)$ is stationary in the sense that it commutes with the Haar\nfunctionals. We present several positive and negative results on this subject.\nWe also discuss similar questions for the discrete quantum groups, proving a\nstationarity result for the discrete dual of the twisted orthogonal group\n$O_2^{-1}$.\n",
"title": "Quasi-flat representations of uniform groups and quantum groups"
} | null | null | null | null | true | null | 20758 | null | Default | null | null |
null | {
"abstract": " We consider the task of collaborative preference completion: given a pool of\nitems, a pool of users and a partially observed item-user rating matrix, the\ngoal is to recover the \\emph{personalized ranking} of each user over all of the\nitems. Our approach is nonparametric: we assume that each item $i$ and each\nuser $u$ have unobserved features $x_i$ and $y_u$, and that the associated\nrating is given by $g_u(f(x_i,y_u))$ where $f$ is Lipschitz and $g_u$ is a\nmonotonic transformation that depends on the user. We propose a $k$-nearest\nneighbors-like algorithm and prove that it is consistent. To the best of our\nknowledge, this is the first consistency result for the collaborative\npreference completion problem in a nonparametric setting. Finally, we\ndemonstrate the performance of our algorithm with experiments on the Netflix\nand Movielens datasets.\n",
"title": "Nonparametric Preference Completion"
} | null | null | null | null | true | null | 20759 | null | Default | null | null |
null | {
"abstract": " For the emerging Internet of Things (IoT), one of the most critical problems\nis the real-time reconstruction of signals from a set of aged measurements.\nDuring the reconstruction, distortion occurs between the observed signal and\nthe reconstructed signal due to sampling and transmission. In this paper, we\nfocus on minimizing the average distortion defined as the 1-norm of the\ndifference of the two signals under the scenario that a Poisson counting\nprocess is reconstructed in real-time on a remote monitor. Especially, we\nconsider the reconstruction under uniform sampling policy and two non-uniform\nsampling policies, i.e., the threshold-based policy and the zero-wait policy.\nFor each of the policy, we derive the closed-form expression of the average\ndistortion by dividing the overall distortion area into polygons and analyzing\ntheir structures. It turns out that the polygons are built up by sub-polygons\nthat account for distortions caused by sampling and transmission. The\nclosed-form expressions of the average distortion help us find the optimal\nsampling parameters that achieve the minimum distortion. Simulation results are\nprovided to validate our conclusion.\n",
"title": "Real-Time Reconstruction of Counting Process through Queues"
} | null | null | null | null | true | null | 20760 | null | Default | null | null |
null | {
"abstract": " We study the multi-armed bandit problem where the rewards are realizations of\ngeneral non-stationary stochastic processes, a setting that generalizes many\nexisting lines of work and analyses. In particular, we present a theoretical\nanalysis and derive regret guarantees for rested bandits in which the reward\ndistribution of each arm changes only when we pull that arm. Remarkably, our\nregret bounds are logarithmic in the number of rounds under several natural\nconditions. We introduce a new algorithm based on classical UCB ideas combined\nwith the notion of weighted discrepancy, a useful tool for measuring the\nnon-stationarity of a stochastic process. We show that the notion of\ndiscrepancy can be used to design very general algorithms and a unified\nframework for the analysis of multi-armed rested bandit problems with\nnon-stationary rewards. In particular, we show that we can recover the regret\nguarantees of many specific instances of bandit problems with non-stationary\nrewards that have been studied in the literature. We also provide experiments\ndemonstrating that our algorithms can enjoy a significant improvement in\npractice compared to standard benchmarks.\n",
"title": "Discrepancy-Based Algorithms for Non-Stationary Rested Bandits"
} | null | null | null | null | true | null | 20761 | null | Default | null | null |
null | {
"abstract": " The physical properties of an intermetallic compound CeRh2Ga2 have been\ninvestigated by magnetic susceptibility \\chi(T), isothermal magnetization M(H),\nheat capacity C_p(T), electrical resistivity \\rho(T), thermal conductivity\n\\kappa(T) and thermopower S(T) measurements. CeRh2Ga2 is found to crystallize\nwith CaBe2Ge2-type primitive tetragonal structure (space group P4/nmm). No\nevidence of long range magnetic order is seen down to 1.8 K. The \\chi(T) data\nshow paramagnetic behavior with an effective moment \\mu_eff ~ 2.5 \\mu_B/Ce\nindicating Ce^3+ valence state of Ce ions. The \\rho(T) data exhibit Kondo\nlattice behavior with a metallic ground state. The low-T C_p(T) data yield an\nenhanced Sommerfeld coefficient \\gamma = 130(2) mJ/mol K^2 characterizing\nCeRh2Ga2 as a moderate heavy fermion system. The high-T C_p(T) and \\rho(T) show\nan anomaly near 255 K, reflecting a phase transition. The \\kappa(T) suggests\nphonon dominated thermal transport with considerably higher values of Lorenz\nnumber L(T) compared to the theoretical Sommerfeld value L_0.\n",
"title": "Kondo lattice heavy fermion behavior in CeRh2Ga2"
} | null | null | null | null | true | null | 20762 | null | Default | null | null |
null | {
"abstract": " Based on the results of the second author, we define an equivariant version\nof Lee and Bar-Natan homology for periodic links and show that there exists an\nequivariant spectral sequence from the equivariant Khovanov homology to\nequivariant Lee homology. As a result we obtain new obstructions for a link to\nbe periodic. These obstructions generalize previous results of Przytycki and of\nthe second author.\n",
"title": "Khovanov homology and periodic links"
} | null | null | null | null | true | null | 20763 | null | Default | null | null |
null | {
"abstract": " Spectral clustering and Singular Value Decomposition (SVD) are both widely\nused technique for analyzing graph data. In this note, I will present their\nconnections using simple linear algebra, aiming to provide some in-depth\nunderstanding for future research.\n",
"title": "A Note on Spectral Clustering and SVD of Graph Data"
} | null | null | null | null | true | null | 20764 | null | Default | null | null |
null | {
"abstract": " Solving linear programs by using entropic penalization has recently attracted\nnew interest in the optimization community, since this strategy forms the basis\nfor the fastest-known algorithms for the optimal transport problem, with many\napplications in modern large-scale machine learning. Crucial to these\napplications has been an analysis of how quickly solutions to the penalized\nprogram approach true optima to the original linear program. More than 20 years\nago, Cominetti and San Martín showed that this convergence is exponentially\nfast; however, their proof is asymptotic and does not give any indication of\nhow accurately the entropic program approximates the original program for any\nparticular choice of the penalization parameter. We close this long-standing\ngap in the literature regarding entropic penalization by giving a new proof of\nthe exponential convergence, valid for any linear program. Our proof is\nnon-asymptotic, yields explicit constants, and has the virtue of being\nextremely simple. We provide matching lower bounds and show that the entropic\napproach does not lead to a near-linear time approximation scheme for the\nlinear assignment problem.\n",
"title": "An explicit analysis of the entropic penalty in linear programming"
} | null | null | null | null | true | null | 20765 | null | Default | null | null |
null | {
"abstract": " Network latencies have become increasingly important for the performance of\nweb servers and cloud computing platforms. Identifying network-related tail\nlatencies and reasoning about their potential causes is especially important to\ngauge application run-time in online data-intensive applications, where the\n99th percentile latency of individual operations can significantly affect the\nthe overall latency of requests.\nThis paper deconstructs the \"tail at scale\" effect across TCP-IP, UDP-IP, and\nRDMA network protocols. Prior scholarly works have analyzed tail latencies\ncaused by extrinsic network parameters like network congestion and flow\nfairness. Contrary to existing literature, we identify surprising rare tails in\nTCP-IP round-trip measurements that are as enormous as 110x higher than the\nmedian latency. Our experimental design eliminates network congestion as a\ntail-inducing factor. Moreover, we observe similar extreme tails in UDP-IP\npacket exchanges, ruling out additional TCP-IP protocol operations as the root\ncause of tail latency. However, we are unable to reproduce similar tail\nlatencies in RDMA packet exchanges, which leads us to conclude that the TCP/UDP\nprotocol stack within the operating system kernel is likely the primary source\nof extreme latency tails.\n",
"title": "Deconstructing the Tail at Scale Effect Across Network Protocols"
} | null | null | null | null | true | null | 20766 | null | Default | null | null |
null | {
"abstract": " Accurate diagnosis of psychiatric disorders plays a critical role in\nimproving quality of life for patients and potentially supports the development\nof new treatments. Many studies have been conducted on machine learning\ntechniques that seek brain imaging data for specific biomarkers of disorders.\nThese studies have encountered the following dilemma: An end-to-end\nclassification overfits to a small number of high-dimensional samples but\nunsupervised feature-extraction has the risk of extracting a signal of no\ninterest. In addition, such studies often provided only diagnoses for patients\nwithout presenting the reasons for these diagnoses. This study proposed a deep\nneural generative model of resting-state functional magnetic resonance imaging\n(fMRI) data. The proposed model is conditioned by the assumption of the\nsubject's state and estimates the posterior probability of the subject's state\ngiven the imaging data, using Bayes' rule. This study applied the proposed\nmodel to diagnose schizophrenia and bipolar disorders. Diagnosis accuracy was\nimproved by a large margin over competitive approaches, namely a support vector\nmachine, logistic regression, and multilayer perceptron with or without\nunsupervised feature-extractors in addition to a Gaussian mixture model. The\nproposed model visualizes brain regions largely related to the disorders, thus\nmotivating further biological investigation.\n",
"title": "Deep Neural Generative Model of Functional MRI Images for Psychiatric Disorder Diagnosis"
} | null | null | null | null | true | null | 20767 | null | Default | null | null |
null | {
"abstract": " An accurate model of patient-specific kidney graft survival distributions can\nhelp to improve shared-decision making in the treatment and care of patients.\nIn this paper, we propose a deep learning method that directly models the\nsurvival function instead of estimating the hazard function to predict survival\ntimes for graft patients based on the principle of multi-task learning. By\nlearning to jointly predict the time of the event, and its rank in the cox\npartial log likelihood framework, our deep learning approach outperforms, in\nterms of survival time prediction quality and concordance index, other common\nmethods for survival analysis, including the Cox Proportional Hazards model and\na network trained on the cox partial log-likelihood.\n",
"title": "Deep Learning for Patient-Specific Kidney Graft Survival Analysis"
} | null | null | null | null | true | null | 20768 | null | Default | null | null |
null | {
"abstract": " We introduce a framework for the modeling of sequential data capturing\npathways of varying lengths observed in a network. Such data are important,\ne.g., when studying click streams in information networks, travel patterns in\ntransportation systems, information cascades in social networks, biological\npathways or time-stamped social interactions. While it is common to apply graph\nanalytics and network analysis to such data, recent works have shown that\ntemporal correlations can invalidate the results of such methods. This raises a\nfundamental question: when is a network abstraction of sequential data\njustified? Addressing this open question, we propose a framework which combines\nMarkov chains of multiple, higher orders into a multi-layer graphical model\nthat captures temporal correlations in pathways at multiple length scales\nsimultaneously. We develop a model selection technique to infer the optimal\nnumber of layers of such a model and show that it outperforms previously used\nMarkov order detection techniques. An application to eight real-world data sets\non pathways and temporal networks shows that it allows to infer graphical\nmodels which capture both topological and temporal characteristics of such\ndata. Our work highlights fallacies of network abstractions and provides a\nprincipled answer to the open question when they are justified. Generalizing\nnetwork representations to multi-order graphical models, it opens perspectives\nfor new data mining and knowledge discovery algorithms.\n",
"title": "When is a Network a Network? Multi-Order Graphical Model Selection in Pathways and Temporal Networks"
} | null | null | null | null | true | null | 20769 | null | Default | null | null |
null | {
"abstract": " We develop efficient algorithms for estimating low-degree moments of unknown\ndistributions in the presence of adversarial outliers. The guarantees of our\nalgorithms improve in many cases significantly over the best previous ones,\nobtained in recent works of Diakonikolas et al, Lai et al, and Charikar et al.\nWe also show that the guarantees of our algorithms match information-theoretic\nlower-bounds for the class of distributions we consider. These improved\nguarantees allow us to give improved algorithms for independent component\nanalysis and learning mixtures of Gaussians in the presence of outliers.\nOur algorithms are based on a standard sum-of-squares relaxation of the\nfollowing conceptually-simple optimization problem: Among all distributions\nwhose moments are bounded in the same way as for the unknown distribution, find\nthe one that is closest in statistical distance to the empirical distribution\nof the adversarially-corrupted sample.\n",
"title": "Outlier-robust moment-estimation via sum-of-squares"
} | null | null | null | null | true | null | 20770 | null | Default | null | null |
null | {
"abstract": " We present data streaming algorithms for the $k$-median problem in\nhigh-dimensional dynamic geometric data streams, i.e. streams allowing both\ninsertions and deletions of points from a discrete Euclidean space $\\{1, 2,\n\\ldots \\Delta\\}^d$. Our algorithms use $k \\epsilon^{-2} poly(d \\log \\Delta)$\nspace/time and maintain with high probability a small weighted set of points (a\ncoreset) such that for every set of $k$ centers the cost of the coreset\n$(1+\\epsilon)$-approximates the cost of the streamed point set. We also provide\nalgorithms that guarantee only positive weights in the coreset with additional\nlogarithmic factors in the space and time complexities. We can use this\npositively-weighted coreset to compute a $(1+\\epsilon)$-approximation for the\n$k$-median problem by any efficient offline $k$-median algorithm. All previous\nalgorithms for computing a $(1+\\epsilon)$-approximation for the $k$-median\nproblem over dynamic data streams required space and time exponential in $d$.\nOur algorithms can be generalized to metric spaces of bounded doubling\ndimension.\n",
"title": "Clustering High Dimensional Dynamic Data Streams"
} | null | null | null | null | true | null | 20771 | null | Default | null | null |
null | {
"abstract": " We consider capillary condensation transitions occurring in open slits of\nwidth $L$ and finite height $H$ immersed in a reservoir of vapour. In this case\nthe pressure at which condensation occurs is closer to saturation compared to\nthat occurring in an infinite slit ($H=\\infty$) due to the presence of two\nmenisci which are pinned near the open ends. Using macroscopic arguments we\nderive a modified Kelvin equation for the pressure, $p_{cc}(L;H)$, at which\ncondensation occurs and show that the two menisci are characterised by an edge\ncontact angle $\\theta_e$ which is always larger than the equilibrium contact\nangle $\\theta$, only equal to it in the limit of macroscopic $H$. For walls\nwhich are completely wet ($\\theta=0$) the edge contact angle depends only on\nthe aspect ratio of the capillary and is well described by $\\theta_e\\approx\n\\sqrt{\\pi L/2H}$ for large $H$. Similar results apply for condensation in\ncylindrical pores of finite length. We have tested these predictions against\nnumerical results obtained using a microscopic density functional model where\nthe presence of an edge contact angle characterising the shape of the menisci\nis clearly visible from the density profiles. Below the wetting temperature\n$T_w$ we find very good agreement for slit pores of widths of just a few tens\nof molecular diameters while above $T_w$ the modified Kelvin equation only\nbecomes accurate for much larger systems.\n",
"title": "Edge contact angle and modified Kelvin equation for condensation in open pores"
} | null | null | null | null | true | null | 20772 | null | Default | null | null |
null | {
"abstract": " The Melan equation for suspension bridges is derived by assuming small\ndisplacements of the deck and inextensible hangers. We determine the thresholds\nfor the validity of the Melan equation when the hangers slacken, thereby\nviolating the inextensibility assumption. To this end, we preliminarily study\nthe possible shortening of the cables: it turns out that there is a striking\ndifference between even and odd vibrating modes since the former never shorten\nthe cables. These problems are studied both on beams and plates.\n",
"title": "Thresholds for hanger slackening and cable shortening in the Melan equation for suspension bridges"
} | null | null | null | null | true | null | 20773 | null | Default | null | null |
null | {
"abstract": " Optimal path planning problems for rigid and deformable (bendable) cuboid\nrobots are considered by providing an analytic safety constraint using\ngeneralized $L_p$ norms. For regular cuboid robots, level sets of weighted\n$L_p$ norms generate implicit approximations of their surfaces. For bendable\ncuboid robots a weighted $L_p$ norm in polar coordinates implicitly\napproximates the surface boundary through a specified level set. Obstacle\nvolumes, in the environment to navigate within, are presumed to be\napproximately described as sub-level sets of weighted $L_p$ norms. Using these\napproximate surface models, the optimal safe path planning problem is\nreformulated as a two stage optimization problem, where the safety constraint\ndepends on a point on the robot which is closest to the obstacle in the\nobstacle's distance metric. A set of equality and inequality constraints are\nderived to replace the closest point problem, which is then defines additional\nanalytic constraints on the original path planning problem. Combining all the\nanalytic constraints with logical AND operations leads to a general optimal\nsafe path planning problem. Numerically solving the problem involve conversion\nto a nonlinear programing problem. Simulations for rigid and bendable cuboid\nrobot verify the proposed method.\n",
"title": "Bendable Cuboid Robot Path Planning with Collision Avoidance using Generalized $L_p$ Norms"
} | null | null | null | null | true | null | 20774 | null | Default | null | null |
null | {
"abstract": " We give a sufficient condition for a Verdier quotient $\\ct/\\cs$ of a\ntriangulated category $\\ct$ by a thick subcategory $\\cs$ to be realized inside\nof $\\ct$ as an ideal quotient. As applications, we deduce three significant\nresults by Buchweitz, Orlov and Amiot--Guo--Keller.\n",
"title": "Quotients of triangulated categories and Equivalences of Buchweitz, Orlov and Amiot--Guo--Keller"
} | null | null | null | null | true | null | 20775 | null | Default | null | null |
null | {
"abstract": " A sum where each of the $N$ summands can be independently chosen from two\nchoices yields $2^N$ possible summation outcomes. There is an\n$\\mathcal{O}(K^2)$-algorithm that finds the $K$ smallest/largest of these sums\nby evading the enumeration of all sums.\n",
"title": "Sorting sums of binary decision summands"
} | null | null | null | null | true | null | 20776 | null | Default | null | null |
null | {
"abstract": " Detecting activities in untrimmed videos is an important but challenging\ntask. The performance of existing methods remains unsatisfactory, e.g., they\noften meet difficulties in locating the beginning and end of a long complex\naction. In this paper, we propose a generic framework that can accurately\ndetect a wide variety of activities from untrimmed videos. Our first\ncontribution is a novel proposal scheme that can efficiently generate\ncandidates with accurate temporal boundaries. The other contribution is a\ncascaded classification pipeline that explicitly distinguishes between\nrelevance and completeness of a candidate instance. On two challenging temporal\nactivity detection datasets, THUMOS14 and ActivityNet, the proposed framework\nsignificantly outperforms the existing state-of-the-art methods, demonstrating\nsuperior accuracy and strong adaptivity in handling activities with various\ntemporal structures.\n",
"title": "A Pursuit of Temporal Accuracy in General Activity Detection"
} | null | null | null | null | true | null | 20777 | null | Default | null | null |
null | {
"abstract": " Principal component analysis continues to be a powerful tool in dimension\nreduction of high dimensional data. We assume a variance-diverging model and\nuse the high-dimension, low-sample-size asymptotics to show that even though\nthe principal component directions are not consistent, the sample and\nprediction principal component scores can be useful in revealing the population\nstructure. We further show that these scores are biased, and the bias is\nasymptotically decomposed into rotation and scaling parts. We propose methods\nof bias-adjustment that are shown to be consistent and work well in the finite\nbut high dimensional situations with small sample sizes. The potential\nadvantage of bias-adjustment is demonstrated in a classification setting.\n",
"title": "Adjusting systematic bias in high dimensional principal component scores"
} | null | null | null | null | true | null | 20778 | null | Default | null | null |
null | {
"abstract": " Nonlinear systems, whose outputs are not directly proportional to their\ninputs, are well known to exhibit many interesting and important phenomena\nwhich have profoundly changed our technological landscape over the last 50\nyears. Recently the ability to engineer quantum metamaterials through\nhybridisation has allowed to explore these nonlinear effects in systems with no\nnatural analogue. Here we investigate amplitude bistability, which is one of\nthe most fundamental nonlinear phenomena, in a hybrid system composed of a\nsuperconducting resonator inductively coupled to an ensemble of\nnitrogen-vacancy centres. One of the exciting properties of this spin system is\nits extremely long spin life-time, more than ten orders of magnitude longer\nthan other relevant timescales of the hybrid system. This allows us to\ndynamically explore this nonlinear regime of cavity quantum electrodynamics\n(cQED) and demonstrate a critical slowing down of the cavity population on the\norder of several tens of thousands of seconds - a timescale much longer than\nobserved so far for this effect. Our results provide the foundation for future\nquantum technologies based on nonlinear phenomena.\n",
"title": "Dynamical Exploration of Amplitude Bistability in Engineered Quantum Systems"
} | null | null | null | null | true | null | 20779 | null | Default | null | null |
null | {
"abstract": " In this work, we present a novel strategy for correcting imperfections in\noccupancy grid maps called map decay. The objective of map decay is to correct\ninvalid occupancy probabilities of map cells that are unobservable by sensors.\nThe strategy was inspired by an analogy between the memory architecture\nbelieved to exist in the human brain and the maps maintained by an autonomous\nvehicle. It consists in merging sensory information obtained during runtime\n(online) with a priori data from a high-precision map constructed offline. In\nmap decay, cells observed by sensors are updated using traditional occupancy\ngrid mapping techniques and unobserved cells are adjusted so that their\noccupancy probabilities tend to the values found in the offline map. This\nstrategy is grounded in the idea that the most precise information available\nabout an unobservable cell is the value found in the high-precision offline\nmap. Map decay was successfully tested and is still in use in the IARA\nautonomous vehicle from Universidade Federal do Espírito Santo.\n",
"title": "Map Memorization and Forgetting in the IARA Autonomous Car"
} | null | null | null | null | true | null | 20780 | null | Default | null | null |
null | {
"abstract": " This paper studies effective separability for subgroups of finitely generated\nnilpotent groups and more broadly effective subgroup separability of finitely\ngenerated nilpotent groups. We provide upper and lower bounds that are\npolynomial with respect to the logarithm of the word length for infinite index\nsubgroups of nilpotent groups. In the case of normal subgroups, we provide an\nexact computation generalizing work of the second author. We introduce a\nfunction that quantifies subgroup separability, and we provide polynomial upper\nand lower bounds. We finish by demonstrating that our results extend to\nvirtually nilpotent groups.\n",
"title": "Effective Subgroup Separability of Finitely Generated Nilpotent Groups"
} | null | null | null | null | true | null | 20781 | null | Default | null | null |
null | {
"abstract": " Most of existing image denoising methods learn image priors from either\nexternal data or the noisy image itself to remove noise. However, priors\nlearned from external data may not be adaptive to the image to be denoised,\nwhile priors learned from the given noisy image may not be accurate due to the\ninterference of corrupted noise. Meanwhile, the noise in real-world noisy\nimages is very complex, which is hard to be described by simple distributions\nsuch as Gaussian distribution, making real-world noisy image denoising a very\nchallenging problem. We propose to exploit the information in both external\ndata and the given noisy image, and develop an external prior guided internal\nprior learning method for real-world noisy image denoising. We first learn\nexternal priors from an independent set of clean natural images. With the aid\nof learned external priors, we then learn internal priors from the given noisy\nimage to refine the prior model. The external and internal priors are\nformulated as a set of orthogonal dictionaries to efficiently reconstruct the\ndesired image. Extensive experiments are performed on several real-world noisy\nimage datasets. The proposed method demonstrates highly competitive denoising\nperformance, outperforming state-of-the-art denoising methods including those\ndesigned for real-world noisy images.\n",
"title": "External Prior Guided Internal Prior Learning for Real-World Noisy Image Denoising"
} | null | null | [
"Computer Science"
]
| null | true | null | 20782 | null | Validated | null | null |
null | {
"abstract": " MAC address randomization is a privacy technique whereby mobile devices\nrotate through random hardware addresses in order to prevent observers from\nsingling out their traffic or physical location from other nearby devices.\nAdoption of this technology, however, has been sporadic and varied across\ndevice manufacturers. In this paper, we present the first wide-scale study of\nMAC address randomization in the wild, including a detailed breakdown of\ndifferent randomization techniques by operating system, manufacturer, and model\nof device.\nWe then identify multiple flaws in these implementations which can be\nexploited to defeat randomization as performed by existing devices. First, we\nshow that devices commonly make improper use of randomization by sending\nwireless frames with the true, global address when they should be using a\nrandomized address. We move on to extend the passive identification techniques\nof Vanhoef et al. to effectively defeat randomization in ~96% of Android\nphones. Finally, we show a method that can be used to track 100% of devices\nusing randomization, regardless of manufacturer, by exploiting a previously\nunknown flaw in the way existing wireless chipsets handle low-level control\nframes.\n",
"title": "A Study of MAC Address Randomization in Mobile Devices and When it Fails"
} | null | null | null | null | true | null | 20783 | null | Default | null | null |
null | {
"abstract": " In this paper, we proposed an procedure to construct the completion of the\nintegrable system by adding a perturbation to the generalized matrix problem,\nwhich can be used to continuous integrable couplings, discrete integrable\ncouplings and super integrable couplings. As example, we construct the\ncompletion of the Kaup-Newell (KN) integrable coupling, the\nWadati-Konno-Ichikawa (WKI) integrable couplingsis, vector\nAblowitz-Kaup-Newell-Segur (vAKNS) integrable couplings, the Volterra\nintegrable couplings, Dirac type integrable couplings and NLS-mKdV type\nintegrable couplings.\n",
"title": "Completion of the integrable coupling systems"
} | null | null | null | null | true | null | 20784 | null | Default | null | null |
null | {
"abstract": " In this note, we investigate the representation type of the cambrian lattices\nand some other related lattices. The result is expressed as a very simple\ntrichotomy. When the rank of the underlined Coxeter group is at most 2, the\nlattices are of finite representation type. When the Coxeter group is a\nreducible group of type A 3 1 , the lattices are of tame representation type.\nIn all the other cases they are of wild representation type.\n",
"title": "On the wildness of cambrian lattices"
} | null | null | [
"Mathematics"
]
| null | true | null | 20785 | null | Validated | null | null |
null | {
"abstract": " Fitting linear regression models can be computationally very expensive in\nlarge-scale data analysis tasks if the sample size and the number of variables\nare very large. Random projections are extensively used as a dimension\nreduction tool in machine learning and statistics. We discuss the applications\nof random projections in linear regression problems, developed to decrease\ncomputational costs, and give an overview of the theoretical guarantees of the\ngeneralization error. It can be shown that the combination of random\nprojections with least squares regression leads to similar recovery as ridge\nregression and principal component regression. We also discuss possible\nimprovements when averaging over multiple random projections, an approach that\nlends itself easily to parallel implementation.\n",
"title": "Random Projections For Large-Scale Regression"
} | null | null | null | null | true | null | 20786 | null | Default | null | null |
null | {
"abstract": " We consider linear groups which do not contain unipotent elements of infinite\norder, which includes all linear groups in positive characteristic, and show\nthat this class of groups has good properties which resemble those held by\ngroups of non positive curvature and which do not hold for arbitrary\ncharacteristic zero linear groups. In particular if such a linear group is\nfinitely generated then centralisers virtually split and all finitely generated\nabelian subgroups are undistorted. If further the group is virtually torsion\nfree (which always holds in characteristic zero) then we have a strong property\non small subgroups: any subgroup either contains a non abelian free group or is\nfinitely generated and virtually abelian, hence also undistorted. We present\napplications, including that the mapping class group of a surface having genus\nat least 3 has no faithful linear representation which is complex unitary or\nover any field of positive characteristic.\n",
"title": "Properties of linear groups with restricted unipotent elements"
} | null | null | null | null | true | null | 20787 | null | Default | null | null |
null | {
"abstract": " The discrete Laplace operator is ubiquitous in spectral shape analysis, since\nits eigenfunctions are provably optimal in representing smooth functions\ndefined on the surface of the shape. Indeed, subspaces defined by its\neigenfunctions have been utilized for shape compression, treating the\ncoordinates as smooth functions defined on the given surface. However, surfaces\nof shapes in nature often contain geometric structures for which the general\nsmoothness assumption may fail to hold. At the other end, some explicit mesh\ncompression algorithms utilize the order by which vertices that represent the\nsurface are traversed, a property which has been ignored in spectral\napproaches. Here, we incorporate the order of vertices into an operator that\ndefines a novel spectral domain. We propose a method for representing 3D meshes\nusing the spectral geometry of the Hamiltonian operator, integrated within a\nsparse approximation framework. We adapt the concept of a potential function\nfrom quantum physics and incorporate vertex ordering information into the\npotential, yielding a novel data-dependent operator. The potential function\nmodifies the spectral geometry of the Laplacian to focus on regions with finer\ndetails of the given surface. By sparsely encoding the geometry of the shape\nusing the proposed data-dependent basis, we improve compression performance\ncompared to previous results that use the standard Laplacian basis and spectral\ngraph wavelets.\n",
"title": "Sparse Approximation of 3D Meshes using the Spectral Geometry of the Hamiltonian Operator"
} | null | null | null | null | true | null | 20788 | null | Default | null | null |
null | {
"abstract": " In this paper, we develop a class of decentralized algorithms for solving a\nconvex resource allocation problem in a network of $n$ agents, where the agent\nobjectives are decoupled while the resource constraints are coupled. The agents\ncommunicate over a connected undirected graph, and they want to collaboratively\ndetermine a solution to the overall network problem, while each agent only\ncommunicates with its neighbors. We first study the connection between the\ndecentralized resource allocation problem and the decentralized consensus\noptimization problem. Then, using a class of algorithms for solving consensus\noptimization problems, we propose a novel class of decentralized schemes for\nsolving resource allocation problems in a distributed manner. Specifically, we\nfirst propose an algorithm for solving the resource allocation problem with an\n$o(1/k)$ convergence rate guarantee when the agents' objective functions are\ngenerally convex (could be nondifferentiable) and per agent local convex\nconstraints are allowed; We then propose a gradient-based algorithm for solving\nthe resource allocation problem when per agent local constraints are absent and\nshow that such scheme can achieve geometric rate when the objective functions\nare strongly convex and have Lipschitz continuous gradients. We have also\nprovided scalability/network dependency analysis. Based on these two\nalgorithms, we have further proposed a gradient projection-based algorithm\nwhich can handle smooth objective and simple constraints more efficiently.\nNumerical experiments demonstrates the viability and performance of all the\nproposed algorithms.\n",
"title": "Improved Convergence Rates for Distributed Resource Allocation"
} | null | null | null | null | true | null | 20789 | null | Default | null | null |
null | {
"abstract": " Today digital sources supply an unprecedented component of human sensorimotor\ndata, the consumption of which is correlated with poorly understood maladies\nsuch as Internet Addiction Disorder and Internet Gaming Disorder. This paper\noffers a mathematical understanding of human sensorimotor processing as\nmultiscale, continuous-time vibratory interaction. We quantify human\ninformational needs using the signal processing metrics of entropy, noise,\ndimensionality, continuity, latency, and bandwidth. Using these metrics, we\ndefine the trust humans experience as a primitive statistical algorithm\nprocessing finely grained sensorimotor data from neuromechanical interaction.\nThis definition of neuromechanical trust implies that artificial sensorimotor\ninputs and interactions that attract low-level attention through frequent\ndiscontinuities and enhanced coherence will decalibrate a brain's\nrepresentation of its world over the long term by violating the implicit\nstatistical contract for which self-calibration evolved. This approach allows\nus to model addiction in general as the result of homeostatic regulation gone\nawry in novel environments and digital dependency as a sub-case in which the\ndecalibration caused by digital sensorimotor data spurs yet more consumption of\nthem. We predict that institutions can use these sensorimotor metrics to\nquantify media richness to improve employee well-being; that dyads and\nfamily-size groups will bond and heal best through low-latency, high-resolution\nmultisensory interaction such as shared meals and reciprocated touch; and that\nindividuals can improve sensory and sociosensory resolution through deliberate\nsensory reintegration practices. We conclude that we humans are the victims of\nour own success, our hands so skilled they fill the world with captivating\nthings, our eyes so innocent they follow eagerly.\n",
"title": "Sensory Metrics of Neuromechanical Trust"
} | null | null | null | null | true | null | 20790 | null | Default | null | null |
null | {
"abstract": " Gaussian processes (GPs) are important models in supervised machine learning.\nTraining in Gaussian processes refers to selecting the covariance functions and\nthe associated parameters in order to improve the outcome of predictions, the\ncore of which amounts to evaluating the logarithm of the marginal likelihood\n(LML) of a given model. LML gives a concrete measure of the quality of\nprediction that a GP model is expected to achieve. The classical computation of\nLML typically carries a polynomial time overhead with respect to the input\nsize. We propose a quantum algorithm that computes the logarithm of the\ndeterminant of a Hermitian matrix, which runs in logarithmic time for sparse\nmatrices. This is applied in conjunction with a variant of the quantum linear\nsystem algorithm that allows for logarithmic time computation of the form\n$\\mathbf{y}^TA^{-1}\\mathbf{y}$, where $\\mathbf{y}$ is a dense vector and $A$ is\nthe covariance matrix. We hence show that quantum computing can be used to\nestimate the LML of a GP with exponentially improved efficiency under certain\nconditions.\n",
"title": "Quantum algorithms for training Gaussian Processes"
} | null | null | [
"Statistics"
]
| null | true | null | 20791 | null | Validated | null | null |
null | {
"abstract": " Most state-of-the-art text detection methods are specific to horizontal Latin\ntext and are not fast enough for real-time applications. We introduce Segment\nLinking (SegLink), an oriented text detection method. The main idea is to\ndecompose text into two locally detectable elements, namely segments and links.\nA segment is an oriented box covering a part of a word or text line; A link\nconnects two adjacent segments, indicating that they belong to the same word or\ntext line. Both elements are detected densely at multiple scales by an\nend-to-end trained, fully-convolutional neural network. Final detections are\nproduced by combining segments connected by links. Compared with previous\nmethods, SegLink improves along the dimensions of accuracy, speed, and ease of\ntraining. It achieves an f-measure of 75.0% on the standard ICDAR 2015\nIncidental (Challenge 4) benchmark, outperforming the previous best by a large\nmargin. It runs at over 20 FPS on 512x512 images. Moreover, without\nmodification, SegLink is able to detect long lines of non-Latin text, such as\nChinese.\n",
"title": "Detecting Oriented Text in Natural Images by Linking Segments"
} | null | null | [
"Computer Science"
]
| null | true | null | 20792 | null | Validated | null | null |
null | {
"abstract": " The aim of this paper is to introduce and study a large class of\n$\\mathfrak{g}$-module algebras which we call factorizable by generalizing the\nGauss factorization of (square or rectangular) matrices. This class includes\ncoordinate algebras of corresponding reductive groups $G$, their parabolic\nsubgroups, basic affine spaces and many others. It turns out that tensor\nproducts of factorizable algebras are also factorizable and it is easy to\ncreate a factorizable algebra out of virtually any $\\mathfrak{g}$-module\nalgebra. We also have quantum versions of all these constructions in the\ncategory of $U_q(\\mathfrak{g})$-module algebras. Quite surprisingly, our\nquantum factorizable algebras are naturally acted on by the quantized\nenveloping algebra $U_q(\\mathfrak{g}^*)$ of the dual Lie bialgebra\n$\\mathfrak{g}^*$ of $\\mathfrak{g}$.\n",
"title": "Factorizable Module Algebras"
} | null | null | null | null | true | null | 20793 | null | Default | null | null |
null | {
"abstract": " Robot awareness of human actions is an essential research problem in robotics\nwith many important real-world applications, including human-robot\ncollaboration and teaming. Over the past few years, depth sensors have become a\nstandard device widely used by intelligent robots for 3D perception, which can\nalso offer human skeletal data in 3D space. Several methods based on skeletal\ndata were designed to enable robot awareness of human actions with satisfactory\naccuracy. However, previous methods treated all body parts and features equally\nimportant, without the capability to identify discriminative body parts and\nfeatures. In this paper, we propose a novel simultaneous Feature And Body-part\nLearning (FABL) approach that simultaneously identifies discriminative body\nparts and features, and efficiently integrates all available information\ntogether to enable real-time robot awareness of human behaviors. We formulate\nFABL as a regression-like optimization problem with structured\nsparsity-inducing norms to model interrelationships of body parts and features.\nWe also develop an optimization algorithm to solve the formulated problem,\nwhich possesses a theoretical guarantee to find the optimal solution. To\nevaluate FABL, three experiments were performed using public benchmark\ndatasets, including the MSR Action3D and CAD-60 datasets, as well as a Baxter\nrobot in practical assistive living applications. Experimental results show\nthat our FABL approach obtains a high recognition accuracy with a processing\nspeed of the order-of-magnitude of 10e4 Hz, which makes FABL a promising method\nto enable real-time robot awareness of human behaviors in practical robotics\napplications.\n",
"title": "Simultaneous Feature and Body-Part Learning for Real-Time Robot Awareness of Human Behaviors"
} | null | null | null | null | true | null | 20794 | null | Default | null | null |
null | {
"abstract": " We obtain some necessary and sufficient conditions for the boundedness of a\nfamily of positive operators defined on symmetric cones, we then deduce\noff-diagonal boundedness of associated Bergman-type operators in tube domains\nover symmetric cones.\n",
"title": "Off-diagonal estimates of some Bergman-type operators on tube domains over symmetric cones"
} | null | null | null | null | true | null | 20795 | null | Default | null | null |
null | {
"abstract": " Many methods for automated software test generation, including some that\nexplicitly use machine learning (and some that use ML more broadly conceived)\nderive new tests from existing tests (often referred to as seeds). Often, the\nseed tests from which new tests are derived are manually constructed, or at\nleast simpler than the tests that are produced as the final outputs of such\ntest generators. We propose annotation of generated tests with a provenance\n(trail) showing how individual generated tests of interest (especially failing\ntests) derive from seed tests, and how the population of generated tests\nrelates to the original seed tests. In some cases, post-processing of generated\ntests can invalidate provenance information, in which case we also propose a\nmethod for attempting to construct \"pseudo-provenance\" describing how the tests\ncould have been (partly) generated from seeds.\n",
"title": "Provenance and Pseudo-Provenance for Seeded Learning-Based Automated Test Generation"
} | null | null | null | null | true | null | 20796 | null | Default | null | null |
null | {
"abstract": " It is expected that progress toward true artificial intelligence will be\nachieved through the emergence of a system that integrates representation\nlearning and complex reasoning (LeCun et al. 2015). In response to this\nprediction, research has been conducted on implementing the symbolic reasoning\nof a von Neumann computer in an artificial neural network (Graves et al. 2016;\nGraves et al. 2014; Reed et al. 2015). However, these studies have many\nlimitations in realizing neural-symbolic integration (Jaeger. 2016). Here, we\npresent a new learning paradigm: a learning solving procedure (LSP) that learns\nthe procedure for solving complex problems. This is not accomplished merely by\nlearning input-output data, but by learning algorithms through a solving\nprocedure that obtains the output as a sequence of tasks for a given input\nproblem. The LSP neural network system not only learns simple problems of\naddition and multiplication, but also the algorithms of complicated problems,\nsuch as complex arithmetic expression, sorting, and Hanoi Tower. To realize\nthis, the LSP neural network structure consists of a deep neural network and\nlong short-term memory, which are recursively combined. Through\nexperimentation, we demonstrate the efficiency and scalability of LSP and its\nvalidity as a mechanism of complex reasoning.\n",
"title": "Learning Solving Procedure for Artificial Neural Network"
} | null | null | null | null | true | null | 20797 | null | Default | null | null |
null | {
"abstract": " Recently, it has been claimed that inflationary models with an inflection\npoint in the scalar potential can produce a large resonance in the power\nspectrum of curvature perturbation. In this paper however we show that the\nprevious analyses are incorrect. The reason is twofold: firstly, the inflaton\nis over-shot from a stage of standard inflation and so deviates from the\nslow-roll attractor before reaching the inflection. Secondly, on the (or close\nto) the inflection point, the ultra-slow-roll trajectory supersede the\nslow-roll one and thus, the slow-roll approximations used in the literature\ncannot be used. We then reconsider the model and provide a recipe for how to\nproduce nevertheless a large peak in the matter power spectrum via fine-tuning\nof parameters.\n",
"title": "On primordial black holes from an inflection point"
} | null | null | null | null | true | null | 20798 | null | Default | null | null |
null | {
"abstract": " Regularization for matrix factorization (MF) and approximation problems has\nbeen carried out in many different ways. Due to its popularity in deep\nlearning, dropout has been applied also for this class of problems. Despite its\nsolid empirical performance, the theoretical properties of dropout as a\nregularizer remain quite elusive for this class of problems. In this paper, we\npresent a theoretical analysis of dropout for MF, where Bernoulli random\nvariables are used to drop columns of the factors. We demonstrate the\nequivalence between dropout and a fully deterministic model for MF in which the\nfactors are regularized by the sum of the product of squared Euclidean norms of\nthe columns. Additionally, we inspect the case of a variable sized\nfactorization and we prove that dropout achieves the global minimum of a convex\napproximation problem with (squared) nuclear norm regularization. As a result,\nwe conclude that dropout can be used as a low-rank regularizer with data\ndependent singular-value thresholding.\n",
"title": "Dropout as a Low-Rank Regularizer for Matrix Factorization"
} | null | null | null | null | true | null | 20799 | null | Default | null | null |
null | {
"abstract": " A temporal graph is a data structure, consisting of nodes and edges in which\nthe edges are associated with time labels. To analyze the temporal graph, the\nfirst step is to find a proper graph dataset/benchmark. While many temporal\ngraph datasets exist online, none could be found that used the interval labels\nin which each edge is associated with a starting and ending time. Therefore we\ncreate a temporal graph data based on Wikipedia reference graph for temporal\nanalysis. This report aims to provide more details of this graph benchmark to\nthose who are interested in using it.\n",
"title": "Introduction to a Temporal Graph Benchmark"
} | null | null | [
"Computer Science",
"Physics"
]
| null | true | null | 20800 | null | Validated | null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.