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": " Traditional recommendation systems rely on past usage data in order to\ngenerate new recommendations. Those approaches fail to generate sensible\nrecommendations for new users and items into the system due to missing\ninformation about their past interactions. In this paper, we propose a solution\nfor successfully addressing item-cold start problem which uses model-based\napproach and recent advances in deep learning. In particular, we use latent\nfactor model for recommendation, and predict the latent factors from item's\ndescriptions using convolutional neural network when they cannot be obtained\nfrom usage data. Latent factors obtained by applying matrix factorization to\nthe available usage data are used as ground truth to train the convolutional\nneural network. To create latent factor representations for the new items, the\nconvolutional neural network uses their textual description. The results from\nthe experiments reveal that the proposed approach significantly outperforms\nseveral baseline estimators.\n",
"title": "Addressing Item-Cold Start Problem in Recommendation Systems using Model Based Approach and Deep Learning"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
15501
| null |
Validated
| null | null |
null |
{
"abstract": " Recently, end-to-end models have become a popular approach as an alternative\nto traditional hybrid models in automatic speech recognition (ASR). The\nmulti-speaker speech separation and recognition task is a central task in\ncocktail party problem. In this paper, we present a state-of-the-art monaural\nmulti-speaker end-to-end automatic speech recognition model. In contrast to\nprevious studies on the monaural multi-speaker speech recognition, this\nend-to-end framework is trained to recognize multiple label sequences\ncompletely from scratch. The system only requires the speech mixture and\ncorresponding label sequences, without needing any indeterminate supervisions\nobtained from non-mixture speech or corresponding labels/alignments. Moreover,\nwe exploited using the individual attention module for each separated speaker\nand the scheduled sampling to further improve the performance. Finally, we\nevaluate the proposed model on the 2-speaker mixed speech generated from the\nWSJ corpus and the wsj0-2mix dataset, which is a speech separation and\nrecognition benchmark. The experiments demonstrate that the proposed methods\ncan improve the performance of the end-to-end model in separating the\noverlapping speech and recognizing the separated streams. From the results, the\nproposed model leads to ~10.0% relative performance gains in terms of CER and\nWER respectively.\n",
"title": "End-to-End Monaural Multi-speaker ASR System without Pretraining"
}
| null | null | null | null | true | null |
15502
| null |
Default
| null | null |
null |
{
"abstract": " Sheng and Zuo's characteristic forms are invariants of a variation of Hodge\nstructure. We show that they characterize Gross's canonical variations of Hodge\nstructure of Calabi-Yau type over (Hermitian symmetric) tube domains.\n",
"title": "Characterization of Calabi--Yau variations of Hodge structure over tube domains by characteristic forms"
}
| null | null |
[
"Mathematics"
] | null | true | null |
15503
| null |
Validated
| null | null |
null |
{
"abstract": " We investigate quantum graphs with infinitely many vertices and edges without\nthe common restriction on the geometry of the underlying metric graph that\nthere is a positive lower bound on the lengths of its edges. Our central result\nis a close connection between spectral properties of a quantum graph and the\ncorresponding properties of a certain weighted discrete Laplacian on the\nunderlying discrete graph. Using this connection together with spectral theory\nof (unbounded) discrete Laplacians on infinite graphs, we prove a number of new\nresults on spectral properties of quantum graphs. Namely, we prove several\nself-adjointness results including a Gaffney type theorem. We investigate the\nproblem of lower semiboundedness, prove several spectral estimates (bounds for\nthe bottom of spectra and essential spectra of quantum graphs, CLR-type\nestimates) and study spectral types.\n",
"title": "Spectral Theory of Infinite Quantum Graphs"
}
| null | null | null | null | true | null |
15504
| null |
Default
| null | null |
null |
{
"abstract": " Recently, Ciufolini et al. reported on a test of the general relativistic\ngravitomagnetic Lense-Thirring effect by analyzing about 3.5 years of laser\nranging data to the LAGEOS, LAGEOS II, LARES geodetic satellites orbiting the\nEarth. By using the GRACE-based GGM05S Earth's global gravity model and a\nlinear combination of the nodes $\\Omega$ of the three satellites designed to\nremove the impact of errors in the first two even zonal harmonic coefficients\n$J_2,~J_4$ of the multipolar expansion of the Newtonian part of the Earth's\ngravitational potential, they claimed an overall accuracy of $5\\%$ for the\nLense-Thirring caused node motion. We show that the scatter in the nominal\nvalues of the uncancelled even zonals of degree $\\ell = 6,~8,~10$ from some of\nthe most recent global gravity models does not yet allow to reach unambiguously\nand univocally the expected $\\approx 1\\%$ level, being large up to $\\lesssim\n15\\%~(\\ell=6),~6\\%~(\\ell=8),~36\\%~(\\ell=10)$ for some pairs of models.\n",
"title": "A comment on \"A test of general relativity using the LARES and LAGEOS satellites and a GRACE Earth gravity model\", by I. Ciufolini et al"
}
| null | null | null | null | true | null |
15505
| null |
Default
| null | null |
null |
{
"abstract": " NGC 7793 P13 is an ultraluminous X-ray source harboring an accreting pulsar.\nWe report on the detection of a ~65 d period X-ray modulation with Swift\nobservations in this system. The modulation period found in the X-ray band is\nP=65.05+/-0.10 d and the profile is asymmetric with a fast rise and a slower\ndecay. On the other hand, the u-band light curve collected by Swift UVOT\nconfirmed an optical modulation with a period of P=64.24+/-0.13 d. We explored\nthe phase evolution of the X-ray and optical periodicities and propose two\nsolutions. A superorbital modulation with a period of ~2,700-4,700 d probably\ncaused by the precession of a warped accretion disk is necessary to interpret\nthe phase drift of the optical data. We further discuss the implication if this\n~65d periodicity is caused by the superorbital modulation. Estimated from the\nrelationship between the spin-orbital and orbital-superorbital periods of known\ndisk-fed high-mass X-ray binaries, the orbital period of P13 is roughly\nestimated as 3-7 d. In this case, an unknown mechanism with a much longer time\nscale is needed to interpret the phase drift. Further studies on the stability\nof these two periodicities with a long-term monitoring could help us to probe\ntheir physical origins.\n",
"title": "SWIFT Detection of a 65-Day X-ray Period from the Ultraluminous Pulsar NGC 7793 P13"
}
| null | null | null | null | true | null |
15506
| null |
Default
| null | null |
null |
{
"abstract": " A new large-scale parallel multiconfigurational self-consistent field (MCSCF)\nimplementation in the open-source NWChem computational chemistry code is\npresented. The generalized active space (GAS) approach is used to partition\nlarge configuration interaction (CI) vectors and generate a sufficient number\nof batches that can be distributed to the available nodes. Massively parallel\nCI calculations with large active spaces can be treated. The performance of the\nnew parallel MCSCF implementation is presented for the chromium trimer and for\nan active space of 20 electrons in 20 orbitals. Unprecedented CI calculations\nwith an active space of 22 electrons in 22 orbitals for the pentacene systems\nwere performed and a single CI iteration calculation with an active space of 24\nelectrons in 24 orbitals for the chromium tetramer was possible. The chromium\ntetramer corresponds to a CI expansion of one trillion SDs (914 058 513 424)\nand is largest conventional CI calculation attempted up to date.\n",
"title": "Pushing Configuration-Interaction to the Limit: Towards Massively Parallel MCSCF Calculations"
}
| null | null | null | null | true | null |
15507
| null |
Default
| null | null |
null |
{
"abstract": " With the start of the Gaia era, the time has come to address the major\nchallenge of deriving the star formation history and evolution of the disk of\nour MilkyWay. Here we review our present knowledge of the outer regions of the\nMilky Way disk population. Its stellar content, its structure and its dynamical\nand chemical evolution are summarized, focussing on our lack of understanding\nboth from an observational and a theoretical viewpoint. We describe the\nunprecedented data that Gaia and the upcoming ground-based spectroscopic\nsurveys will provide in the next decade. More in detail, we quantify the expect\naccuracy in position, velocity and astrophysical parameters of some of the key\ntracers of the stellar populations in the outer Galactic disk. Some insights on\nthe future capability of these surveys to answer crucial and fundamental issues\nare discussed, such as the mechanisms driving the spiral arms and the warp\nformation. Our Galaxy, theMilkyWay, is our cosmological laboratory for\nunderstanding the process of formation and evolution of disk galaxies. What we\nlearn in the next decades will be naturally transferred to the extragalactic\ndomain.\n",
"title": "Outer Regions of the Milky Way"
}
| null | null |
[
"Physics"
] | null | true | null |
15508
| null |
Validated
| null | null |
null |
{
"abstract": " We propose a multi-objective framework to learn both secondary targets not\ndirectly related to the intended task of speech enhancement (SE) and the\nprimary target of the clean log-power spectra (LPS) features to be used\ndirectly for constructing the enhanced speech signals. In deep neural network\n(DNN) based SE we introduce an auxiliary structure to learn secondary\ncontinuous features, such as mel-frequency cepstral coefficients (MFCCs), and\ncategorical information, such as the ideal binary mask (IBM), and integrate it\ninto the original DNN architecture for joint optimization of all the\nparameters. This joint estimation scheme imposes additional constraints not\navailable in the direct prediction of LPS, and potentially improves the\nlearning of the primary target. Furthermore, the learned secondary information\nas a byproduct can be used for other purposes, e.g., the IBM-based\npost-processing in this work. A series of experiments show that joint LPS and\nMFCC learning improves the SE performance, and IBM-based post-processing\nfurther enhances listening quality of the reconstructed speech.\n",
"title": "Multi-Objective Learning and Mask-Based Post-Processing for Deep Neural Network Based Speech Enhancement"
}
| null | null | null | null | true | null |
15509
| null |
Default
| null | null |
null |
{
"abstract": " Diarization of audio recordings from ad-hoc mobile devices using spatial\ninformation is considered in this paper. A two-channel synchronous recording is\nassumed for each mobile device, which is used to compute directional statistics\nseparately at each device in a frame-wise manner. The recordings across the\nmobile devices are asynchronous, but a coarse synchronization is performed by\naligning the signals using acoustic events, or real-time clock. Direction\nstatistics computed for all the devices, are then modeled jointly using a\nDirichlet mixture model, and the posterior probability over the mixture\ncomponents is used to derive the diarization information. Experiments on real\nlife recordings using mobile phones show a diarization error rate of less than\n14%.\n",
"title": "Latent variable approach to diarization of audio recordings using ad-hoc randomly placed mobile devices"
}
| null | null | null | null | true | null |
15510
| null |
Default
| null | null |
null |
{
"abstract": " A model of incentive salience as a function of stimulus value and\ninteroceptive state has been previously proposed. In that model, the function\ndiffers depending on whether the stimulus is appetitive or aversive; it is\nmultiplicative for appetitive stimuli and additive for aversive stimuli. The\nauthors argued it was necessary to capture data on how extreme changes in salt\nappetite could move evaluation of an extreme salt solution from negative to\npositive. We demonstrate that arbitrarily varying this function is unnecessary,\nand that a multiplicative function is sufficient if one assumes the incentive\nsalience function for an incentive (such as salt) is comprised of multiple\nstimulus features and multiple interoceptive signals. We show that it is also\nunnecessary considering the dual-structure approach-aversive nature of the\nreward system, which results in separate weighting of appetitive and aversive\nstimulus features.\n",
"title": "A multiple attribute model resolves a conflict between additive and multiplicative models of incentive salience"
}
| null | null |
[
"Quantitative Biology"
] | null | true | null |
15511
| null |
Validated
| null | null |
null |
{
"abstract": " These are lecture notes based on three lectures given by Antonello\nScardicchio at the December 2016 Topical School on Many-Body-Localization\norganized by the Statistical Physics Group of the Institute Jean Lamour in\nNancy. They were compiled and put in a coherent logical form by Thimothée\nThiery.\n",
"title": "Perturbation theory approaches to Anderson and Many-Body Localization: some lecture notes"
}
| null | null | null | null | true | null |
15512
| null |
Default
| null | null |
null |
{
"abstract": " We report the detection of extended Halpha emission from the tip of the HI\ndisk of the nearby edge-on galaxy UGC 7321, observed with the Multi Unit\nSpectroscopic Explorer (MUSE) instrument at the Very Large Telescope. The\nHalpha surface brightness fades rapidly where the HI column density drops below\nN(HI) = 10^19 cm^-2 , consistent with fluorescence arising at the ionisation\nfront from gas that is photoionized by the extragalactic ultraviolet background\n(UVB). The surface brightness measured at this location is (1.2 +/- 0.5)x10^-19\nerg/s/cm^2/arcsec^2, where the error is mostly systematic and results from the\nproximity of the signal to the edge of the MUSE field of view, and from the\npresence of a sky line next to the redshifted Halpha wavelength. By combining\nthe Halpha and the HI 21 cm maps with a radiative transfer calculation of an\nexponential disk illuminated by the UVB, we derive a value for the HI\nphotoionization rate of Gamma ~ (6-8)x10^-14 1/s . This value is consistent\nwith transmission statistics of the Lyalpha forest and with recent models of a\nUVB which is dominated by quasars.\n",
"title": "A measurement of the z = 0 UV background from H$α$ fluorescence"
}
| null | null | null | null | true | null |
15513
| null |
Default
| null | null |
null |
{
"abstract": " We consider the problem of estimating a large rank-one tensor ${\\boldsymbol\nu}^{\\otimes k}\\in({\\mathbb R}^{n})^{\\otimes k}$, $k\\ge 3$ in Gaussian noise.\nEarlier work characterized a critical signal-to-noise ratio $\\lambda_{Bayes}=\nO(1)$ above which an ideal estimator achieves strictly positive correlation\nwith the unknown vector of interest. Remarkably no polynomial-time algorithm is\nknown that achieved this goal unless $\\lambda\\ge C n^{(k-2)/4}$ and even\npowerful semidefinite programming relaxations appear to fail for $1\\ll\n\\lambda\\ll n^{(k-2)/4}$.\nIn order to elucidate this behavior, we consider the maximum likelihood\nestimator, which requires maximizing a degree-$k$ homogeneous polynomial over\nthe unit sphere in $n$ dimensions. We compute the expected number of critical\npoints and local maxima of this objective function and show that it is\nexponential in the dimensions $n$, and give exact formulas for the exponential\ngrowth rate. We show that (for $\\lambda$ larger than a constant) critical\npoints are either very close to the unknown vector ${\\boldsymbol u}$, or are\nconfined in a band of width $\\Theta(\\lambda^{-1/(k-1)})$ around the maximum\ncircle that is orthogonal to ${\\boldsymbol u}$. For local maxima, this band\nshrinks to be of size $\\Theta(\\lambda^{-1/(k-2)})$. These `uninformative' local\nmaxima are likely to cause the failure of optimization algorithms.\n",
"title": "The landscape of the spiked tensor model"
}
| null | null | null | null | true | null |
15514
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, the parameter estimation problem for a multi-timescale\nadaptive threshold (MAT) neuronal model is investigated. By manipulating the\nsystem dynamics, which comprise of a non-resetting leaky integrator coupled\nwith an adaptive threshold, the threshold voltage can be obtained as a\nrealizable model that is linear in the unknown parameters. This linearly\nparametrized realizable model is then utilized inside a prediction error based\nframework to identify the threshold parameters with the purpose of predicting\nsingle neuron precise firing times. The iterative linear least squares\nestimation scheme is evaluated using both synthetic data obtained from an exact\nmodel as well as experimental data obtained from in vitro rat somatosensory\ncortical neurons. Results show the ability of this approach to fit the MAT\nmodel to different types of fluctuating reference data. The performance of the\nproposed approach is seen to be superior when comparing with existing\nidentification approaches used by the neuronal community.\n",
"title": "System Identification of a Multi-timescale Adaptive Threshold Neuronal Model"
}
| null | null | null | null | true | null |
15515
| null |
Default
| null | null |
null |
{
"abstract": " We introduce a new method to qualify the goodness of fit parameter estimation\nof compound Wishart models. Our method based on the free deterministic\nequivalent Z-score, which we introduce in this paper. Furthermore, an\napplication to two dimensional autoregressive moving-average model is provided.\nOur proposal method is a generalization of statistical hypothesis testing to\none dimensional moving average model based on fluctuations of real compound\nWishart matrices, which is a recent result by Hasegawa, Sakuma and Yoshida.\n",
"title": "Free deterministic equivalent Z-scores of compound Wishart models: A goodness of fit test of 2DARMA models"
}
| null | null | null | null | true | null |
15516
| null |
Default
| null | null |
null |
{
"abstract": " For non-Gaussian stochastic dynamical systems, mean exit time and escape\nprobability are important deterministic quantities, which can be obtained from\nintegro-differential (nonlocal) equations. We develop an efficient and\nconvergent numerical method for the mean first exit time and escape probability\nfor stochastic systems with an asymmetric Lévy motion, and analyze the\nproperties of the solutions of the nonlocal equations. We also investigate the\neffects of different system factors on the mean exit time and escape\nprobability, including the skewness parameter, the size of the domain, the\ndrift term and the intensity of Gaussian and non-Gaussian noises. We find that\nthe behavior of the mean exit time and the escape probability has dramatic\ndifference at the boundary of the domain when the index of stability crosses\nthe critical value of one.\n",
"title": "Numerical algorithms for mean exit time and escape probability of stochastic systems with asymmetric Lévy motion"
}
| null | null | null | null | true | null |
15517
| null |
Default
| null | null |
null |
{
"abstract": " We study modal team logic MTL, the team-semantical extension of modal logic\nML closed under Boolean negation. Its fragments, such as modal dependence,\nindependence, and inclusion logic, are well-understood. However, due to the\nunrestricted Boolean negation, the satisfiability problem of full MTL has been\nnotoriously resistant to a complexity theoretical classification. In our\napproach, we introduce the notion of canonical models into the team-semantical\nsetting. By construction of such a model, we reduce the satisfiability problem\nof MTL to simple model checking. Afterwards, we show that this approach is\noptimal in the sense that MTL-formulas can efficiently enforce canonicity.\nFurthermore, to capture these results in terms of complexity, we introduce a\nnon-elementary complexity class, TOWER(poly), and prove that it contains\nsatisfiability and validity of MTL as complete problems. We also prove that the\nfragments of MTL with bounded modal depth are complete for the levels of the\nelementary hierarchy (with polynomially many alternations). The respective\nhardness results hold for both strict or lax semantics of the modal operators\nand the splitting disjunction, and also over the class of reflexive and\ntransitive frames.\n",
"title": "Canonical Models and the Complexity of Modal Team Logic"
}
| null | null | null | null | true | null |
15518
| null |
Default
| null | null |
null |
{
"abstract": " We propose Cooperative Training (CoT) for training generative models that\nmeasure a tractable density for discrete data. CoT coordinately trains a\ngenerator $G$ and an auxiliary predictive mediator $M$. The training target of\n$M$ is to estimate a mixture density of the learned distribution $G$ and the\ntarget distribution $P$, and that of $G$ is to minimize the Jensen-Shannon\ndivergence estimated through $M$. CoT achieves independent success without the\nnecessity of pre-training via Maximum Likelihood Estimation or involving\nhigh-variance algorithms like REINFORCE. This low-variance algorithm is\ntheoretically proved to be unbiased for both generative and predictive tasks.\nWe also theoretically and empirically show the superiority of CoT over most\nprevious algorithms in terms of generative quality and diversity, predictive\ngeneralization ability and computational cost.\n",
"title": "CoT: Cooperative Training for Generative Modeling of Discrete Data"
}
| null | null | null | null | true | null |
15519
| null |
Default
| null | null |
null |
{
"abstract": " We associate to each iterated function system consisting of\nphi-max-contractions an operator (on the space of continuous functions from the\nshift space on the metric space corresponding to the system) having a unique\nfixed point whose image turns out to be the attractor of the system. Moreover,\nwe prove that the unique fixed point of the operator associated to an iterated\nfunction system consisting of convex contractions is the canonical projection\nfrom the shift space on the attractor of the system.\n",
"title": "Iterated function systems consisting of phi-max-contractions have attractor"
}
| null | null | null | null | true | null |
15520
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we study the energy decay for the thermoelastic Bresse system\nin the whole line with two different dissipative mechanism, given by heat\nconduction (Types I and III). We prove that the decay rate of the solutions are\nvery slow. More precisely, we show that the solutions decay with the rate of\n$(1+t)^{-\\frac{1}{8}}$ in the $L^2$-norm, whenever the initial data belongs to\n$L^1(R) \\cap H^{s}(R)$ for a suitable $s$. The wave speeds of propagation have\ninfluence on the decay rate with respect to the regularity of the initial data.\nThis phenomenon is known as \\textit{regularity-loss}. The main tool used to\nprove our results is the energy method in the Fourier space.\n",
"title": "Decay Rates of the Solutions to the Thermoelastic Bresse System of Types I and III"
}
| null | null |
[
"Mathematics"
] | null | true | null |
15521
| null |
Validated
| null | null |
null |
{
"abstract": " We study a scenario in which the baryon asymmetry of the universe arises from\na cosmological phase transition where lepton-number is spontaneously broken. If\nthe phase transition is first order, a lepton-number asymmetry can arise at the\nbubble wall, through dynamics similar to electroweak baryogenesis, but\ninvolving right-handed neutrinos. In addition to the usual neutrinoless double\nbeta decay in nuclear experiments, the model may be probed through a variety of\n\"baryogenesis by-products,\" which include a stochastic background of\ngravitational waves created by the colliding bubbles. Depending on the model,\nother aspects may include a network of topological defects that produce their\nown gravitational waves, additional contribution to dark radiation, and a light\npseudo-Goldstone boson (majoron) as dark matter candidate.\n",
"title": "Baryogenesis at a Lepton-Number-Breaking Phase Transition"
}
| null | null | null | null | true | null |
15522
| null |
Default
| null | null |
null |
{
"abstract": " No-insulation (NI) REBCO magnets have many advantages. They are\nself-protecting, therefore do not need quench detection and protection which\ncan be very challenging in a high Tc superconducting magnet. Moreover, by\nremoving insulation and allowing thinner copper stabilizer, NI REBCO magnets\nhave significantly higher engineering current density and higher mechanical\nstrength. On the other hand, NI REBCO magnets have drawbacks of long magnet\ncharging time and high field-ramp-loss. In principle, these drawbacks can be\nmitigated by managing the turn-to-turn contact resistivity (Rc). Evidently the\nfirst step toward managing Rc is to establish a reliable method of accurate Rc\nmeasurement. In this paper, we present experimental Rc measurements of REBCO\ntapes as a function of mechanical load up to 144 MPa and load cycles up to 14\ntimes. We found that Rc is in the range of 26-100 uOhm-cm2; it decreases with\nincreasing pressure, and gradually increases with number of load cycles. The\nresults are discussed in the framework of Holm's electric contact theory.\n",
"title": "Contact resistance between two REBCO tapes under load and load-cycles"
}
| null | null | null | null | true | null |
15523
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we propose an implicit gradient descent algorithm for the\nclassic $k$-means problem. The implicit gradient step or backward Euler is\nsolved via stochastic fixed-point iteration, in which we randomly sample a\nmini-batch gradient in every iteration. It is the average of the fixed-point\ntrajectory that is carried over to the next gradient step. We draw connections\nbetween the proposed stochastic backward Euler and the recent entropy\nstochastic gradient descent (Entropy-SGD) for improving the training of deep\nneural networks. Numerical experiments on various synthetic and real datasets\nshow that the proposed algorithm provides better clustering results compared to\n$k$-means algorithms in the sense that it decreased the objective function (the\ncluster) and is much more robust to initialization.\n",
"title": "Stochastic Backward Euler: An Implicit Gradient Descent Algorithm for $k$-means Clustering"
}
| null | null | null | null | true | null |
15524
| null |
Default
| null | null |
null |
{
"abstract": " We compute the Frobenius number for sequences of triangular and tetrahedral\nnumbers. In addition, we study some properties of the numerical semigroups\nassociated to those sequences.\n",
"title": "The Frobenius number for sequences of triangular and tetrahedral numbers"
}
| null | null | null | null | true | null |
15525
| null |
Default
| null | null |
null |
{
"abstract": " Solitary waves propagation of baryonic density perturbations, ruled by the\nKorteweg--de Vries equation in a mean-field quark-gluon plasma model, are\ninvestigated from the point of view of the theory of information. A recently\nproposed continuous logarithmic measure of information, called configurational\nentropy, is used to derive the soliton width, defining the pulse, for which the\ninformational content of the soliton spatial profile is more compressed, in the\nShannon's sense.\n",
"title": "Information-entropic analysis of Korteweg--de Vries solitons in the quark-gluon plasma"
}
| null | null | null | null | true | null |
15526
| null |
Default
| null | null |
null |
{
"abstract": " Recent Einstein-Podolsky-Rosen-Bohm experiments [M. Giustina et al. Phys.\nRev. Lett. 115, 250401 (2015); L. K. Shalm et al. Phys. Rev. Lett. 115, 250402\n(2015)] that claim to be loophole free are scrutinized and are shown to suffer\na photon identification loophole. The combination of a digital computer and\ndiscrete-event simulation is used to construct a minimal but faithful model of\nthe most perfected realization of these laboratory experiments. In contrast to\nprior simulations, all photon selections are strictly made, as they are in the\nactual experiments, at the local station and no other \"post-selection\" is\ninvolved. The simulation results demonstrate that a manifestly non-quantum\nmodel that identifies photons in the same local manner as in these experiments\ncan produce correlations that are in excellent agreement with those of the\nquantum theoretical description of the corresponding thought experiment, in\nconflict with Bell's theorem. The failure of Bell's theorem is possible because\nof our recognition of the photon identification loophole. Such identification\nmeasurement-procedures are necessarily included in all actual experiments but\nare not included in the theory of Bell and his followers.\n",
"title": "The photon identification loophole in EPRB experiments: computer models with single-wing selection"
}
| null | null | null | null | true | null |
15527
| null |
Default
| null | null |
null |
{
"abstract": " There exist tilings of the plane with pairwise noncongruent triangles of\nequal area and bounded perimeter. Analogously, there exist tilings with\ntriangles of equal perimeter, the areas of which are bounded from below by a\npositive constant. This solves a problem of Nandakumar.\n",
"title": "Tilings of the plane with unit area triangles of bounded diameter"
}
| null | null | null | null | true | null |
15528
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, we study the Cauchy problem for radially symmetric homogeneous\nnon-cutoff Boltzmann equation with Maxwellian molecules, the initial datum\nbelongs to Shubin space of the negative index which can be characterized by\nspectral decomposition of the harmonic oscillators. The Shubin space of the\nnegative index contains the measure functions. Based on this spectral\ndecomposition, we construct the weak solution with Shubin class initial datum,\nwe also prove that the Cauchy problem enjoys Gelfand-Shilov smoothing effect,\nmeaning that the smoothing properties are the same as the Cauchy problem\ndefined by the evolution equation associated to a fractional harmonic\noscillator.\n",
"title": "The cauchy problem for radially symmetric homogeneous boltzmann equation with shubin class initial datum and gelfand-shilov smoothing effect"
}
| null | null | null | null | true | null |
15529
| null |
Default
| null | null |
null |
{
"abstract": " We propose a general framework called Network Dissection for quantifying the\ninterpretability of latent representations of CNNs by evaluating the alignment\nbetween individual hidden units and a set of semantic concepts. Given any CNN\nmodel, the proposed method draws on a broad data set of visual concepts to\nscore the semantics of hidden units at each intermediate convolutional layer.\nThe units with semantics are given labels across a range of objects, parts,\nscenes, textures, materials, and colors. We use the proposed method to test the\nhypothesis that interpretability of units is equivalent to random linear\ncombinations of units, then we apply our method to compare the latent\nrepresentations of various networks when trained to solve different supervised\nand self-supervised training tasks. We further analyze the effect of training\niterations, compare networks trained with different initializations, examine\nthe impact of network depth and width, and measure the effect of dropout and\nbatch normalization on the interpretability of deep visual representations. We\ndemonstrate that the proposed method can shed light on characteristics of CNN\nmodels and training methods that go beyond measurements of their discriminative\npower.\n",
"title": "Network Dissection: Quantifying Interpretability of Deep Visual Representations"
}
| null | null | null | null | true | null |
15530
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we study the systole growth of arithmetic locally symmetric\nspaces up congruence covers and show that this growth is at least logarithmic\nin volume. This generalizes previous work of Buser and Sarnak as well as Katz,\nSchaps and Vishne where the case of compact hyperbolic 2- and 3-manifolds was\nconsidered.\n",
"title": "Systole inequalities for arithmetic locally symmetric spaces"
}
| null | null | null | null | true | null |
15531
| null |
Default
| null | null |
null |
{
"abstract": " A featured transition system is a transition system in which the transitions\nare annotated with feature expressions: Boolean expressions on a finite number\nof given features. Depending on its feature expression, each individual\ntransition can be enabled when some features are present, and disabled for\nother sets of features. The behavior of a featured transition system hence\ndepends on a given set of features. There are algorithms for featured\ntransition systems which can check their properties for all sets of features at\nonce, for example for LTL or CTL properties.\nHere we introduce a model of featured weighted automata which combines\nfeatured transition systems and (semiring-) weighted automata. We show that\nmethods and techniques from weighted automata extend to featured weighted\nautomata and devise algorithms to compute quantitative properties of featured\nweighted automata for all sets of features at once. We show applications to\nminimum reachability and to energy properties.\n",
"title": "Featured Weighted Automata"
}
| null | null | null | null | true | null |
15532
| null |
Default
| null | null |
null |
{
"abstract": " We introduce a Schrödinger model for the unitary irreducible\nrepresentations of a Heisenberg motion group and we show that the usual Weyl\nquantization then provides a Stratonovich-Weyl correspondence.\n",
"title": "Schrödinger model and Stratonovich-Weyl correspondence for Heisenberg motion groups"
}
| null | null | null | null | true | null |
15533
| null |
Default
| null | null |
null |
{
"abstract": " When analysing new emerging infectious disease outbreaks one typically has\nobservational data over a limited period of time and several parameters to\nestimate, such as growth rate, R0, serial or generation interval distribution,\nlatent and incubation times or case fatality rates. Also parameters describing\nthe temporal relations between appearance of symptoms, notification, death and\nrecovery/discharge will be of interest. These parameters form the basis for\npredicting the future outbreak, planning preventive measures and monitoring the\nprogress of the disease. We study the problem of making inference during the\nemerging phase of an outbreak and point out potential sources of bias related\nto contact tracing, replacing generation times by serial intervals, multiple\npotential infectors or truncation effects amplified by exponential growth.\nThese biases directly affect the estimation of e.g. the generation time\ndistribution and the case fatality rate, but can then propagate to other\nestimates, e.g. of R0 and growth rate. Many of the traditionally used\nestimation methods in disease epidemiology may suffer from these biases when\napplied to the emerging disease outbreak situation. We show how to avoid these\nbiases based on proper statistical modelling. We illustrate the theory by\nnumerical examples and simulations based on the recent 2014-15 Ebola outbreak\nto quantify possible estimation biases, which may be up to 20% underestimation\nof R0, if the epidemic growth rate is fitted to observed data or, conversely,\nup to 62% overestimation of the growth rate if the correct R0 is used in\nconjunction with the Euler-Lotka equation.\n",
"title": "Estimation in emerging epidemics: biases and remedies"
}
| null | null | null | null | true | null |
15534
| null |
Default
| null | null |
null |
{
"abstract": " We evolve binary mux-6 trees for up to 100000 generations evolving some\nprograms with more than a hundred million nodes. Our unbounded Long-Term\nEvolution Experiment LTEE GP appears not to evolve building blocks but does\nsuggests a limit to bloat. We do see periods of tens even hundreds of\ngenerations where the population is 100 percent functionally converged. The\ndistribution of tree sizes is not as predicted by theory.\n",
"title": "Long-Term Evolution of Genetic Programming Populations"
}
| null | null |
[
"Computer Science"
] | null | true | null |
15535
| null |
Validated
| null | null |
null |
{
"abstract": " OpenStreetMap offers a valuable source of worldwide geospatial data useful to\nurban researchers. This study uses the OSMnx software to automatically download\nand analyze 27,000 US street networks from OpenStreetMap at metropolitan,\nmunicipal, and neighborhood scales - namely, every US city and town, census\nurbanized area, and Zillow-defined neighborhood. It presents empirical findings\non US urban form and street network characteristics, emphasizing measures\nrelevant to graph theory, transportation, urban design, and morphology such as\nstructure, connectedness, density, centrality, and resilience. In the past,\nstreet network data acquisition and processing have been challenging and ad\nhoc. This study illustrates the use of OSMnx and OpenStreetMap to consistently\nconduct street network analysis with extremely large sample sizes, with clearly\ndefined network definitions and extents for reproducibility, and using\nnonplanar, directed graphs. These street networks and measures data have been\nshared in a public repository for other researchers to use.\n",
"title": "A Multi-Scale Analysis of 27,000 Urban Street Networks: Every US City, Town, Urbanized Area, and Zillow Neighborhood"
}
| null | null | null | null | true | null |
15536
| null |
Default
| null | null |
null |
{
"abstract": " Detection of a planetary ring of exoplanets remains as one of the most\nattractive but challenging goals in the field. We present a methodology of a\nsystematic search for exoplanetary rings via transit photometry of long-period\nplanets. The methodology relies on a precise integration scheme we develop to\ncompute a transit light curve of a ringed planet. We apply the methodology to\n89 long-period planet candidates from the Kepler data so as to estimate, and/or\nset upper limits on, the parameters of possible rings. While a majority of our\nsamples do not have a sufficiently good signal-to-noise ratio for meaningful\nconstraints on ring parameters, we find that six systems with a higher\nsignal-to-noise ratio are inconsistent with the presence of a ring larger than\n1.5 times the planetary radius assuming a grazing orbit and a tilted ring.\nFurthermore, we identify five preliminary candidate systems whose light curves\nexhibit ring-like features. After removing four false positives due to the\ncontamination from nearby stars, we identify KIC 10403228 as a reasonable\ncandidate for a ringed planet. A systematic parameter fit of its light curve\nwith a ringed planet model indicates two possible solutions corresponding to a\nSaturn-like planet with a tilted ring. There also remain other two possible\nscenarios accounting for the data; a circumstellar disk and a hierarchical\ntriple. Due to large uncertain factors, we cannot choose one specific model\namong the three.\n",
"title": "Towards Detection of Exoplanetary Rings Via Transit Photometry: Methodology and a Possible Candidate"
}
| null | null | null | null | true | null |
15537
| null |
Default
| null | null |
null |
{
"abstract": " Scientific discovery via numerical simulations is important in modern\nastrophysics. This relatively new branch of astrophysics has become possible\ndue to the development of reliable numerical algorithms and the high\nperformance of modern computing technologies. These enable the analysis of\nlarge collections of observational data and the acquisition of new data via\nsimulations at unprecedented accuracy and resolution. Ideally, simulations run\nuntil they reach some pre-determined termination condition, but often other\nfactors cause extensive numerical approaches to break down at an earlier stage.\nIn those cases, processes tend to be interrupted due to unexpected events in\nthe software or the hardware. In those cases, the scientist handles the\ninterrupt manually, which is time-consuming and prone to errors. We present the\nSimulation Monitor (SiMon) to automatize the farming of large and extensive\nsimulation processes. Our method is light-weight, it fully automates the entire\nworkflow management, operates concurrently across multiple platforms and can be\ninstalled in user space. Inspired by the process of crop farming, we perceive\neach simulation as a crop in the field and running simulation becomes analogous\nto growing crops. With the development of SiMon we relax the technical aspects\nof simulation management. The initial package was developed for extensive\nparameter searchers in numerical simulations, but it turns out to work equally\nwell for automating the computational processing and reduction of observational\ndata reduction.\n",
"title": "SiMon: Simulation Monitor for Computational Astrophysics"
}
| null | null | null | null | true | null |
15538
| null |
Default
| null | null |
null |
{
"abstract": " Dimension reduction is often a preliminary step in the analysis of large data\nsets. The so-called non-Gaussian component analysis searches for a projection\nonto the non-Gaussian part of the data, and it is then important to know the\ncorrect dimension of the non-Gaussian signal subspace. In this paper we develop\nasymptotic as well as bootstrap tests for the dimension based on the popular\nfourth order blind identification (FOBI) method.\n",
"title": "Asymptotic and bootstrap tests for the dimension of the non-Gaussian subspace"
}
| null | null |
[
"Mathematics",
"Statistics"
] | null | true | null |
15539
| null |
Validated
| null | null |
null |
{
"abstract": " A flag domain of a real from $G_0$ of a complex semismiple Lie group $G$ is\nan open $G_0$-orbit $D$ in a (compact) $G$-flag manifold. In the usual way one\nreduces to the case where $G_0$ is simple. It is known that if $D$ possesses\nnon-constant holomorphic functions, then it is the product of a compact flag\nmanifold and a Hermitian symmetric bounded domain. This pseudoconvex case is\nrare in the geography of flag domains. Here it is shown that otherwise, i.e.,\nwhen $\\mathcal{O}(D)\\cong\\mathbb{C}$, the flag domain $D$ is pseudoconcave. In\na rather general setting the degree of the pseudoconcavity is estimated in\nterms of root invariants. This estimate is explicitly computed for domains in\ncertain Grassmannians.\n",
"title": "Pseudoconcavity of flag domains: The method of supporting cycles"
}
| null | null | null | null | true | null |
15540
| null |
Default
| null | null |
null |
{
"abstract": " The nominal transition systems (NTSs) of Parrow et al. describe the\noperational semantics of nominal process calculi. We study NTSs in terms of the\nnominal residual transition systems (NRTSs) that we introduce. We provide rule\nformats for the specifications of NRTSs that ensure that the associated NRTS is\nan NTS and apply them to the operational specifications of the early and late\npi-calculus. We also explore alternative specifications of the NTSs in which we\nallow residuals of abstraction sort, and introduce translations between the\nsystems with and without residuals of abstraction sort. Our study stems from\nthe Nominal SOS of Cimini et al. and from earlier works in nominal sets and\nnominal logic by Gabbay, Pitts and their collaborators.\n",
"title": "Rule Formats for Nominal Process Calculi"
}
| null | null |
[
"Computer Science"
] | null | true | null |
15541
| null |
Validated
| null | null |
null |
{
"abstract": " We solve a problem of R. Nandakumar by proving that there is no tiling of the\nplane with pairwise noncongruent triangles of equal area and equal perimeter.\nWe also show that no convex polygon with more than three sides can be tiled\nwith finitely many triangles such that no pair of them share a full side.\n",
"title": "Tilings with noncongruent triangles"
}
| null | null | null | null | true | null |
15542
| null |
Default
| null | null |
null |
{
"abstract": " Phase limitations of both continuous-time and discrete-time Zames-Falb\nmultipliers and their relation with the Kalman conjecture are analysed. A phase\nlimitation for continuous-time multipliers given by Megretski is generalised\nand its applicability is clarified; its relation to the Kalman conjecture is\nillustrated with a classical example from the literature. It is demonstrated\nthat there exist fourth-order plants where the existence of a suitable\nZames-Falb multiplier can be discarded and for which simulations show unstable\nbehavior. A novel phase-limitation for discrete-time Zames-Falb multipliers is\ndeveloped. Its application is demonstrated with a second-order counterexample\nto the Kalman conjecture. Finally, the discrete-time limitation is used to show\nthat there can be no direct counterpart of the off-axis circle criterion in the\ndiscrete-time domain.\n",
"title": "Phase limitations of Zames-Falb multipliers"
}
| null | null | null | null | true | null |
15543
| null |
Default
| null | null |
null |
{
"abstract": " We investigate, using 3D hydrodynamic simulations, the fragmentation of\npressure-confined, vertically stratified, self-gravitating gaseous layers. The\nconfining pressure is either thermal pressure acting on both surfaces, or\nthermal pressure acting on one surface and ram-pressure on the other. In the\nlinear regime of fragmentation, the dispersion relation we obtain agrees well\nwith that derived by Elmegreen & Elmegreen (1978), and consequently deviates\nfrom the dispersion relations based on the thin shell approximation (Vishniac\n1983) or pressure assisted gravitational instability (Wünsch et al. 2010). In\nthe non-linear regime, the relative importance of the confining pressure to the\nself-gravity is a crucial parameter controlling the qualitative course of\nfragmentation. When confinement of the layer is dominated by external pressure,\nself- gravitating condensations are delivered by a two-stage process: first the\nlayer fragments into gravitationally bound but stable clumps, and then these\nclumps coalesce until they assemble enough mass to collapse. In contrast, when\nexternal pressure makes a small contribution to confinement of the layer, the\nlayer fragments monolithically into gravitationally unstable clumps and there\nis no coalescence. This dichotomy persists whether the external pressure is\nthermal or ram. We apply these results to fragments forming in a shell swept up\nby an expanding H II region, and find that, unless the swept up gas is quite\nhot or the surrounding medium has low density, the fragments have low-mass ( ~<\n3 M_Sun ), and therefore they are unlikely to spawn stars that are sufficiently\nmassive to promote sequential self-propagating star formation.\n",
"title": "Fragmentation of vertically stratified gaseous layers: monolithic or coalescence-driven collapse"
}
| null | null | null | null | true | null |
15544
| null |
Default
| null | null |
null |
{
"abstract": " We propose a simple and generic layer formulation that extends the properties\nof convolutional layers to any domain that can be described by a graph. Namely,\nwe use the support of its adjacency matrix to design learnable weight sharing\nfilters able to exploit the underlying structure of signals in the same fashion\nas for images. The proposed formulation makes it possible to learn the weights\nof the filter as well as a scheme that controls how they are shared across the\ngraph. We perform validation experiments with image datasets and show that\nthese filters offer performances comparable with convolutional ones.\n",
"title": "Learning Local Receptive Fields and their Weight Sharing Scheme on Graphs"
}
| null | null | null | null | true | null |
15545
| null |
Default
| null | null |
null |
{
"abstract": " Stellar shells are low surface brightness arcs of overdense stellar regions,\nextending to large galactocentric distances. In a companion study, we\nidentified 39 shell galaxies in a sample of 220 massive ellipticals\n($\\mathrm{M}_{\\mathrm{200crit}}>6\\times10^{12}\\,\\mathrm{M}_\\odot$) from the\nIllustris cosmological simulation. We used stellar history catalogs to trace\nthe history of each individual star particle inside the shell substructures,\nand we found that shells in high-mass galaxies form through mergers with\nmassive satellites (stellar mass ratios $\\mu_{\\mathrm{stars}}\\gtrsim1:10$).\nUsing the same sample of shell galaxies, the current study extends the stellar\nhistory catalogs in order to investigate the metallicity of stellar shells\naround massive galaxies. Our results indicate that outer shells are often times\nmore metal-rich than the surrounding stellar material in a galaxy's halo. For a\ngalaxy with two different satellites forming $z=0$ shells, we find a\nsignificant difference in the metallicity of the shells produced by each\nprogenitor. We also find that shell galaxies have higher mass-weighted\nlogarithmic metallicities ([Z/H]) at $2$-$4\\,\\mathrm{R}_{\\mathrm{eff}}$\ncompared to galaxies without shells. Our results indicate that observations\ncomparing the metallicities of stars in tidal features, such as shells, to the\naverage metallicities in the stellar halo can provide information about the\nassembly histories of galaxies.\n",
"title": "Galaxies with Shells in the Illustris Simulation: Metallicity Signatures"
}
| null | null |
[
"Physics"
] | null | true | null |
15546
| null |
Validated
| null | null |
null |
{
"abstract": " We present a monitoring approach for verifying systems at runtime. Our\napproach targets systems whose components communicate with the monitors over\nunreliable channels, where messages can be delayed or lost. In contrast to\nprior works, whose property specification languages are limited to\npropositional temporal logics, our approach handles an extension of the\nreal-time logic MTL with freeze quantifiers for reasoning about data values. We\npresent its underlying theory based on a new three-valued semantics that is\nwell suited to soundly and completely reason online about event streams in the\npresence of message delay or loss. We also evaluate our approach\nexperimentally. Our prototype implementation processes hundreds of events per\nsecond in settings where messages are received out of order.\n",
"title": "Runtime Verification of Temporal Properties over Out-of-order Data Streams"
}
| null | null |
[
"Computer Science"
] | null | true | null |
15547
| null |
Validated
| null | null |
null |
{
"abstract": " In recent years, deep learning algorithms have become increasingly more\nprominent for their unparalleled ability to automatically learn discriminant\nfeatures from large amounts of data. However, within the field of\nelectromyography-based gesture recognition, deep learning algorithms are seldom\nemployed as they require an unreasonable amount of effort from a single person,\nto generate tens of thousands of examples.\nThis work's hypothesis is that general, informative features can be learned\nfrom the large amounts of data generated by aggregating the signals of multiple\nusers, thus reducing the recording burden while enhancing gesture recognition.\nConsequently, this paper proposes applying transfer learning on aggregated data\nfrom multiple users, while leveraging the capacity of deep learning algorithms\nto learn discriminant features from large datasets. Two datasets comprised of\n19 and 17 able-bodied participants respectively (the first one is employed for\npre-training) were recorded for this work, using the Myo Armband. A third Myo\nArmband dataset was taken from the NinaPro database and is comprised of 10\nable-bodied participants. Three different deep learning networks employing\nthree different modalities as input (raw EMG, Spectrograms and Continuous\nWavelet Transform (CWT)) are tested on the second and third dataset. The\nproposed transfer learning scheme is shown to systematically and significantly\nenhance the performance for all three networks on the two datasets, achieving\nan offline accuracy of 98.31% for 7 gestures over 17 participants for the\nCWT-based ConvNet and 68.98% for 18 gestures over 10 participants for the raw\nEMG-based ConvNet. Finally, a use-case study employing eight able-bodied\nparticipants suggests that real-time feedback allows users to adapt their\nmuscle activation strategy which reduces the degradation in accuracy normally\nexperienced over time.\n",
"title": "Deep Learning for Electromyographic Hand Gesture Signal Classification Using Transfer Learning"
}
| null | null | null | null | true | null |
15548
| null |
Default
| null | null |
null |
{
"abstract": " In this paper, adaptive non-uniform compressive sampling (ANCS) of\ntime-varying signals, which are sparse in a proper basis, is introduced. ANCS\nemploys the measurements of previous time steps to distribute the sensing\nenergy among coefficients more intelligently. To this aim, a Bayesian inference\nmethod is proposed that does not require any prior knowledge of importance\nlevels of coefficients or sparsity of the signal. Our numerical simulations\nshow that ANCS is able to achieve the desired non-uniform recovery of the\nsignal. Moreover, if the signal is sparse in canonical basis, ANCS can reduce\nthe number of required measurements significantly.\n",
"title": "Adaptive Non-uniform Compressive Sampling for Time-varying Signals"
}
| null | null | null | null | true | null |
15549
| null |
Default
| null | null |
null |
{
"abstract": " In order to automate verification process, regulatory rules written in\nnatural language need to be translated into a format that machines can\nunderstand. However, none of the existing formalisms can fully represent the\nelements that appear in legal norms. For instance, most of these formalisms do\nnot provide features to capture the behavior of deontic effects, which is an\nimportant aspect in automated compliance checking. This paper presents an\napproach for transforming legal norms represented using LegalRuleML to a\nvariant of Modal Defeasible Logic (and vice versa) such that a legal statement\nrepresented using LegalRuleML can be transformed into a machine-readable format\nthat can be understood and reasoned about depending upon the client's\npreferences.\n",
"title": "Enabling Reasoning with LegalRuleML"
}
| null | null | null | null | true | null |
15550
| null |
Default
| null | null |
null |
{
"abstract": " We survey problems and results from combinatorial geometry in normed spaces,\nconcentrating on problems that involve distances. These include various\nproperties of unit-distance graphs, minimum-distance graphs, diameter graphs,\nas well as minimum spanning trees and Steiner minimum trees. In particular, we\ndiscuss translative kissing (or Hadwiger) numbers, equilateral sets, and the\nBorsuk problem in normed spaces. We show how to use the angular measure of\nPeter Brass to prove various statements about Hadwiger and blocking numbers of\nconvex bodies in the plane, including some new results. We also include some\nnew results on thin cones and their application to distinct distances and other\ncombinatorial problems for normed spaces.\n",
"title": "Combinatorial distance geometry in normed spaces"
}
| null | null | null | null | true | null |
15551
| null |
Default
| null | null |
null |
{
"abstract": " A recently proposed learning algorithm for massive network-structured data\nsets (big data over networks) is the network Lasso (nLasso), which extends the\nwell- known Lasso estimator from sparse models to network-structured datasets.\nEfficient implementations of the nLasso have been presented using modern convex\noptimization methods. In this paper, we provide sufficient conditions on the\nnetwork structure and available label information such that nLasso accurately\nlearns a vector-valued graph signal (representing label information) from the\ninformation provided by the labels of a few data points.\n",
"title": "When is Network Lasso Accurate: The Vector Case"
}
| null | null | null | null | true | null |
15552
| null |
Default
| null | null |
null |
{
"abstract": " We consider the recovery of regression coefficients, denoted by\n$\\boldsymbol{\\beta}_0$, for a single index model (SIM) relating a binary\noutcome $Y$ to a set of possibly high dimensional covariates $\\boldsymbol{X}$,\nbased on a large but 'unlabeled' dataset $\\mathcal{U}$, with $Y$ never\nobserved. On $\\mathcal{U}$, we fully observe $\\boldsymbol{X}$ and additionally,\na surrogate $S$ which, while not being strongly predictive of $Y$ throughout\nthe entirety of its support, can forecast it with high accuracy when it assumes\nextreme values. Such datasets arise naturally in modern studies involving large\ndatabases such as electronic medical records (EMR) where $Y$, unlike\n$(\\boldsymbol{X}, S)$, is difficult and/or expensive to obtain. In EMR studies,\nan example of $Y$ and $S$ would be the true disease phenotype and the count of\nthe associated diagnostic codes respectively. Assuming another SIM for $S$\ngiven $\\boldsymbol{X}$, we show that under sparsity assumptions, we can recover\n$\\boldsymbol{\\beta}_0$ proportionally by simply fitting a least squares LASSO\nestimator to the subset of the observed data on $(\\boldsymbol{X}, S)$\nrestricted to the extreme sets of $S$, with $Y$ imputed using the surrogacy of\n$S$. We obtain sharp finite sample performance bounds for our estimator,\nincluding deterministic deviation bounds and probabilistic guarantees. We\ndemonstrate the effectiveness of our approach through multiple simulation\nstudies, as well as by application to real data from an EMR study conducted at\nthe Partners HealthCare Systems.\n",
"title": "Surrogate Aided Unsupervised Recovery of Sparse Signals in Single Index Models for Binary Outcomes"
}
| null | null | null | null | true | null |
15553
| null |
Default
| null | null |
null |
{
"abstract": " By way of the nonequilibrium Green's function simulations and first\nprinciples calculations, we report that borophene, a single layer of boron\natoms that was fabricated recently, possesses an extraordinarily high lattice\nthermal conductance in the ballistic transport regime, which even exceeds\ngraphene. In addition to the obvious reasons of light mass and strong bonding\nof boron atoms, the superior thermal conductance is mainly rooted in its strong\nstructural anisotropy and unusual phonon transmission. For low-frequency\nphonons, the phonon transmission within borophene is nearly isotropic, similar\nto that of graphene. For high frequency phonons, however, the transmission is\none dimensional, that is, all the phonons travel in one direction, giving rise\nto its ultrahigh thermal conductance. The present study suggests that borophene\nis promising for applications in efficient heat dissipation and thermal\nmanagement, and also an ideal material for revealing fundamentals of\ndimensionality effect on phonon transport in ballistic regime.\n",
"title": "Superior lattice thermal conductance of single layer borophene"
}
| null | null | null | null | true | null |
15554
| null |
Default
| null | null |
null |
{
"abstract": " The liar paradox is widely seen as not a serious problem. I try to explain\nwhy this view is mistaken.\n",
"title": "The liar paradox is a real problem"
}
| null | null |
[
"Mathematics"
] | null | true | null |
15555
| null |
Validated
| null | null |
null |
{
"abstract": " This Chapter, \"High-dimensional ABC\", is to appear in the forthcoming\nHandbook of Approximate Bayesian Computation (2018). It details the main ideas\nand concepts behind extending ABC methods to higher dimensions, with supporting\nexamples and illustrations.\n",
"title": "High-dimensional ABC"
}
| null | null | null | null | true | null |
15556
| null |
Default
| null | null |
null |
{
"abstract": " Unlike the Web where each web page has a global URL to reach, a specific\n\"content page\" inside a mobile app cannot be opened unless the user explores\nthe app with several operations from the landing page. Recently, deep links\nhave been advocated by major companies to enable targeting and opening a\nspecific page of an app externally with an accessible uniform resource\nidentifier (URI). To empirically investigate the state of the practice on\nadopting deep links, in this article, we present the largest empirical study of\ndeep links over 20,000 Android apps, and find that deep links do not get wide\nadoption among current Android apps, and non-trivial manual efforts are\nrequired for app developers to support deep links. To address such an issue, we\npropose the Aladdin approach and supporting tool to release deep links to\naccess arbitrary location of existing apps. Aladdin instantiates our novel\ncooperative framework to synergically combine static analysis and dynamic\nanalysis while minimally engaging developers to provide inputs to the framework\nfor automation, without requiring any coding efforts or additional deployment\nefforts. We evaluate Aladdin with popular apps and demonstrate its\neffectiveness and performance.\n",
"title": "Automating Release of Deep Link APIs for Android Applications"
}
| null | null | null | null | true | null |
15557
| null |
Default
| null | null |
null |
{
"abstract": " In this chapter, we introduce digital holographic microscopy (DHM) as a\nmarker-free method to determine the refractive index of single, spherical cells\nin suspension. The refractive index is a conclusive measure in a biological\ncontext. Cell conditions, such as differentiation or infection, are known to\nyield significant changes in the refractive index. Furthermore, the refractive\nindex of biological tissue determines the way it interacts with light. Besides\nthe biological relevance of this interaction in the retina, a lot of methods\nused in biology, including microscopy, rely on light-tissue or light-cell\ninteractions. Hence, determining the refractive index of cells using DHM is\nvaluable in many biological applications. This chapter covers the main topics\nwhich are important for the implementation of DHM: setup, sample preparation\nand analysis. First, the optical setup is described in detail including notes\nand suggestions for the implementation. Following that, a protocol for the\nsample and measurement preparation is explained. In the analysis section, an\nalgorithm for the determination of the quantitative phase map is described.\nSubsequently, all intermediate steps for the calculation of the refractive\nindex of suspended cells are presented, exploiting their spherical shape. In\nthe last section, a discussion of possible extensions to the setup, further\nmeasurement configurations and additional analysis methods are given.\nThroughout this chapter, we describe a simple, robust, and thus easily\nreproducible implementation of DHM. The different possibilities for extensions\nshow the diverse fields of application for this technique.\n",
"title": "Refractive index measurements of single, spherical cells using digital holographic microscopy"
}
| null | null | null | null | true | null |
15558
| null |
Default
| null | null |
null |
{
"abstract": " Data compression is a popular technique for improving the efficiency of data\nprocessing workloads such as SQL queries and more recently, machine learning\n(ML) with classical batch gradient methods. But the efficacy of such ideas for\nmini-batch stochastic gradient descent (MGD), arguably the workhorse algorithm\nof modern ML, is an open question. MGD's unique data access pattern renders\nprior art, including those designed for batch gradient methods, less effective.\nWe fill this crucial research gap by proposing a new lossless compression\nscheme we call tuple-oriented compression (TOC) that is inspired by an unlikely\nsource, the string/text compression scheme Lempel-Ziv-Welch, but tailored to\nMGD in a way that preserves tuple boundaries within mini-batches. We then\npresent a suite of novel compressed matrix operation execution techniques\ntailored to the TOC compression scheme that operate directly over the\ncompressed data representation and avoid decompression overheads. An extensive\nempirical evaluation with real-world datasets shows that TOC consistently\nachieves substantial compression ratios by up to 51x and reduces runtimes for\nMGD workloads by up to 10.2x in popular ML systems.\n",
"title": "Tuple-oriented Compression for Large-scale Mini-batch Stochastic Gradient Descent"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
15559
| null |
Validated
| null | null |
null |
{
"abstract": " The task of translating between programming languages differs from the\nchallenge of translating natural languages in that programming languages are\ndesigned with a far more rigid set of structural and grammatical rules.\nPrevious work has used a tree-to-tree encoder/decoder model to take advantage\nof the inherent tree structure of programs during translation. Neural decoders,\nhowever, by default do not exploit known grammar rules of the target language.\nIn this paper, we describe a tree decoder that leverages knowledge of a\nlanguage's grammar rules to exclusively generate syntactically correct\nprograms. We find that this grammar-based tree-to-tree model outperforms the\nstate of the art tree-to-tree model in translating between two programming\nlanguages on a previously used synthetic task.\n",
"title": "Program Language Translation Using a Grammar-Driven Tree-to-Tree Model"
}
| null | null | null | null | true | null |
15560
| null |
Default
| null | null |
null |
{
"abstract": " Initializing all elements of an array to a specified value is a basic\noperation that frequently appears in numerous algorithms and programs.\nInitializable arrays are abstract arrays that support initialization as well as\nreading and writing of any element of the array in less than linear time\nproportional to the length of the array. On the word RAM model with $w$ bits\nword size, we propose an in-place algorithm using only 1 extra bit which\nimplements an initializable array of length $N$ each of whose elements can\nstore $\\ell \\in O(w)$ bits value, and supports all operations in constant worst\ncase time. We also show that our algorithm is not only time optimal but also\nspace optimal. Our algorithm significantly improves upon the previous best\nalgorithm [Navarro, CSUR 2014] using $N + \\ell + o(N)$ extra bits supporting\nall operations in constant worst case time.\nMoreover, for a special cast that $\\ell \\ge 2 \\lceil \\log N \\rceil$ and $\\ell\n\\in O(w)$, we also propose an algorithm so that each element of initializable\narray can store $2^\\ell$ normal states and a one optional state, which uses\n$\\ell + \\lceil \\log N \\rceil + 1$ extra bits and supports all operations in\nconstant worst case time.\n",
"title": "In-Place Initializable Arrays"
}
| null | null | null | null | true | null |
15561
| null |
Default
| null | null |
null |
{
"abstract": " We present a simple generative framework for learning to predict previously\nunseen classes, based on estimating class-attribute-gated class-conditional\ndistributions. We model each class-conditional distribution as an exponential\nfamily distribution and the parameters of the distribution of each seen/unseen\nclass are defined as functions of the respective observed class attributes.\nThese functions can be learned using only the seen class data and can be used\nto predict the parameters of the class-conditional distribution of each unseen\nclass. Unlike most existing methods for zero-shot learning that represent\nclasses as fixed embeddings in some vector space, our generative model\nnaturally represents each class as a probability distribution. It is simple to\nimplement and also allows leveraging additional unlabeled data from unseen\nclasses to improve the estimates of their class-conditional distributions using\ntransductive/semi-supervised learning. Moreover, it extends seamlessly to\nfew-shot learning by easily updating these distributions when provided with a\nsmall number of additional labelled examples from unseen classes. Through a\ncomprehensive set of experiments on several benchmark data sets, we demonstrate\nthe efficacy of our framework.\n",
"title": "A Simple Exponential Family Framework for Zero-Shot Learning"
}
| null | null | null | null | true | null |
15562
| null |
Default
| null | null |
null |
{
"abstract": " This paper presents a probabilistic method for capturing non-monotonic\nbehavior under the biphasic dose-response regime observed in many biological\nsystems experiencing different types of stress. The proposed method is based on\nthe rolling-pin method introduced earlier to estimate highly nonlinear and\nnon-monotonic joint probability distributions from continuous domain data. We\nshow that the proposed method outperforms the conventional parametric methods\nin terms of the error (namely RMSE) and it needs fewer parameters to be\nestimated a priori, while offering high flexibility. The application and\nperformance of the proposed method are shown through an example.\n",
"title": "Modeling Hormesis Using a Non-Monotonic Copula Method"
}
| null | null | null | null | true | null |
15563
| null |
Default
| null | null |
null |
{
"abstract": " Mobile-Edge Computing (MEC) is an emerging paradigm that provides a capillary\ndistribution of cloud computing capabilities to the edge of the wireless access\nnetwork, enabling rich services and applications in close proximity to the end\nusers. In this article, a MEC enabled multi-cell wireless network is considered\nwhere each Base Station (BS) is equipped with a MEC server that can assist\nmobile users in executing computation-intensive tasks via task offloading. The\nproblem of Joint Task Offloading and Resource Allocation (JTORA) is studied in\norder to maximize the users' task offloading gains, which is measured by the\nreduction in task completion time and energy consumption. The considered\nproblem is formulated as a Mixed Integer Non-linear Program (MINLP) that\ninvolves jointly optimizing the task offloading decision, uplink transmission\npower of mobile users, and computing resource allocation at the MEC servers.\nDue to the NP-hardness of this problem, solving for optimal solution is\ndifficult and impractical for a large-scale network. To overcome this drawback,\nour approach is to decompose the original problem into (i) a Resource\nAllocation (RA) problem with fixed task offloading decision and (ii) a Task\nOffloading (TO) problem that optimizes the optimal-value function corresponding\nto the RA problem. We address the RA problem using convex and quasi-convex\noptimization techniques, and propose a novel heuristic algorithm to the TO\nproblem that achieves a suboptimal solution in polynomial time. Numerical\nsimulation results show that our algorithm performs closely to the optimal\nsolution and that it significantly improves the users' offloading utility over\ntraditional approaches.\n",
"title": "Joint Task Offloading and Resource Allocation for Multi-Server Mobile-Edge Computing Networks"
}
| null | null | null | null | true | null |
15564
| null |
Default
| null | null |
null |
{
"abstract": " We discuss the three spacetime dimensional $\\mathbb{C}\\mathbb{P}^N$ model and\nspecialize to the $\\mathbb{C}\\mathbb{P}^1$ model. Because of the Hopf map\n$\\pi_3(\\mathbb{C}\\mathbb{P}^1)=\\mathbb{Z}$ one might try to couple the model to\na periodic $\\theta$ parameter. However, we argue that only the values\n$\\theta=0$ and $\\theta=\\pi$ are consistent. For these values the Skyrmions in\nthe model are bosons and fermions respectively, rather than being anyons. We\nalso extend the model by coupling it to a topological quantum field theory,\nsuch that the Skyrmions are anyons. We use techniques from geometry and\ntopology to construct the $\\theta =\\pi $ theory on arbitrary 3-manifolds, and\nuse recent results about invertible field theories to prove that no other\nvalues of $\\theta $ satisfy the necessary locality.\n",
"title": "The Sum Over Topological Sectors and $θ$ in the 2+1-Dimensional $\\mathbb{C}\\mathbb{P}^1$ $σ$-Model"
}
| null | null | null | null | true | null |
15565
| null |
Default
| null | null |
null |
{
"abstract": " We observed 15 of the solar-type binaries within 67 pc of the Sun previously\nobserved by the Robo-AO system in the visible, with the PHARO near-IR camera\nand the PALM-3000 adaptive optics system on the 5 m Hale telescope. The\nphysical status of the binaries is confirmed through common proper motion and\ndetection of orbital motion. In the process we detected a new candidate\ncompanion to HIP 95309. We also resolved the primary of HIP 110626 into a close\nbinary making that system a triple. These detections increase the completeness\nof the multiplicity survey of the solar-type stars within 67 pc of the Sun.\nCombining our observations of HIP 103455 with archival astrometric measurements\nand RV measurements, we are able to compute the first orbit of HIP 103455\nshowing that the binary has a 68 yr period. We place the components on a\ncolor-magnitude diagram and discuss each multiple system individually.\n",
"title": "Continued Kinematic and Photometric Investigations of Hierarchical Solar-Type Multiple Star Systems"
}
| null | null |
[
"Physics"
] | null | true | null |
15566
| null |
Validated
| null | null |
null |
{
"abstract": " For integers $n$ and $k$, the density Hales-Jewett number $c_{n,k}$ is\ndefined as the maximal size of a subset of $[k]^n$ that contains no\ncombinatorial line. We show that for $k \\ge 3$ the density Hales-Jewett number\n$c_{n,k}$ is equal to the maximal size of a cylinder intersection in the\nproblem $Part_{n,k}$ of testing whether $k$ subsets of $[n]$ form a partition.\nIt follows that the communication complexity, in the Number On the Forehead\n(NOF) model, of $Part_{n,k}$, is equal to the minimal size of a partition of\n$[k]^n$ into subsets that do not contain a combinatorial line. Thus, the bound\nin \\cite{chattopadhyay2007languages} on $Part_{n,k}$ using the Hales-Jewett\ntheorem is in fact tight, and the density Hales-Jewett number can be thought of\nas a quantity in communication complexity. This gives a new angle to this well\nstudied quantity.\nAs a simple application we prove a lower bound on $c_{n,k}$, similar to the\nlower bound in \\cite{polymath2010moser} which is roughly $c_{n,k}/k^n \\ge\n\\exp(-O(\\log n)^{1/\\lceil \\log_2 k\\rceil})$. This lower bound follows from a\nprotocol for $Part_{n,k}$. It is interesting to better understand the\ncommunication complexity of $Part_{n,k}$ as this will also lead to the better\nunderstanding of the Hales-Jewett number. The main purpose of this note is to\nmotivate this study.\n",
"title": "A Note on Multiparty Communication Complexity and the Hales-Jewett Theorem"
}
| null | null | null | null | true | null |
15567
| null |
Default
| null | null |
null |
{
"abstract": " This paper explores the entertainment experience and learning experience in\nScrabble. It proposes a new measure from the educational point of view, which\nwe call learning coefficient, based on the balance between the learner's skill\nand the challenge in Scrabble. Scrabble variants, generated using different\nsize of board and dictionary, are analyzed with two measures of game refinement\nand learning coefficient. The results show that 13x13 Scrabble yields the best\nentertainment experience and 15x15 (standard) Scrabble with 4% of original\ndictionary size yields the most effective environment for language learners.\nMoreover, 15x15 Scrabble with 10% of original dictionary size has a good\nbalance between entertainment and learning experience.\n",
"title": "First Results from Using Game Refinement Measure and Learning Coefficient in Scrabble"
}
| null | null | null | null | true | null |
15568
| null |
Default
| null | null |
null |
{
"abstract": " Blind deconvolution is a ubiquitous problem of recovering two unknown signals\nfrom their convolution. Unfortunately, this is an ill-posed problem in general.\nThis paper focuses on the {\\em short and sparse} blind deconvolution problem,\nwhere the one unknown signal is short and the other one is sparsely and\nrandomly supported. This variant captures the structure of the unknown signals\nin several important applications. We assume the short signal to have unit\n$\\ell^2$ norm and cast the blind deconvolution problem as a nonconvex\noptimization problem over the sphere. We demonstrate that (i) in a certain\nregion of the sphere, every local optimum is close to some shift truncation of\nthe ground truth, and (ii) for a generic short signal of length $k$, when the\nsparsity of activation signal $\\theta\\lesssim k^{-2/3}$ and number of\nmeasurements $m\\gtrsim poly(k)$, a simple initialization method together with a\ndescent algorithm which escapes strict saddle points recovers a near shift\ntruncation of the ground truth kernel.\n",
"title": "Structured Local Optima in Sparse Blind Deconvolution"
}
| null | null | null | null | true | null |
15569
| null |
Default
| null | null |
null |
{
"abstract": " We illustrate the potential applications in machine learning of the\nChristoffel function, or more precisely, its empirical counterpart associated\nwith a counting measure uniformly supported on a finite set of points. Firstly,\nwe provide a thresholding scheme which allows to approximate the support of a\nmeasure from a finite subset of its moments with strong asymptotic guaranties.\nSecondly, we provide a consistency result which relates the empirical\nChristoffel function and its population counterpart in the limit of large\nsamples. Finally, we illustrate the relevance of our results on simulated and\nreal world datasets for several applications in statistics and machine\nlearning: (a) density and support estimation from finite samples, (b) outlier\nand novelty detection and (c) affine matching.\n",
"title": "The empirical Christoffel function with applications in data analysis"
}
| null | null | null | null | true | null |
15570
| null |
Default
| null | null |
null |
{
"abstract": " In the paper we investigate power law for PageRank components for the\nBuckley-Osthus model for web graph. We compare different numerical methods for\nPageRank calculation. With the best method we do a lot of numerical\nexperiments. These experiments confirm the hypothesis about power law. At the\nend we discuss real model of web-ranking based on the classical PageRank\napproach.\n",
"title": "Around power law for PageRank components in Buckley-Osthus model of web graph"
}
| null | null | null | null | true | null |
15571
| null |
Default
| null | null |
null |
{
"abstract": " Wholesale electricity market designs in practice do not provide the market\nparticipants with adequate mechanisms to hedge their financial risks. Demanders\nand suppliers will likely face even greater risks with the deepening\npenetration of variable renewable resources like wind and solar. This paper\nexplores the design of a centralized cash-settled call option market to\nmitigate such risks. A cash-settled call option is a financial instrument that\nallows its holder the right to claim a monetary reward equal to the positive\ndifference between the real-time price of an underlying commodity and a\npre-negotiated strike price for an upfront fee. Through an example, we\nillustrate that a bilateral call option can reduce the payment volatility of\nmarket participants. Then, we design a centralized clearing mechanism for call\noptions that generalizes the bilateral trade. We illustrate through an example\nhow the centralized clearing mechanism generalizes the bilateral trade.\nFinally, the effect of risk preference of the market participants, as well as\nsome generalizations are discussed.\n",
"title": "Cash-settled options for wholesale electricity markets"
}
| null | null |
[
"Computer Science",
"Mathematics"
] | null | true | null |
15572
| null |
Validated
| null | null |
null |
{
"abstract": " Fan et al. (2015) recently introduced a remarkable method for increasing\nasymptotic power of tests in high-dimensional testing problems. If applicable\nto a given test, their power enhancement principle leads to an improved test\nthat has the same asymptotic size, uniformly non-inferior asymptotic power, and\nis consistent against a strictly broader range of alternatives than the\ninitially given test. We study under which conditions this method can be\napplied and show the following: In asymptotic regimes where the dimensionality\nof the parameter space is fixed as sample size increases, there often exist\ntests that can not be further improved with the power enhancement principle.\nHowever, when the dimensionality of the parameter space increases sufficiently\nslowly with sample size and a marginal local asymptotic normality (LAN)\ncondition is satisfied, every test with asymptotic size smaller than one can be\nimproved with the power enhancement principle. While the marginal LAN condition\nalone does not allow one to extend the latter statement to all rates at which\nthe dimensionality increases with sample size, we give sufficient conditions\nunder which this is the case.\n",
"title": "Power in High-Dimensional Testing Problems"
}
| null | null | null | null | true | null |
15573
| null |
Default
| null | null |
null |
{
"abstract": " With the goal of making high-resolution forecasts of regional rainfall,\nprecipitation nowcasting has become an important and fundamental technology\nunderlying various public services ranging from rainstorm warnings to flight\nsafety. Recently, the Convolutional LSTM (ConvLSTM) model has been shown to\noutperform traditional optical flow based methods for precipitation nowcasting,\nsuggesting that deep learning models have a huge potential for solving the\nproblem. However, the convolutional recurrence structure in ConvLSTM-based\nmodels is location-invariant while natural motion and transformation (e.g.,\nrotation) are location-variant in general. Furthermore, since\ndeep-learning-based precipitation nowcasting is a newly emerging area, clear\nevaluation protocols have not yet been established. To address these problems,\nwe propose both a new model and a benchmark for precipitation nowcasting.\nSpecifically, we go beyond ConvLSTM and propose the Trajectory GRU (TrajGRU)\nmodel that can actively learn the location-variant structure for recurrent\nconnections. Besides, we provide a benchmark that includes a real-world\nlarge-scale dataset from the Hong Kong Observatory, a new training loss, and a\ncomprehensive evaluation protocol to facilitate future research and gauge the\nstate of the art.\n",
"title": "Deep Learning for Precipitation Nowcasting: A Benchmark and A New Model"
}
| null | null | null | null | true | null |
15574
| null |
Default
| null | null |
null |
{
"abstract": " PAWS is a tool to analyse the behaviour of weighted automata and conditional\ntransition systems. At its core PAWS is based on a generic implementation of\nalgorithms for checking language equivalence in weighted automata and\nbisimulation in conditional transition systems. This architecture allows for\nthe use of arbitrary user-defined semirings. New semirings can be generated\nduring run-time and the user can rely on numerous automatisation techniques to\ncreate new semiring structures for PAWS' algorithms. Basic semirings such as\ndistributive complete lattices and fields of fractions can be defined by\nspecifying few parameters, more exotic semirings can be generated from other\nsemirings or defined from scratch using a built-in semiring generator. In the\nmost general case, users can define new semirings by programming (in C#) the\nbase operations of the semiring and a procedure to solve linear equations and\nuse their newly generated semiring in the analysis tools that PAWS offers.\n",
"title": "PAWS: A Tool for the Analysis of Weighted Systems"
}
| null | null | null | null | true | null |
15575
| null |
Default
| null | null |
null |
{
"abstract": " Measurements of plasma electric fields are essential to the advancement of\nplasma science and applications. Methods for non-invasive in situ measurements\nof plasma fields on sub-millimeter length scales with high sensitivity over a\nlarge field range remain an outstanding challenge. Here, we introduce and\ndemonstrate a new method for plasma electric field measurement that employs\nelectromagnetically induced transparency as a high-resolution quantum-optical\nprobe for the Stark energy level shifts of plasma-embedded Rydberg atoms, which\nserve as highly-sensitive field sensors with a large dynamic range. The method\nis applied in diagnostics of plasmas photo-excited out of a cesium vapor. The\nplasma electric fields are extracted from spatially-resolved measurements of\nfield-induced shape changes and shifts of Rydberg resonances in rubidium tracer\natoms. Measurement capabilities over a range of plasma densities and\ntemperatures are exploited to characterize plasmas in applied magnetic fields\nand to image electric-field distributions in cyclotron-heated plasmas.\n",
"title": "Quantum-optical spectroscopy for plasma electric field measurements and diagnostics"
}
| null | null | null | null | true | null |
15576
| null |
Default
| null | null |
null |
{
"abstract": " R is a popular language and programming environment for data scientists. It\nis increasingly co-packaged with both relational and Hadoop-based data\nplatforms and can often be the most dominant computational component in data\nanalytics pipelines. Recent work has highlighted inefficiencies in executing R\nprograms, both in terms of execution time and memory requirements, which in\npractice limit the size of data that can be analyzed by R. This paper presents\nROSA, a static analysis framework to improve the performance and space\nefficiency of R programs. ROSA analyzes input programs to determine program\nproperties such as reaching definitions, live variables, aliased variables, and\ntypes of variables. These inferred properties enable program transformations\nsuch as C++ code translation, strength reduction, vectorization, code motion,\nin addition to interpretive optimizations such as avoiding redundant object\ncopies and performing in-place evaluations. An empirical evaluation shows\nsubstantial reductions by ROSA in execution time and memory consumption over\nboth CRAN R and Microsoft R Open.\n",
"title": "ROSA: R Optimizations with Static Analysis"
}
| null | null | null | null | true | null |
15577
| null |
Default
| null | null |
null |
{
"abstract": " Humans are routinely asked to evaluate the performance of other individuals,\nseparating success from failure and affecting outcomes from science to\neducation and sports. Yet, in many contexts, the metrics driving the human\nevaluation process remain unclear. Here we analyse a massive dataset capturing\nplayers' evaluations by human judges to explore human perception of performance\nin soccer, the world's most popular sport. We use machine learning to design an\nartificial judge which accurately reproduces human evaluation, allowing us to\ndemonstrate how human observers are biased towards diverse contextual features.\nBy investigating the structure of the artificial judge, we uncover the aspects\nof the players' behavior which attract the attention of human judges,\ndemonstrating that human evaluation is based on a noticeability heuristic where\nonly feature values far from the norm are considered to rate an individual's\nperformance.\n",
"title": "Human Perception of Performance"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
15578
| null |
Validated
| null | null |
null |
{
"abstract": " Path planning is an important problem in robotics. One way to plan a path\nbetween two points $x,y$ within a (not necessarily simply-connected) planar\ndomain $\\Omega$, is to define a non-negative distance function $d(x,y)$ on\n$\\Omega\\times\\Omega$ such that following the (descending) gradient of this\ndistance function traces such a path. This presents two equally important\nchallenges: A mathematical challenge -- to define $d$ such that $d(x,y)$ has a\nsingle minimum for any fixed $y$ (and this is when $x=y$), since a local\nminimum is in effect a \"dead end\", A computational challenge -- to define $d$\nsuch that it may be computed efficiently. In this paper, given a description of\n$\\Omega$, we show how to assign coordinates to each point of $\\Omega$ and\ndefine a family of distance functions between points using these coordinates,\nsuch that both the mathematical and the computational challenges are met. This\nis done using the concepts of \\emph{harmonic measure} and\n\\emph{$f$-divergences}.\nIn practice, path planning is done on a discrete network defined on a finite\nset of \\emph{sites} sampled from $\\Omega$, so any method that works well on the\ncontinuous domain must be adapted so that it still works well on the discrete\ndomain. Given a set of sites sampled from $\\Omega$, we show how to define a\nnetwork connecting these sites such that a \\emph{greedy routing} algorithm\n(which is the discrete equivalent of continuous gradient descent) based on the\ndistance function mentioned above is guaranteed to generate a path in the\nnetwork between any two such sites. In many cases, this network is close to a\n(desirable) planar graph, especially if the set of sites is dense.\n",
"title": "Practical Distance Functions for Path-Planning in Planar Domains"
}
| null | null | null | null | true | null |
15579
| null |
Default
| null | null |
null |
{
"abstract": " A deep-learning inference accelerator is synthesized from a C-language\nsoftware program parallelized with Pthreads. The software implementation uses\nthe well-known producer/consumer model with parallel threads interconnected by\nFIFO queues. The LegUp high-level synthesis (HLS) tool synthesizes threads into\nparallel FPGA hardware, translating software parallelism into spatial\nparallelism. A complete system is generated where convolution, pooling and\npadding are realized in the synthesized accelerator, with remaining tasks\nexecuting on an embedded ARM processor. The accelerator incorporates reduced\nprecision, and a novel approach for zero-weight-skipping in convolution. On a\nmid-sized Intel Arria 10 SoC FPGA, peak performance on VGG-16 is 138 effective\nGOPS.\n",
"title": "FPGA-Based CNN Inference Accelerator Synthesized from Multi-Threaded C Software"
}
| null | null |
[
"Computer Science",
"Statistics"
] | null | true | null |
15580
| null |
Validated
| null | null |
null |
{
"abstract": " The problem of determining those multiplets of forces, or sets of force\nmultiplets, acting at a set of points, such that there exists a truss\nstructure, or wire web, that can support these force multiplets with all the\nelements of the truss or wire web being under tension, is considered. The\ntwo-dimensional problem where the points are at the vertices of a convex\npolygon is essentially solved: each multiplet of forces must be such that the\nnet anticlockwise torque around any vertex of the forces summed over any number\nof consecutive points clockwise past the vertex must be non-negative; and one\ncan find a truss structure that supports under tension, and only supports,\nthose force multiplets in a convex polyhedron of force multiplets that is\ngenerated by a finite number of force multiplets each satisfying the torque\ncondition. Progress is also made on the problem where only a subset of the\npoints are at the vertices of a convex polygon, and the other points are\ninside. In particular, in the case where only one point is inside, an explicit\nprocedure is described for constructing a suitable truss, if one exists. An\nalternative recipe to that provided by Guevara-Vasquez, Milton, and Onofrei\n(2011), based on earlier work of Camar Eddine and Seppecher (2003), is given\nfor constructing a truss structure, with elements under either compression or\ntension, that supports an arbitrary collection of balanced forces at the\nvertices of a convex polygon. Finally some constraints are given on the forces\nthat a three-dimension truss, or wire web, under tension must satisfy.\n",
"title": "The set of forces that ideal trusses, or wire webs, under tension can support"
}
| null | null |
[
"Physics"
] | null | true | null |
15581
| null |
Validated
| null | null |
null |
{
"abstract": " Over the last decade, researchers and engineers have developed a vast body of\nmethodologies and technologies in requirements engineering for self-adaptive\nsystems. Although existing studies have explored various aspects of this topic,\nfew of them have categorized and summarized these areas of research in\nrequire-ments modeling and analysis. This study aims to investigate the\nresearch themes based on the utilized modeling methods and RE activities. We\nconduct a thematic study in the systematic literature review. The results are\nderived by synthesizing the extracted data with statistical methods. This paper\nprovides an updated review of the research literature, enabling researchers and\npractitioners to better understand the research themes in these areas and\nidentify research gaps which need to be further studied.\n",
"title": "A Thematic Study of Requirements Modeling and Analysis for Self-Adaptive Systems"
}
| null | null |
[
"Computer Science"
] | null | true | null |
15582
| null |
Validated
| null | null |
null |
{
"abstract": " Based on independently distributed $X_1 \\sim N_p(\\theta_1, \\sigma^2_1 I_p)$\nand $X_2 \\sim N_p(\\theta_2, \\sigma^2_2 I_p)$, we consider the efficiency of\nvarious predictive density estimators for $Y_1 \\sim N_p(\\theta_1, \\sigma^2_Y\nI_p)$, with the additional information $\\theta_1 - \\theta_2 \\in A$ and known\n$\\sigma^2_1, \\sigma^2_2, \\sigma^2_Y$. We provide improvements on benchmark\npredictive densities such as plug-in, the maximum likelihood, and the minimum\nrisk equivariant predictive densities. Dominance results are obtained for\n$\\alpha-$divergence losses and include Bayesian improvements for reverse\nKullback-Leibler loss, and Kullback-Leibler (KL) loss in the univariate case\n($p=1$). An ensemble of techniques are exploited, including variance expansion\n(for KL loss), point estimation duality, and concave inequalities.\nRepresentations for Bayesian predictive densities, and in particular for\n$\\hat{q}_{\\pi_{U,A}}$ associated with a uniform prior for $\\theta=(\\theta_1,\n\\theta_2)$ truncated to $\\{\\theta \\in \\mathbb{R}^{2p}: \\theta_1 - \\theta_2 \\in\nA \\}$, are established and are used for the Bayesian dominance findings.\nFinally and interestingly, these Bayesian predictive densities also relate to\nskew-normal distributions, as well as new forms of such distributions.\n",
"title": "On predictive density estimation with additional information"
}
| null | null | null | null | true | null |
15583
| null |
Default
| null | null |
null |
{
"abstract": " Interpretation of electroencephalogram (EEG) signals can be complicated by\nobfuscating artifacts. Artifact detection plays an important role in the\nobservation and analysis of EEG signals. Spatial information contained in the\nplacement of the electrodes can be exploited to accurately detect artifacts.\nHowever, when fewer electrodes are used, less spatial information is available,\nmaking it harder to detect artifacts. In this study, we investigate the\nperformance of a deep learning algorithm, CNN-LSTM, on several channel\nconfigurations. Each configuration was designed to minimize the amount of\nspatial information lost compared to a standard 22-channel EEG. Systems using a\nreduced number of channels ranging from 8 to 20 achieved sensitivities between\n33% and 37% with false alarms in the range of [38, 50] per 24 hours. False\nalarms increased dramatically (e.g., over 300 per 24 hours) when the number of\nchannels was further reduced. Baseline performance of a system that used all 22\nchannels was 39% sensitivity with 23 false alarms. Since the 22-channel system\nwas the only system that included referential channels, the rapid increase in\nthe false alarm rate as the number of channels was reduced underscores the\nimportance of retaining referential channels for artifact reduction. This\ncautionary result is important because one of the biggest differences between\nvarious types of EEGs administered is the type of referential channel used.\n",
"title": "Optimizing Channel Selection for Seizure Detection"
}
| null | null | null | null | true | null |
15584
| null |
Default
| null | null |
null |
{
"abstract": " Feigin-Stoyanovsky's type subspaces for affine Lie algebras of type\n$C_\\ell^{(1)}$ have monomial bases with a nice combinatorial description. We\ndescribe bases of whole standard modules in terms of semi-infinite monomials\nobtained as \"a limit of translations\" of bases for Feigin-Stoyanovsky's type\nsubspaces.\n",
"title": "Bases of standard modules for affine Lie algebras of type $C_\\ell^{(1)}$"
}
| null | null | null | null | true | null |
15585
| null |
Default
| null | null |
null |
{
"abstract": " We present two related methods for deriving connectivity-based brain atlases\nfrom individual connectomes. The proposed methods exploit a previously proposed\ndense connectivity representation, termed continuous connectivity, by first\nperforming graph-based hierarchical clustering of individual brains, and\nsubsequently aggregating the individual parcellations into a consensus\nparcellation. The search for consensus minimizes the sum of cluster membership\ndistances, effectively estimating a pseudo-Karcher mean of individual\nparcellations. We assess the quality of our parcellations using (1)\nKullback-Liebler and Jensen-Shannon divergence with respect to the dense\nconnectome representation, (2) inter-hemispheric symmetry, and (3) performance\nof the simplified connectome in a biological sex classification task. We find\nthat the parcellation based-atlas computed using a greedy search at a\nhierarchical depth 3 outperforms all other parcellation-based atlases as well\nas the standard Dessikan-Killiany anatomical atlas in all three assessments.\n",
"title": "Connectivity-Driven Brain Parcellation via Consensus Clustering"
}
| null | null | null | null | true | null |
15586
| null |
Default
| null | null |
null |
{
"abstract": " We present a measurement of baryon acoustic oscillations (BAO) in the\ncross-correlation of quasars with the Ly$\\alpha$-forest flux-transmission at a\nmean redshift $z=2.40$. The measurement uses the complete SDSS-III data sample:\n168,889 forests and 234,367 quasars from the SDSS Data Release DR12. In\naddition to the statistical improvement on our previous study using DR11, we\nhave implemented numerous improvements at the analysis level allowing a more\naccurate measurement of this cross-correlation. We also developed the first\nsimulations of the cross-correlation allowing us to test different aspects of\nour data analysis and to search for potential systematic errors in the\ndetermination of the BAO peak position. We measure the two ratios\n$D_{H}(z=2.40)/r_{d} = 9.01 \\pm 0.36$ and $D_{M}(z=2.40)/r_{d} = 35.7 \\pm 1.7$,\nwhere the errors include marginalization over the non-linear velocity of\nquasars and the metal - quasar cross-correlation contribution, among other\neffects. These results are within $1.8\\sigma$ of the prediction of the\nflat-$\\Lambda$CDM model describing the observed CMB anisotropies. We combine\nthis study with the Ly$\\alpha$-forest auto-correlation function\n[2017A&A...603A..12B], yielding $D_{H}(z=2.40)/r_{d} = 8.94 \\pm 0.22$ and\n$D_{M}(z=2.40)/r_{d} = 36.6 \\pm 1.2$, within $2.3\\sigma$ of the same\nflat-$\\Lambda$CDM model.\n",
"title": "Baryon acoustic oscillations from the complete SDSS-III Ly$α$-quasar cross-correlation function at $z=2.4$"
}
| null | null | null | null | true | null |
15587
| null |
Default
| null | null |
null |
{
"abstract": " We examine the 2008-2016 $\\gamma$-ray and optical light curves of three\nbright BL Lac objects, 0716+714, MRK 421, BL Lac, which exhibit large\nstructured variability. We searched for periodicities by using a fully Bayesian\napproach. For two out of three sources investigated no significant periodic\nvariability was found. In the case of BL Lac we detected a periodicity of ~ 680\ndays. Although the signal related to this is modest, the coincidence of the\nperiods in both gamma and optical bands is indicative of a physical relevance.\nConsidering previous literature results, possibly related $\\gamma$-ray and\noptical periodicities of about one year time scale are proposed in 4 bright\n$\\gamma$-ray blazars out of the 10 examined in detail. Comparing with results\nfrom periodicity search of optical archives of quasars, the presence of\nquasi-periodicities in blazars might be more frequent by a large factor. This\nsuggests the intriguing possibility that the basic conditions for their\nobservability are related to the relativistic jet in the observer direction,\nbut the overall picture remains uncertain.\n",
"title": "Gamma-ray and Optical Oscillations of 0716+714, MRK 421, and BL Lac"
}
| null | null | null | null | true | null |
15588
| null |
Default
| null | null |
null |
{
"abstract": " The ground state of the diatomic molecules in nature is inevitably bonding,\nand its first excited state is antibonding. We demonstrate theoretically that,\nfor a pair of distant adatoms placed buried in three-dimensional-Dirac\nsemimetals, this natural order of the states can be reversed and an antibonding\nground state occurs at the lowest energy of the so-called bound states in the\ncontinuum. We propose an experimental protocol with the use of a scanning\ntunneling microscope tip to visualize the topographic map of the local density\nof states on the surface of the system to reveal the emerging physics.\n",
"title": "Antibonding Ground state of Adatom Molecules in Bulk Dirac Semimetals"
}
| null | null | null | null | true | null |
15589
| null |
Default
| null | null |
null |
{
"abstract": " T2K (Tokai-to-Kamioka) is a long-baseline neutrino experiment in Japan\ndesigned to study various parameters of neutrino oscillations. A near detector\ncomplex (ND280) is located 280~m downstream of the production target and\nmeasures neutrino beam parameters before any oscillations occur. ND280's\nmeasurements are used to predict the number and spectra of neutrinos in the\nSuper-Kamiokande detector at the distance of 295~km. The difference in the\ntarget material between the far (water) and near (scintillator, hydrocarbon)\ndetectors leads to the main non-cancelling systematic uncertainty for the\noscillation analysis. In order to reduce this uncertainty a new\nWAter-Grid-And-SCintillator detector (WAGASCI) has been developed. A magnetized\niron neutrino detector (Baby MIND) will be used to measure momentum and charge\nidentification of the outgoing muons from charged current interactions. The\nBaby MIND modules are composed of magnetized iron plates and long plastic\nscintillator bars read out at the both ends with wavelength shifting fibers and\nsilicon photomultipliers. The front-end electronics board has been developed to\nperform the readout and digitization of the signals from the scintillator bars.\nDetector elements were tested with cosmic rays and in the PS beam at CERN. The\nobtained results are presented in this paper.\n",
"title": "Baby MIND: A magnetized segmented neutrino detector for the WAGASCI experiment"
}
| null | null | null | null | true | null |
15590
| null |
Default
| null | null |
null |
{
"abstract": " For a general class of contractions of a variety X to a base Y, I discuss\nrecent joint work with M. Wemyss defining a noncommutative enhancement of the\nlocus in Y over which the contraction is not an isomorphism, along with\napplications to the derived symmetries of X. This note is based on a talk given\nat the Kinosaki Symposium in 2016.\n",
"title": "Applications of noncommutative deformations"
}
| null | null | null | null | true | null |
15591
| null |
Default
| null | null |
null |
{
"abstract": " We present the first simultaneous photometric and spectroscopic investigation\nof a large set of RR Lyrae variables in a globular cluster. The radial-velocity\ndata presented comprise the largest sample of RVs of RR Lyrae stars ever\nobtained. The target is M3; $BVI_{\\mathrm{C}}$ time-series of 111 and $b$ flux\ndata of further 64 RRab stars, and RV data of 79 RR Lyrae stars are published.\nBlazhko modulation of the light curves of 47 percent of the RRab stars are\ndetected. The mean value of the center-of-mass velocities of RR Lyrae stars is\n$-146.8$ km s$^{-1}$ with 4.52 km s$^{-1}$ standard deviation, which is in good\nagreement with the results obtained for the red giants of the cluster. The\n${\\Phi_{21}}^{\\mathrm RV}$ phase difference of the RV curves of RRab stars is\nfound to be uniformly constant both for the M3 and for Galactic field RRab\nstars; no period or metallicity dependence of the ${\\Phi_{21}}^{\\mathrm RV}$ is\ndetected. The Baade-Wesselink distances of 26 non-Blazhko variables with the\nbest phase-coverage radial-velocity curves are determined; the corresponding\ndistance of the cluster, $10480\\pm210$ pc, agrees with the previous literature\ninformation. A quadratic formula for the $A_{\\mathrm{puls}}-A_V$ relation of\nRRab stars is given, which is valid for both OoI and OoII variables. We also\nshow that the $(V-I)_0$ of RRab stars measured at light minimum is period\ndependent, there is at least 0.1 mag difference between the colours at minimum\nlight of the shortest- and longest-period variables.\n",
"title": "Photometric and radial-velocity time-series of RR Lyrae stars in M3: analysis of single-mode variables"
}
| null | null | null | null | true | null |
15592
| null |
Default
| null | null |
null |
{
"abstract": " We realize scattering states in a lossy and chaotic two-dimensional microwave\ncavity which follow bundles of classical particle trajectories. To generate\nsuch particlelike scattering states we measure the system's transmission matrix\nand apply an adapted Wigner-Smith time-delay formalism to it. The necessary\nshaping of the incident wave is achieved in situ using phase and amplitude\nregulated microwave antennas. Our experimental findings pave the way for\nestablishing spatially confined communication channels that avoid possible\nintruders or obstacles in wave-based communication systems.\n",
"title": "Particlelike scattering states in a microwave cavity"
}
| null | null | null | null | true | null |
15593
| null |
Default
| null | null |
null |
{
"abstract": " We present a simple model of a non-equilibrium self-organizing market where\nasset prices are partially driven by investment decisions of a bounded-rational\nagent. The agent acts in a stochastic market environment driven by various\nexogenous \"alpha\" signals, agent's own actions (via market impact), and noise.\nUnlike traditional agent-based models, our agent aggregates all traders in the\nmarket, rather than being a representative agent. Therefore, it can be\nidentified with a bounded-rational component of the market itself, providing a\nparticular implementation of an Invisible Hand market mechanism. In such\nsetting, market dynamics are modeled as a fictitious self-play of such\nbounded-rational market-agent in its adversarial stochastic environment. As\nrewards obtained by such self-playing market agent are not observed from market\ndata, we formulate and solve a simple model of such market dynamics based on a\nneuroscience-inspired Bounded Rational Information Theoretic Inverse\nReinforcement Learning (BRIT-IRL). This results in effective asset price\ndynamics with a non-linear mean reversion - which in our model is generated\ndynamically, rather than being postulated. We argue that our model can be used\nin a similar way to the Black-Litterman model. In particular, it represents, in\na simple modeling framework, market views of common predictive signals, market\nimpacts and implied optimal dynamic portfolio allocations, and can be used to\nassess values of private signals. Moreover, it allows one to quantify a\n\"market-implied\" optimal investment strategy, along with a measure of market\nrationality. Our approach is numerically light, and can be implemented using\nstandard off-the-shelf software such as TensorFlow.\n",
"title": "Market Self-Learning of Signals, Impact and Optimal Trading: Invisible Hand Inference with Free Energy"
}
| null | null |
[
"Quantitative Finance"
] | null | true | null |
15594
| null |
Validated
| null | null |
null |
{
"abstract": " A theoretical study of the current-driven dynamics of magnetic skyrmions in\ndisordered perpendicularly-magnetized ultrathin films is presented. The\ndisorder is simulated as a granular structure in which the local anisotropy\nvaries randomly from grain to grain. The skyrmion velocity is computed for\ndifferent disorder parameters and ensembles. Similar behavior is seen for\nspin-torques due to in-plane currents and the spin Hall effect, where a pinning\nregime can be identified at low currents with a transition towards the\ndisorder-free case at higher currents, similar to domain wall motion in\ndisordered films. Moreover, a current-dependent skyrmion Hall effect and\nfluctuations in the core radius are found, which result from the interaction\nwith the pinning potential.\n",
"title": "Current-driven skyrmion dynamics in disordered films"
}
| null | null | null | null | true | null |
15595
| null |
Default
| null | null |
null |
{
"abstract": " In this paper we prove a rigidity result for the equality case of the Penrose\ninequality on $3$-dimensional asymptotically flat manifolds with nonnegative\nscalar curvature and corners. Our result also has deep connections with the\nequality cases of Theorem 1 in \\cite{Miao2} and Theorem 1.1 in \\cite{LM}.\n",
"title": "On the Rigidity of Riemannian-Penrose Inequality for Asymptotically Flat 3-manifolds with Corners"
}
| null | null | null | null | true | null |
15596
| null |
Default
| null | null |
null |
{
"abstract": " We introduce an algebra model to study higher order sum rules for orthogonal\npolynomials on the unit circle. We build the relation between the algebra model\nand sum rules, and prove an equivalent expression on the algebra side for the\nsum rules, involving a Hall-Littlewood type polynomial. By this expression, we\nrecover an earlier result by Golinskii and Zlatǒs, and prove a new case -\nhalf of the Lukic conjecture in the case of a single critical point with\narbitrary order.\n",
"title": "An Algebra Model for the Higher Order Sum Rules"
}
| null | null |
[
"Mathematics"
] | null | true | null |
15597
| null |
Validated
| null | null |
null |
{
"abstract": " For the problems of nonparametric hypothesis testing we introduce the notion\nof maxisets and maxispace. We point out the maxisets of $\\chi^2-$tests,\nCramer-von Mises tests, tests generated $\\mathbb{L}_2$- norms of kernel\nestimators and tests generated quadratic forms of estimators of Fourier\ncoefficients. For these tests we show that, if sequence of alternatives having\ngiven rates of convergence to hypothesis is consistent, then each altehrnative\ncan be broken down into the sum of two parts: a function belonging to maxiset\nand orthogonal function. Sequence of functions belonging to maxiset is\nconsistent sequence of alternatives.\nWe point out asymptotically minimax tests if sets of alternatives are maxiset\nwith deleted \"small\" $\\mathbb{L}_2$-balls.\n",
"title": "On maxispaces of nonparametric tests"
}
| null | null | null | null | true | null |
15598
| null |
Default
| null | null |
null |
{
"abstract": " A major goal of unsupervised learning is to discover data representations\nthat are useful for subsequent tasks, without access to supervised labels\nduring training. Typically, this goal is approached by minimizing a surrogate\nobjective, such as the negative log likelihood of a generative model, with the\nhope that representations useful for subsequent tasks will arise incidentally.\nIn this work, we propose instead to directly target a later desired task by\nmeta-learning an unsupervised learning rule, which leads to representations\nuseful for that task. Here, our desired task (meta-objective) is the\nperformance of the representation on semi-supervised classification, and we\nmeta-learn an algorithm -- an unsupervised weight update rule -- that produces\nrepresentations that perform well under this meta-objective. Additionally, we\nconstrain our unsupervised update rule to a be a biologically-motivated,\nneuron-local function, which enables it to generalize to novel neural network\narchitectures. We show that the meta-learned update rule produces useful\nfeatures and sometimes outperforms existing unsupervised learning techniques.\nWe further show that the meta-learned unsupervised update rule generalizes to\ntrain networks with different widths, depths, and nonlinearities. It also\ngeneralizes to train on data with randomly permuted input dimensions and even\ngeneralizes from image datasets to a text task.\n",
"title": "Learning Unsupervised Learning Rules"
}
| null | null | null | null | true | null |
15599
| null |
Default
| null | null |
null |
{
"abstract": " A conformal coating technique with nanocarbon was developed to enhance the\nsurface properties of alumina nanoparticles for bio-applications. The\nultra-thin carbon layer induces new surface properties such as water\ndispersion, cytocompatibility and tuneable surface chemistry, while maintaining\nthe optical properties of the core particle. The possibility of using these\nparticles as agents for DNA sensing was demonstrated in a competitive assay.\nAdditionally, the inherent fluorescence of the core alumina particles provided\na unique platform for localization and monitoring of living organisms, allowing\nsimultaneous cell monitoring and intra-cellular sensing. Nanoparticles were\nable to carry genes to the cells and release them in an environment where\nspecific biomarkers were present.\n",
"title": "Conformal Nanocarbon Coating of Alumina Nanocrystals for Biosensing and Bioimaging"
}
| null | null | null | null | true | null |
15600
| null |
Default
| null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.