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": " Orbifold equivalence is a notion of symmetry that does not rely on group\nactions. Among other applications, it leads to surprising connections between\nhitherto unrelated singularities. While the concept can be defined in a very\ngeneral category-theoretic language, we focus on the most explicit setting in\nterms of matrix factorisations, where orbifold equivalences arise from defects\nwith special properties. Examples are relatively difficult to construct, but we\nuncover some structural features that distinguish orbifold equivalences -- most\nnotably a finite perturbation expansion. We use those properties to devise a\nsearch algorithm, then present some new examples including Arnold\nsingularities.\n",
"title": "Orbifold equivalence: structure and new examples"
} | null | null | null | null | true | null | 20801 | null | Default | null | null |
null | {
"abstract": " The standard content-based attention mechanism typically used in\nsequence-to-sequence models is computationally expensive as it requires the\ncomparison of large encoder and decoder states at each time step. In this work,\nwe propose an alternative attention mechanism based on a fixed size memory\nrepresentation that is more efficient. Our technique predicts a compact set of\nK attention contexts during encoding and lets the decoder compute an efficient\nlookup that does not need to consult the memory. We show that our approach\nperforms on-par with the standard attention mechanism while yielding inference\nspeedups of 20% for real-world translation tasks and more for tasks with longer\nsequences. By visualizing attention scores we demonstrate that our models learn\ndistinct, meaningful alignments.\n",
"title": "Efficient Attention using a Fixed-Size Memory Representation"
} | null | null | null | null | true | null | 20802 | null | Default | null | null |
null | {
"abstract": " The local multiplicities of the Maxwell sets in the spaces of versal\ndeformations of Pham holomorphic function singularities are calculated. A\nsimilar calculation for some other bifurcation sets (generalized Stokes' sets)\ndefined by more complicated relations between the critical values is given.\nAplications to the complexity of algorithms enumerating topologically distinct\nmorsifications of complicated real function singularities are discussed.\n",
"title": "Multiplicities of bifurcation sets of Pham singularities"
} | null | null | [
"Mathematics"
]
| null | true | null | 20803 | null | Validated | null | null |
null | {
"abstract": " The growing field of large-scale time domain astronomy requires methods for\nprobabilistic data analysis that are computationally tractable, even with large\ndatasets. Gaussian Processes are a popular class of models used for this\npurpose but, since the computational cost scales, in general, as the cube of\nthe number of data points, their application has been limited to small\ndatasets. In this paper, we present a novel method for Gaussian Process\nmodeling in one-dimension where the computational requirements scale linearly\nwith the size of the dataset. We demonstrate the method by applying it to\nsimulated and real astronomical time series datasets. These demonstrations are\nexamples of probabilistic inference of stellar rotation periods, asteroseismic\noscillation spectra, and transiting planet parameters. The method exploits\nstructure in the problem when the covariance function is expressed as a mixture\nof complex exponentials, without requiring evenly spaced observations or\nuniform noise. This form of covariance arises naturally when the process is a\nmixture of stochastically-driven damped harmonic oscillators -- providing a\nphysical motivation for and interpretation of this choice -- but we also\ndemonstrate that it can be a useful effective model in some other cases. We\npresent a mathematical description of the method and compare it to existing\nscalable Gaussian Process methods. The method is fast and interpretable, with a\nrange of potential applications within astronomical data analysis and beyond.\nWe provide well-tested and documented open-source implementations of this\nmethod in C++, Python, and Julia.\n",
"title": "Fast and scalable Gaussian process modeling with applications to astronomical time series"
} | null | null | null | null | true | null | 20804 | null | Default | null | null |
null | {
"abstract": " Motivations Short-read accuracy is important for downstream analyses such as\ngenome assembly and hybrid long-read correction. Despite much work on\nshort-read correction, present-day correctors either do not scale well on large\ndata sets or consider reads as mere suites of k-mers, without taking into\naccount their full-length read information. Results We propose a new method to\ncorrect short reads using de Bruijn graphs, and implement it as a tool called\nBcool. As a first st ep, Bcool constructs a compacted de Bruijn graph from the\nreads. This graph is filtered on the basis of k-mer abundance then of unitig\nabundance, thereby removing from most sequencing errors. The cleaned graph is\nthen used as a reference on which the reads are mapped to correct them. We show\nthat this approach yields more accurate reads than k-mer-spectrum correctors\nwhile being scalable to human-size genomic datasets and beyond. Availability\nand Implementation The implementation is open source and available at http:\n//github.com/Malfoy/BCOOL under the Affero GPL license. Contact Antoine\nLimasset [email protected] & Jean-François Flot [email protected] &\nPierre Peterlongo [email protected]\n",
"title": "Toward perfect reads: self-correction of short reads via mapping on de Bruijn graphs"
} | null | null | null | null | true | null | 20805 | null | Default | null | null |
null | {
"abstract": " Patient pain can be detected highly reliably from facial expressions using a\nset of facial muscle-based action units (AUs) defined by the Facial Action\nCoding System (FACS). A key characteristic of facial expression of pain is the\nsimultaneous occurrence of pain-related AU combinations, whose automated\ndetection would be highly beneficial for efficient and practical pain\nmonitoring. Existing general Automated Facial Expression Recognition (AFER)\nsystems prove inadequate when applied specifically for detecting pain as they\neither focus on detecting individual pain-related AUs but not on combinations\nor they seek to bypass AU detection by training a binary pain classifier\ndirectly on pain intensity data but are limited by lack of enough labeled data\nfor satisfactory training. In this paper, we propose a new approach that mimics\nthe strategy of human coders of decoupling pain detection into two consecutive\ntasks: one performed at the individual video-frame level and the other at\nvideo-sequence level. Using state-of-the-art AFER tools to detect single AUs at\nthe frame level, we propose two novel data structures to encode AU combinations\nfrom single AU scores. Two weakly supervised learning frameworks namely\nmultiple instance learning (MIL) and multiple clustered instance learning\n(MCIL) are employed corresponding to each data structure to learn pain from\nvideo sequences. Experimental results show an 87% pain recognition accuracy\nwith 0.94 AUC (Area Under Curve) on the UNBC-McMaster Shoulder Pain Expression\ndataset. Tests on long videos in a lung cancer patient video dataset\ndemonstrates the potential value of the proposed system for pain monitoring in\nclinical settings.\n",
"title": "Learning Pain from Action Unit Combinations: A Weakly Supervised Approach via Multiple Instance Learning"
} | null | null | null | null | true | null | 20806 | null | Default | null | null |
null | {
"abstract": " Tracking humans that are interacting with the other subjects or environment\nremains unsolved in visual tracking, because the visibility of the human of\ninterests in videos is unknown and might vary over time. In particular, it is\nstill difficult for state-of-the-art human trackers to recover complete human\ntrajectories in crowded scenes with frequent human interactions. In this work,\nwe consider the visibility status of a subject as a fluent variable, whose\nchange is mostly attributed to the subject's interaction with the surrounding,\ne.g., crossing behind another object, entering a building, or getting into a\nvehicle, etc. We introduce a Causal And-Or Graph (C-AOG) to represent the\ncausal-effect relations between an object's visibility fluent and its\nactivities, and develop a probabilistic graph model to jointly reason the\nvisibility fluent change (e.g., from visible to invisible) and track humans in\nvideos. We formulate this joint task as an iterative search of a feasible\ncausal graph structure that enables fast search algorithm, e.g., dynamic\nprogramming method. We apply the proposed method on challenging video sequences\nto evaluate its capabilities of estimating visibility fluent changes of\nsubjects and tracking subjects of interests over time. Results with comparisons\ndemonstrate that our method outperforms the alternative trackers and can\nrecover complete trajectories of humans in complicated scenarios with frequent\nhuman interactions.\n",
"title": "A Causal And-Or Graph Model for Visibility Fluent Reasoning in Tracking Interacting Objects"
} | null | null | null | null | true | null | 20807 | null | Default | null | null |
null | {
"abstract": " We introduce an affine generalization of counter automata, and analyze their\nability as well as affine finite automata. Our contributions are as follows. We\nshow that there is a language that can be recognized by exact realtime affine\ncounter automata but by neither 1-way deterministic pushdown automata nor\nrealtime deterministic k-counter automata. We also show that a certain promise\nproblem, which is conjectured not to be solved by two-way quantum finite\nautomata in polynomial time, can be solved by Las Vegas affine finite automata.\nLastly, we show that how a counter helps for affine finite automata by showing\nthat the language MANYTWINS, which is conjectured not to be recognized by\naffine, quantum or classical finite state models in polynomial time, can be\nrecognized by affine counter automata with one-sided bounded-error in realtime.\n",
"title": "Exact Affine Counter Automata"
} | null | null | null | null | true | null | 20808 | null | Default | null | null |
null | {
"abstract": " In this paper, we characterize several lower separation axioms $C_0, C_D$,\n$C_R$, $C_N$, $\\lambda$-space, nested, $S_{YS}$, $S_{YY}$, $S_{YS}$, and\n$S_{\\delta}$ using pre-order. To analyze topological properties of (resp.\ndynamical systems) foliations, we introduce notions of topology (resp.\ndynamical systems) for foliations. Then proper (resp. compact, minimal,\nrecurrent) foliations are characterized by separation axioms. Conversely, lower\nseparation axioms are interpreted into the condition for foliations and several\nrelations of them are described. Moreover, we introduce some notions for\ntopologies from dynamical systems and foliation theory.\n",
"title": "Preorder characterizations of lower separation axioms and their applications to foliations and flows"
} | null | null | null | null | true | null | 20809 | null | Default | null | null |
null | {
"abstract": " We consider row sequences of vector valued Padé-Faber approximants\n(simultaneous Padé-Faber approximants) and prove a Montessus de Ballore\ntype theorem.\n",
"title": "Convergence of row sequences of simultaneous Padé-Faber approximants"
} | null | null | null | null | true | null | 20810 | null | Default | null | null |
null | {
"abstract": " In this work, we present a novel background subtraction system that uses a\ndeep Convolutional Neural Network (CNN) to perform the segmentation. With this\napproach, feature engineering and parameter tuning become unnecessary since the\nnetwork parameters can be learned from data by training a single CNN that can\nhandle various video scenes. Additionally, we propose a new approach to\nestimate background model from video. For the training of the CNN, we employed\nrandomly 5 percent video frames and their ground truth segmentations taken from\nthe Change Detection challenge 2014(CDnet 2014). We also utilized\nspatial-median filtering as the post-processing of the network outputs. Our\nmethod is evaluated with different data-sets, and the network outperforms the\nexisting algorithms with respect to the average ranking over different\nevaluation metrics. Furthermore, due to the network architecture, our CNN is\ncapable of real time processing.\n",
"title": "A Deep Convolutional Neural Network for Background Subtraction"
} | null | null | null | null | true | null | 20811 | null | Default | null | null |
null | {
"abstract": " In this paper, we establish a second main theorem for holomorphic curve\nintersecting hypersurfaces in general position in projective space with level\nof truncation. As an application, we reduce the number hypersurfaces in\nuniqueness problem for holomorphic curve of authors before.\n",
"title": "A second main theorem for holomorphic curve intersecting hypersurfaces"
} | null | null | [
"Mathematics"
]
| null | true | null | 20812 | null | Validated | null | null |
null | {
"abstract": " We prove that for any partially hyperbolic diffeomorphism with one\ndimensional neutral center on a 3-manifold, the center stable and center\nunstable foliations are complete; moreover, each leaf of center stable and\ncenter unstable foliations is a cylinder, a M$\\ddot{o}$bius band or a plane.\nFurther properties of the Bonatti-Parwani-Potrie type of partially hyperbolic\ndiffeomorphisms are studied. Such examples are obtained by composing the time\n$m$-map (for $m>0$ large) of a non-transitive Anosov flow $\\phi_t$ on an\norientable 3-manifold with Dehn twists along some transverse tori, and the\nexamples are partially hyperbolic with one-dimensional neutral center. We prove\nthat the center foliation gives a topologically Anosov flow which is\ntopologically equivalent to $\\phi_t$. We also prove that for the precise\nexample constructed by Bonatti-Parwani-Potrie, the center stable and center\nunstable foliations are robustly complete.\n",
"title": "Partially hyperbolic diffeomorphisms with one-dimensional neutral center on 3-manifolds"
} | null | null | [
"Mathematics"
]
| null | true | null | 20813 | null | Validated | null | null |
null | {
"abstract": " Speech enhancement (SE) aims to reduce noise in speech signals. Most SE\ntechniques focus on addressing audio information only. In this work, inspired\nby multimodal learning, which utilizes data from different modalities, and the\nrecent success of convolutional neural networks (CNNs) in SE, we propose an\naudio-visual deep CNN (AVDCNN) SE model, which incorporates audio and visual\nstreams into a unified network model. In the proposed AVDCNN SE model, audio\nand visual data are first processed using individual CNNs, and then, fused into\na joint network to generate enhanced speech at the output layer. The AVDCNN\nmodel is trained in an end-to-end manner, and parameters are jointly learned\nthrough back-propagation. We evaluate enhanced speech using five objective\ncriteria. Results show that the AVDCNN yields notably better performance,\ncompared with an audio-only CNN-based SE model and two conventional SE\napproaches, confirming the effectiveness of integrating visual information into\nthe SE process.\n",
"title": "Audio-Visual Speech Enhancement based on Multimodal Deep Convolutional Neural Network"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20814 | null | Validated | null | null |
null | {
"abstract": " We reexamine interactions between the dark sectors of cosmology, with a focus\non robust constraints that can be obtained using only mildly nonlinear scales.\nWhile it is well known that couplings between dark matter and dark energy can\nbe constrained to the percent level when including the full range of scales\nprobed by future optical surveys, calibrating matter power spectrum emulators\nto all possible choices of potentials and couplings requires many\ncomputationally expensive n-body simulations. Here we show that lensing and\nclustering of galaxies in combination with the Cosmic Microwave Background\n(CMB) is capable of probing the dark sector coupling to the few percent level\nfor a given class of models, using only linear and quasi-linear Fourier modes.\nThese scales can, in principle, be described by semi-analytical techniques such\nas the effective field theory of large-scale structure.\n",
"title": "Finding structure in the dark: coupled dark energy, weak lensing, and the mildly nonlinear regime"
} | null | null | null | null | true | null | 20815 | null | Default | null | null |
null | {
"abstract": " Existing zero-shot learning (ZSL) models typically learn a projection\nfunction from a feature space to a semantic embedding space (e.g.~attribute\nspace). However, such a projection function is only concerned with predicting\nthe training seen class semantic representation (e.g.~attribute prediction) or\nclassification. When applied to test data, which in the context of ZSL contains\ndifferent (unseen) classes without training data, a ZSL model typically suffers\nfrom the project domain shift problem. In this work, we present a novel\nsolution to ZSL based on learning a Semantic AutoEncoder (SAE). Taking the\nencoder-decoder paradigm, an encoder aims to project a visual feature vector\ninto the semantic space as in the existing ZSL models. However, the decoder\nexerts an additional constraint, that is, the projection/code must be able to\nreconstruct the original visual feature. We show that with this additional\nreconstruction constraint, the learned projection function from the seen\nclasses is able to generalise better to the new unseen classes. Importantly,\nthe encoder and decoder are linear and symmetric which enable us to develop an\nextremely efficient learning algorithm. Extensive experiments on six benchmark\ndatasets demonstrate that the proposed SAE outperforms significantly the\nexisting ZSL models with the additional benefit of lower computational cost.\nFurthermore, when the SAE is applied to supervised clustering problem, it also\nbeats the state-of-the-art.\n",
"title": "Semantic Autoencoder for Zero-Shot Learning"
} | null | null | [
"Computer Science"
]
| null | true | null | 20816 | null | Validated | null | null |
null | {
"abstract": " We methodologically address the problem of Q-value overestimation in deep\nreinforcement learning to handle high-dimensional state spaces efficiently. By\nadapting concepts from information theory, we introduce an intrinsic penalty\nsignal encouraging reduced Q-value estimates. The resultant algorithm\nencompasses a wide range of learning outcomes containing deep Q-networks as a\nspecial case. Different learning outcomes can be demonstrated by tuning a\nLagrange multiplier accordingly. We furthermore propose a novel scheduling\nscheme for this Lagrange multiplier to ensure efficient and robust learning. In\nexperiments on Atari, our algorithm outperforms other algorithms (e.g. deep and\ndouble deep Q-networks) in terms of both game-play performance and sample\ncomplexity. These results remain valid under the recently proposed dueling\narchitecture.\n",
"title": "An Information-Theoretic Optimality Principle for Deep Reinforcement Learning"
} | null | null | null | null | true | null | 20817 | null | Default | null | null |
null | {
"abstract": " With the analysis of noise-induced synchronization of opinion dynamics with\nbounded confidence (BC), a natural and fundamental question is what opinion\nstructures can be synchronized by noise. In the traditional Hegselmann-Krause\n(HK) model, each agent examines the opinion values of all the other ones and\nthen choose neighbors to update its own opinion according to the BC scheme. In\nreality, people are more likely to interchange opinions with only some\nindividuals, resulting in a predetermined local discourse relationship as\nintroduced by the DeGroot model. In this paper, we consider an opinion dynamics\nthat combines the schemes of BC and local discourse topology and investigate\nits synchronization induced by noise. The new model endows the heterogeneous HK\nmodel with a time-varying discourse topology. With the proposed definition of\nnoise-synchronizability, it is shown that the compound noisy model is almost\nsurely noise-synchronizable if and only if the time-varying discourse graph is\nuniformly jointly connected, taking the noise-induced synchronization of the\nclassical heterogeneous HK model as a special case. As a natural implication,\nthe result for the first time builds the equivalence between the connectivity\nof discourse graph and the beneficial effect of noise for opinion consensus.\n",
"title": "Noise-synchronizability of opinion dynamics"
} | null | null | null | null | true | null | 20818 | null | Default | null | null |
null | {
"abstract": " We establish that matter-wave interference at near-resonant ultraviolet\noptical gratings can be used to spatially separate individual conformers of\ncomplex molecules. Our calculations show that the conformational purity of the\nprepared beam can be close to 100% and that all molecules remain in their\nelectronic ground state. The proposed technique is independent of the dipole\nmoment and the spin of the molecule and thus paves the way for\nstructure-sensitive experiments with hydrocarbons and biomolecules, such as\nneurotransmitters and hormones, which evaded conformer-pure isolation so far\n",
"title": "Conformer-selection by matter-wave interference"
} | null | null | null | null | true | null | 20819 | null | Default | null | null |
null | {
"abstract": " Recent works have shown that synthetic parallel data automatically generated\nby translation models can be effective for various neural machine translation\n(NMT) issues. In this study, we build NMT systems using only synthetic parallel\ndata. As an efficient alternative to real parallel data, we also present a new\ntype of synthetic parallel corpus. The proposed pseudo parallel data are\ndistinct from previous works in that ground truth and synthetic examples are\nmixed on both sides of sentence pairs. Experiments on Czech-German and\nFrench-German translations demonstrate the efficacy of the proposed pseudo\nparallel corpus, which shows not only enhanced results for bidirectional\ntranslation tasks but also substantial improvement with the aid of a ground\ntruth real parallel corpus.\n",
"title": "Building a Neural Machine Translation System Using Only Synthetic Parallel Data"
} | null | null | null | null | true | null | 20820 | null | Default | null | null |
null | {
"abstract": " We propose an automatic method to infer high dynamic range illumination from\na single, limited field-of-view, low dynamic range photograph of an indoor\nscene. In contrast to previous work that relies on specialized image capture,\nuser input, and/or simple scene models, we train an end-to-end deep neural\nnetwork that directly regresses a limited field-of-view photo to HDR\nillumination, without strong assumptions on scene geometry, material\nproperties, or lighting. We show that this can be accomplished in a three step\nprocess: 1) we train a robust lighting classifier to automatically annotate the\nlocation of light sources in a large dataset of LDR environment maps, 2) we use\nthese annotations to train a deep neural network that predicts the location of\nlights in a scene from a single limited field-of-view photo, and 3) we\nfine-tune this network using a small dataset of HDR environment maps to predict\nlight intensities. This allows us to automatically recover high-quality HDR\nillumination estimates that significantly outperform previous state-of-the-art\nmethods. Consequently, using our illumination estimates for applications like\n3D object insertion, we can achieve results that are photo-realistic, which is\nvalidated via a perceptual user study.\n",
"title": "Learning to Predict Indoor Illumination from a Single Image"
} | null | null | null | null | true | null | 20821 | null | Default | null | null |
null | {
"abstract": " We prove that the negative infinitesimal generator $L$ of a semigroup of\npositive contractions on $L^\\infty$ has a bounded $H^\\infty(S_\\eta^0)$-calculus\non BMO$(\\sqrt L)$ for any angle $\\eta>\\pi/2$, provided the semigroup satisfies\nBakry-Emry's $\\Gamma_2 $ criterion. Our arguments only rely on the properties\nof the underlying semigroup and works well in the noncommutative setting. A key\ningredient of our argument is a quasi monotone property for the subordinated\nsemigroup $T_{t,\\alpha}=e^{-tL^\\alpha},0<\\alpha<1$, that is proved in the first\nhalf of the article.\n",
"title": "$H^\\infty$-calculus for semigroup generators on BMO"
} | null | null | [
"Mathematics"
]
| null | true | null | 20822 | null | Validated | null | null |
null | {
"abstract": " We present a three player Bayesian game for which there is no epsilon\nequilibria in Borel measurable strategies for small enough epsilon, however\nthere are non-measurable equilibria.\n",
"title": "A Bayesian Game without epsilon equilibria"
} | null | null | null | null | true | null | 20823 | null | Default | null | null |
null | {
"abstract": " Most recent work on interpretability of complex machine learning models has\nfocused on estimating $\\textit{a posteriori}$ explanations for previously\ntrained models around specific predictions. $\\textit{Self-explaining}$ models\nwhere interpretability plays a key role already during learning have received\nmuch less attention. We propose three desiderata for explanations in general --\nexplicitness, faithfulness, and stability -- and show that existing methods do\nnot satisfy them. In response, we design self-explaining models in stages,\nprogressively generalizing linear classifiers to complex yet architecturally\nexplicit models. Faithfulness and stability are enforced via regularization\nspecifically tailored to such models. Experimental results across various\nbenchmark datasets show that our framework offers a promising direction for\nreconciling model complexity and interpretability.\n",
"title": "Towards Robust Interpretability with Self-Explaining Neural Networks"
} | null | null | [
"Statistics"
]
| null | true | null | 20824 | null | Validated | null | null |
null | {
"abstract": " Deep learning (DL) creates impactful advances following a virtuous recipe:\nmodel architecture search, creating large training data sets, and scaling\ncomputation. It is widely believed that growing training sets and models should\nimprove accuracy and result in better products. As DL application domains grow,\nwe would like a deeper understanding of the relationships between training set\nsize, computational scale, and model accuracy improvements to advance the\nstate-of-the-art.\nThis paper presents a large scale empirical characterization of\ngeneralization error and model size growth as training sets grow. We introduce\na methodology for this measurement and test four machine learning domains:\nmachine translation, language modeling, image processing, and speech\nrecognition. Our empirical results show power-law generalization error scaling\nacross a breadth of factors, resulting in power-law exponents---the \"steepness\"\nof the learning curve---yet to be explained by theoretical work. Further, model\nimprovements only shift the error but do not appear to affect the power-law\nexponent. We also show that model size scales sublinearly with data size. These\nscaling relationships have significant implications on deep learning research,\npractice, and systems. They can assist model debugging, setting accuracy\ntargets, and decisions about data set growth. They can also guide computing\nsystem design and underscore the importance of continued computational scaling.\n",
"title": "Deep Learning Scaling is Predictable, Empirically"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20825 | null | Validated | null | null |
null | {
"abstract": " We derive macroscopic dynamics for self-propelled particles in a fluid. The\nstarting point is a coupled Vicsek-Stokes system. The Vicsek model describes\nself-propelled agents interacting through alignment. It provides a\nphenomenological description of hydrodynamic interactions between agents at\nhigh density. Stokes equations describe a low Reynolds number fluid. These two\ndynamics are coupled by the interaction between the agents and the fluid. The\nfluid contributes to rotating the particles through Jeffery's equation.\nParticle self-propulsion induces a force dipole on the fluid. After\ncoarse-graining we obtain a coupled Self-Organised Hydrodynamics (SOH)-Stokes\nsystem. We perform a linear stability analysis for this system which shows that\nboth pullers and pushers have unstable modes. We conclude by providing\nextensions of the Vicsek-Stokes model including short-distance repulsion,\nfinite particle inertia and finite Reynolds number fluid regime.\n",
"title": "Coupled Self-Organized Hydrodynamics and Stokes models for suspensions of active particles"
} | null | null | null | null | true | null | 20826 | null | Default | null | null |
null | {
"abstract": " The spin Hall effect (SHE) is found to be strong in heavy transition metals\n(HM), such as Ta and W, in their amorphous and/or high resistivity form. In\nthis work, we show that by employing a Cu-Ta binary alloy as buffer layer in an\namorphous Cu$_{100-x}$Ta$_{x}$-based magnetic heterostructure with\nperpendicular magnetic anisotropy (PMA), the SHE-induced damping-like\nspin-orbit torque (DL-SOT) efficiency $|\\xi_{DL}|$ can be linearly tuned by\nadjusting the buffer layer resistivity. Current-induced SOT switching can also\nbe achieved in these Cu$_{100-x}$Ta$_{x}$-based magnetic heterostructures, and\nwe find the switching behavior better explained by a SOT-assisted domain wall\npropagation picture. Through systematic studies on Cu$_{100-x}$Ta$_{x}$-based\nsamples with various compositions, we determine the lower bound of spin Hall\nconductivity\n$|\\sigma_{SH}|\\approx2.02\\times10^{4}[\\hbar/2e]\\Omega^{-1}\\cdot\\operatorname{m}^{-1}$\nin the Ta-rich regime. Based on the idea of resistivity tuning, we further\ndemonstrate that $|\\xi_{DL}|$ can be enhanced from 0.087 for pure Ta to 0.152\nby employing a resistive TaN buffer layer.\n",
"title": "Tunable Spin-Orbit Torques in Cu-Ta Binary Alloy Heterostructures"
} | null | null | null | null | true | null | 20827 | null | Default | null | null |
null | {
"abstract": " By considering nests on a given space, we explore order-theoretical and\ntopological properties that are closely related to the structure of a nest. In\nparticular, we see how subbases given by two dual nests can be an indicator of\nhow close or far are the properties of the space from the structure of a\nlinearly ordered space. Having in mind that the term interlocking nest is a key\ntool to a general solution of the orderability problem, we give a\ncharacterization of interlocking nest via closed sets in the Alexandroff\ntopology and via lower sets, respectively. We also characterize bounded subsets\nof a given set in terms of nests and, finally, we explore the possibility of\ncharacterizing topological groups via properties of nests. All sections are\nfollowed by a number of open questions, which may give new directions to the\norderability problem.\n",
"title": "On Properties of Nests: Some Answers and Questions"
} | null | null | null | null | true | null | 20828 | null | Default | null | null |
null | {
"abstract": " We introduce differential characters of Drinfeld modules. These are\nfunction-field analogues of Buium's p-adic differential characters of elliptic\ncurves and of Manin's differential characters of elliptic curves in\ndifferential algebra, both of which have had notable Diophantine applications.\nWe determine the structure of the group of differential characters. This shows\nthe existence of a family of interesting differential modular functions on the\nmoduli of Drinfeld modules. It also leads to a canonical $F$-crystal equipped\nwith a map to the de Rham cohomology of the Drinfeld module. This $F$-crystal\nis of a differential-algebraic nature, and the relation to the classical\ncohomological realizations is presently not clear.\n",
"title": "Differential Characters of Drinfeld Modules and de Rham Cohomology"
} | null | null | null | null | true | null | 20829 | null | Default | null | null |
null | {
"abstract": " A strong submeasure on a compact metric space X is a sub-linear and bounded\noperator on the space of continuous functions on X. A strong submeasure is\npositive if it is non-decreasing. By Hahn-Banach theorem, a positive strong\nsubmeasure is the supremum of a non-empty collection of measures whose masses\nare uniformly bounded from above.\nWe give several applications of strong submeasures in various diverse topics,\nthus illustrate the usefulness of this classical but largely overlooked notion.\nThe applications include:\n- Pullback and pushforward of all measures by meromorphic selfmaps of compact\ncomplex varieties.\n- The existence of invariant positive strong submeasures for meromorphic maps\nbetween compact complex varieties, a notion of entropy for such submeasures\n(which coincide with the classical ones in good cases) and a version of the\nVariation Principle.\n- Intersection of every positive closed (1,1) currents on compact Kähler\nmanifolds. Explicit calculations are given for self-intersection of the current\nof integration of some curves $C$ in a compact Kähler surface where the\nself-intersection in cohomology is negative.\nAll of these points are new and have not been previously given in work by\nother authors. In addition, we will apply the same ideas to entropy of\ntranscendental maps of $\\mathbb{C}$ and $\\mathbb{C}^2$.\n",
"title": "Strong submeasures and several applications"
} | null | null | null | null | true | null | 20830 | null | Default | null | null |
null | {
"abstract": " Professional baseball players are increasingly guaranteed expensive long-term\ncontracts, with over 70 deals signed in excess of \\$90 million, mostly in the\nlast decade. These are substantial sums compared to a typical franchise\nvaluation of \\$1-2 billion. Hence, the players to whom a team chooses to give\nsuch a contract can have an enormous impact on both competitiveness and profit.\nDespite this, most published approaches examining career progression in\nbaseball are fairly simplistic. We applied four machine learning algorithms to\nthe problem and soundly improved upon existing approaches, particularly for\nbatting data.\n",
"title": "Understanding Career Progression in Baseball Through Machine Learning"
} | null | null | null | null | true | null | 20831 | null | Default | null | null |
null | {
"abstract": " When deriving a master equation for a multipartite weakly-interacting open\nquantum systems, dissipation is often addressed \\textit{locally} on each\ncomponent, i.e. ignoring the coherent couplings, which are later added `by\nhand'. Although simple, the resulting local master equation (LME) is known to\nbe thermodynamically inconsistent. Otherwise, one may always obtain a\nconsistent \\textit{global} master equation (GME) by working on the energy basis\nof the full interacting Hamiltonian. Here, we consider a two-node `quantum\nwire' connected to two heat baths. The stationary solution of the LME and GME\nare obtained and benchmarked against the exact result. Importantly, in our\nmodel, the validity of the GME is constrained by the underlying secular\napproximation. Whenever this breaks down (for resonant weakly-coupled nodes),\nwe observe that the LME, in spite of being thermodynamically flawed: (a)\npredicts the correct steady state, (b) yields the exact asymptotic heat\ncurrents, and (c) reliably reflects the correlations between the nodes. In\ncontrast, the GME fails at all three tasks. Nonetheless, as the inter-node\ncoupling grows, the LME breaks down whilst the GME becomes correct. Hence, the\nglobal and local approach may be viewed as \\textit{complementary} tools, best\nsuited to different parameter regimes.\n",
"title": "Testing the validity of the local and global GKLS master equations on an exactly solvable model"
} | null | null | null | null | true | null | 20832 | null | Default | null | null |
null | {
"abstract": " This work presents a method for contact state estimation using fuzzy\nclustering to learn contact probability for full, six-dimensional humanoid\ncontacts. The data required for training is solely from proprioceptive sensors\n- endeffector contact wrench sensors and inertial measurement units (IMUs) -\nand the method is completely unsupervised. The resulting cluster means are used\nto efficiently compute the probability of contact in each of the six\nendeffector degrees of freedom (DoFs) independently. This clustering-based\ncontact probability estimator is validated in a kinematics-based base state\nestimator in a simulation environment with realistic added sensor noise for\nlocomotion over rough, low-friction terrain on which the robot is subject to\nfoot slip and rotation. The proposed base state estimator which utilizes these\nsix DoF contact probability estimates is shown to perform considerably better\nthan that which determines kinematic contact constraints purely based on\nmeasured normal force.\n",
"title": "Unsupervised Contact Learning for Humanoid Estimation and Control"
} | null | null | null | null | true | null | 20833 | null | Default | null | null |
null | {
"abstract": " Recent literature in the robotics community has focused on learning robot\nbehaviors that abstract out lower-level details of robot control. To fully\nleverage the efficacy of such behaviors, it is necessary to select and sequence\nthem to achieve a given task. In this paper, we present an approach to both\nlearn and sequence robot behaviors, applied to the problem of visual navigation\nof mobile robots. We construct a layered representation of control policies\ncomposed of low- level behaviors and a meta-level policy. The low-level\nbehaviors enable the robot to locomote in a particular environment while\navoiding obstacles, and the meta-level policy actively selects the low-level\nbehavior most appropriate for the current situation based purely on visual\nfeedback. We demonstrate the effectiveness of our method on three simulated\nrobot navigation tasks: a legged hexapod robot which must successfully traverse\nvarying terrain, a wheeled robot which must navigate a maze-like course while\navoiding obstacles, and finally a wheeled robot navigating in the presence of\ndynamic obstacles. We show that by learning control policies in a layered\nmanner, we gain the ability to successfully traverse new compound environments\ncomposed of distinct sub-environments, and outperform both the low-level\nbehaviors in their respective sub-environments, as well as a hand-crafted\nselection of low-level policies on these compound environments.\n",
"title": "Learning to Sequence Robot Behaviors for Visual Navigation"
} | null | null | [
"Computer Science"
]
| null | true | null | 20834 | null | Validated | null | null |
null | {
"abstract": " Evaluation complexity for convexly constrained optimization is considered and\nit is shown first that the complexity bound of $O(\\epsilon^{-3/2})$ proved by\nCartis, Gould and Toint (IMAJNA 32(4) 2012, pp.1662-1695) for computing an\n$\\epsilon$-approximate first-order critical point can be obtained under\nsignificantly weaker assumptions. Moreover, the result is generalized to the\ncase where high-order derivatives are used, resulting in a bound of\n$O(\\epsilon^{-(p+1)/p})$ evaluations whenever derivatives of order $p$ are\navailable. It is also shown that the bound of\n$O(\\epsilon_P^{-1/2}\\epsilon_D^{-3/2})$ evaluations ($\\epsilon_P$ and\n$\\epsilon_D$ being primal and dual accuracy thresholds) suggested by Cartis,\nGould and Toint (SINUM, 2015) for the general nonconvex case involving both\nequality and inequality constraints can be generalized to a bound of\n$O(\\epsilon_P^{-1/p}\\epsilon_D^{-(p+1)/p})$ evaluations under similarly\nweakened assumptions. This paper is variant of a companion report (NTR-11-2015,\nUniversity of Namur, Belgium) which uses a different first-order criticality\nmeasure to obtain the same complexity bounds.\n",
"title": "Evaluation complexity bounds for smooth constrained nonlinear optimisation using scaled KKT conditions, high-order models and the criticality measure $χ$"
} | null | null | null | null | true | null | 20835 | null | Default | null | null |
null | {
"abstract": " In this paper, we prove some one level density results for the low-lying\nzeros of famliies of quadratic and quartic Hecke $L$-functions of the Gaussian\nfield. As corollaries, we deduce that, respectively, at least $94.27 \\%$ and\n$5\\%$ of the members of the quadratic family and the quartic family do not\nvanish at the central point.\n",
"title": "One level density of low-lying zeros of quadratic and quartic Hecke $L$-functions"
} | null | null | null | null | true | null | 20836 | null | Default | null | null |
null | {
"abstract": " Existing approaches to online convex optimization (OCO) make sequential\none-slot-ahead decisions, which lead to (possibly adversarial) losses that\ndrive subsequent decision iterates. Their performance is evaluated by the\nso-called regret that measures the difference of losses between the online\nsolution and the best yet fixed overall solution in hindsight. The present\npaper deals with online convex optimization involving adversarial loss\nfunctions and adversarial constraints, where the constraints are revealed after\nmaking decisions, and can be tolerable to instantaneous violations but must be\nsatisfied in the long term. Performance of an online algorithm in this setting\nis assessed by: i) the difference of its losses relative to the best dynamic\nsolution with one-slot-ahead information of the loss function and the\nconstraint (that is here termed dynamic regret); and, ii) the accumulated\namount of constraint violations (that is here termed dynamic fit). In this\ncontext, a modified online saddle-point (MOSP) scheme is developed, and proved\nto simultaneously yield sub-linear dynamic regret and fit, provided that the\naccumulated variations of per-slot minimizers and constraints are sub-linearly\ngrowing with time. MOSP is also applied to the dynamic network resource\nallocation task, and it is compared with the well-known stochastic dual\ngradient method. Under various scenarios, numerical experiments demonstrate the\nperformance gain of MOSP relative to the state-of-the-art.\n",
"title": "An Online Convex Optimization Approach to Dynamic Network Resource Allocation"
} | null | null | [
"Computer Science",
"Mathematics",
"Statistics"
]
| null | true | null | 20837 | null | Validated | null | null |
null | {
"abstract": " Causal discovery from empirical data is a fundamental problem in many\nscientific domains. Observational data allows for identifiability only up to\nMarkov equivalence class. In this paper we first propose a polynomial time\nalgorithm for learning the exact correctly-oriented structure of the transitive\nreduction of any causal Bayesian networks with high probability, by using\ninterventional path queries. Each path query takes as input an origin node and\na target node, and answers whether there is a directed path from the origin to\nthe target. This is done by intervening the origin node and observing samples\nfrom the target node. We theoretically show the logarithmic sample complexity\nfor the size of interventional data per path query, for continuous and discrete\nnetworks. We further extend our work to learn the transitive edges using\nlogarithmic sample complexity (albeit in time exponential in the maximum number\nof parents for discrete networks). This allows us to learn the full network. We\nalso provide an analysis of imperfect interventions.\n",
"title": "Learning causal Bayes networks using interventional path queries in polynomial time and sample complexity"
} | null | null | null | null | true | null | 20838 | null | Default | null | null |
null | {
"abstract": " We investigate the Fredholm alternative for the $p$-Laplacian in an exterior\ndomain which is the complement of the closed unit ball in $\\mathbb{R}^N$\n($N\\geq 2$). By employing techniques of Calculus of Variations we obtain the\nmultiplicity of solutions. The striking difference between our case and the\nentire space case is also discussed.\n",
"title": "The Fredholm alternative for the $p$-Laplacian in exterior domains"
} | null | null | null | null | true | null | 20839 | null | Default | null | null |
null | {
"abstract": " Precipitation hardening, which relies on a high density of intermetallic\nprecipitates, is a commonly utilized technique for strengthening structural\nalloys. Structural alloys are commonly strengthened through a high density of\nsmall size intermetallic precipitates. At high temperatures, however, the\nprecipitates coarsen to reduce the excess energy of the interface, resulting in\na significant reduction in the strengthening provided by the precipitates. In\ncertain ternary alloys, the secondary solute segregates to the interface and\nresults in the formation of a high density of nanosize precipitates that\nprovide enhanced strength and are resistant to coarsening. To understand the\nchemical effects involved, and to identify such systems, we develop a\nthermodynamic model using the framework of the regular nanocrystalline solution\nmodel. For various global compositions, temperatures and thermodynamic\nparameters, equilibrium configuration of Mg-Sn-Zn alloy is evaluated by\nminimizing the Gibbs free energy function with respect to the region-specific\n(bulk solid-solution, interface and precipitate) concentrations and sizes. The\nresults show that Mg$_2$Sn precipitates can be stabilized to nanoscale sizes\nthrough Zn segregation to Mg/Mg$_2$Sn interface, and the precipitates can be\nstabilized against coarsening at high-temperatures by providing a larger Zn\nconcentration in the system. Together with the inclusion of elastic strain\nenergy effects and the input of computationally informed interface\nthermodynamic parameters in the future, the model is expected to provide a more\nrealistic prediction of segregation and precipitate stabilization in ternary\nalloys of structural importance.\n",
"title": "Thermodynamic Stabilization of Precipitates through Interface Segregation: Chemical Effects"
} | null | null | null | null | true | null | 20840 | null | Default | null | null |
null | {
"abstract": " Multi-band phase variations in principle allow us to infer the longitudinal\ntemperature distributions of planets as a function of height in their\natmospheres. For example, 3.6 micron emission originates from deeper layers of\nthe atmosphere than 4.5 micron due to greater water vapor absorption at the\nlonger wavelength. Since heat transport efficiency increases with pressure, we\nexpect thermal phase curves at 3.6 micron to exhibit smaller amplitudes and\ngreater phase offsets than at 4.5 micron; this trend is not observed. Of the\nseven hot Jupiters with full-orbit phase curves at 3.6 and 4.5 micron, all have\ngreater phase amplitude at 3.6 micron than at 4.5 micron, while four of seven\nexhibit a greater phase offset at 3.6 micron. We use a 3D\nradiative-hydrodynamic model to calculate theoretical phase curves of HD\n189733b, assuming thermo-chemical equilibrium. The model exhibits temperature,\npressure, and wavelength dependent opacity, primarily driven by carbon\nchemistry: CO is energetically favored on the dayside, while CH4 is favored on\nthe cooler nightside. Infrared opacity therefore changes by orders of magnitude\nbetween day and night, producing dramatic vertical shifts in the\nwavelength-specific photospheres, which would complicate eclipse or phase\nmapping with spectral data. The model predicts greater relative phase amplitude\nand greater phase offset at 3.6 micron than at 4.5 micron, in agreement with\nthe data. Our model qualitatively explains the observed phase curves, but is in\ntension with current thermo-chemical kinetics models that predict zonally\nuniform atmospheric composition due to transport of CO from the hot regions of\nthe atmosphere.\n",
"title": "Wavelength Does Not Equal Pressure: Vertical Contribution Functions and their Implications for Mapping Hot Jupiters"
} | null | null | null | null | true | null | 20841 | null | Default | null | null |
null | {
"abstract": " We introduce a new algorithm, called CDER, for supervised machine learning\nthat merges the multi-scale geometric properties of Cover Trees with the\ninformation-theoretic properties of entropy. CDER applies to a training set of\nlabeled pointclouds embedded in a common Euclidean space. If typical\npointclouds corresponding to distinct labels tend to differ at any scale in any\nsub-region, CDER can identify these differences in (typically) linear time,\ncreating a set of distributional coordinates which act as a feature extraction\nmechanism for supervised learning. We describe theoretical properties and\nimplementation details of CDER, and illustrate its benefits on several\nsynthetic examples.\n",
"title": "Supervised Learning of Labeled Pointcloud Differences via Cover-Tree Entropy Reduction"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20842 | null | Validated | null | null |
null | {
"abstract": " Integrable optics is an innovation in particle accelerator design that\nprovides strong nonlinear focusing while avoiding parametric resonances. One\npromising application of integrable optics is to overcome the traditional\nlimits on accelerator intensity imposed by betatron tune-spread and collective\ninstabilities. The efficacy of high-intensity integrable accelerators will be\nundergo comprehensive testing over the next several years at the Fermilab\nIntegrable Optics Test Accelerator (IOTA) and the University of Maryland\nElectron Ring (UMER). We propose an integrable Rapid-Cycling Synchrotron (iRCS)\nas a replacement for the Fermilab Booster to achieve multi-MW beam power for\nthe Fermilab high-energy neutrino program. We provide a overview of the machine\nparameters and discuss an approach to lattice optimization. Integrable optics\nrequires arcs with integer-pi phase advance followed by drifts with matched\nbeta functions. We provide an example integrable lattice with features of a\nmodern RCS - long dispersion-free drifts, low momentum compaction,\nsuperperiodicity, chromaticity correction, separate-function magnets, and\nbounded beta functions.\n",
"title": "Design Considerations for Proposed Fermilab Integrable RCS"
} | null | null | null | null | true | null | 20843 | null | Default | null | null |
null | {
"abstract": " We propose an empirical estimator of the preferential attachment function $f$\nin the setting of general preferential attachment trees. Using a supercritical\ncontinuous-time branching process framework, we prove the almost sure\nconsistency of the proposed estimator. We perform simulations to study the\nempirical properties of our estimators.\n",
"title": "Consistent Estimation in General Sublinear Preferential Attachment Trees"
} | null | null | null | null | true | null | 20844 | null | Default | null | null |
null | {
"abstract": " Learned boundary maps are known to outperform hand- crafted ones as a basis\nfor the watershed algorithm. We show, for the first time, how to train\nwatershed computation jointly with boundary map prediction. The estimator for\nthe merging priorities is cast as a neural network that is con- volutional\n(over space) and recurrent (over iterations). The latter allows learning of\ncomplex shape priors. The method gives the best known seeded segmentation\nresults on the CREMI segmentation challenge.\n",
"title": "Learned Watershed: End-to-End Learning of Seeded Segmentation"
} | null | null | null | null | true | null | 20845 | null | Default | null | null |
null | {
"abstract": " MAGIC (Major Atmospheric Gamma Imaging Cherenkov) is a system of two 17 m\ndiameter, F/1.03 Imaging Atmospheric Cherenkov Telescopes (IACT). They are\ndedicated to the observation of gamma rays from galactic and extragalactic\nsources in the very high energy range (VHE, 30 GeV to 100 TeV). This submission\ncontains links to the proceedings for the 35th International Cosmic Ray\nConference (ICRC2017), held in Bexco, Busan, Korea from the 12th to the 17th of\nJuly, 2017.\n",
"title": "MAGIC Contributions to the 35th International Cosmic Ray Conference (ICRC2017)"
} | null | null | null | null | true | null | 20846 | null | Default | null | null |
null | {
"abstract": " Predicate encryption is a new paradigm of public key encryption that enables\nsearches on encrypted data. Using the predicate encryption, we can search\nkeywords or attributes on encrypted data without decrypting the ciphertexts. In\npredicate encryption, a ciphertext is associated with attributes and a token\ncorresponds to a predicate. The token that corresponds to a predicate $f$ can\ndecrypt the ciphertext associated with attributes $x$ if and only if $f(x)=1$.\nHidden vector encryption (HVE) is a special kind of predicate encryption. In\nthis thesis, we consider the efficiency, the generality, and the security of\nHVE schemes. The results of this thesis are described as follows.\nThe first results of this thesis are efficient HVE schemes where the token\nconsists of just four group elements and the decryption only requires four\nbilinear map computations, independent of the number of attributes in the\nciphertext. The construction uses composite order bilinear groups and is\nselectively secure under the well-known assumptions. The second results are\nefficient HVE schemes that are secure under any kind of pairing types. To\nachieve our goals, we proposed a general framework that converts HVE schemes\nfrom composite order bilinear groups to prime order bilinear groups. Using the\nframework, we convert the previous HVE schemes from composite order bilinear\ngroups to prime order bilinear groups. The third results are fully secure HVE\nschemes with short tokens. Previous HVE schemes were proven to be secure only\nin the selective security model where the capabilities of the adversaries are\nseverely restricted. Using the dual system encryption techniques, we construct\nfully secure HVE schemes with match revealing property in composite order\ngroups.\n",
"title": "Efficient Hidden Vector Encryptions and Its Applications"
} | null | null | null | null | true | null | 20847 | null | Default | null | null |
null | {
"abstract": " Galaxy clustering on small scales is significantly under-predicted by\nsub-halo abundance matching (SHAM) models that populate (sub-)haloes with\ngalaxies based on peak halo mass, $M_{\\rm peak}$. SHAM models based on the peak\nmaximum circular velocity, $V_{\\rm peak}$, have had much better success. The\nprimary reason $M_{\\rm peak}$ based models fail is the relatively low abundance\nof satellite galaxies produced in these models compared to those based on\n$V_{\\rm peak}$. Despite success in predicting clustering, a simple $V_{\\rm\npeak}$ based SHAM model results in predictions for galaxy growth that are at\nodds with observations. We evaluate three possible remedies that could \"save\"\nmass-based SHAM: (1) SHAM models require a significant population of \"orphan\"\ngalaxies as a result of artificial disruption/merging of sub-haloes in modern\nhigh resolution dark matter simulations; (2) satellites must grow significantly\nafter their accretion; and (3) stellar mass is significantly affected by halo\nassembly history. No solution is entirely satisfactory. However, regardless of\nthe particulars, we show that popular SHAM models based on $M_{\\rm peak}$\ncannot be complete physical models as presented. Either $V_{\\rm peak}$ truly is\na better predictor of stellar mass at $z\\sim 0$ and it remains to be seen how\nthe correlation between stellar mass and $V_{\\rm peak}$ comes about, or SHAM\nmodels are missing vital component(s) that significantly affect galaxy\nclustering.\n",
"title": "The Galaxy Clustering Crisis in Abundance Matching"
} | null | null | null | null | true | null | 20848 | null | Default | null | null |
null | {
"abstract": " We consider the hypothesis that dark matter and dark energy consists of\nultra-light self-interacting scalar particles. It is found that the\nKlein-Gordon equation with only two free parameters (mass and self-coupling) on\na Schwarzschild background, at the galactic length-scales has the solution\nwhich corresponds to Bose-Einstein condensate, behaving as dark matter, while\nthe constant solution at supra-galactic scales can explain dark energy.\n",
"title": "Supplying Dark Energy from Scalar Field Dark Matter"
} | null | null | null | null | true | null | 20849 | null | Default | null | null |
null | {
"abstract": " The demand on mobile electronics to continue to shrink in size while increase\nin efficiency drives the demand on the internal passive components to do the\nsame. Power amplifiers require inductors with small form factors, high quality\nfactors, and high operating frequency in the single-digit GHz range. This work\nexplores the use of magnetic materials to satisfy the needs of power amplifier\ninductor applications. This paper discusses the optimization choices regarding\nmaterial selection, device design, and fabrication methodology. The inductors\nachieved here present the best performance to date for an integrated magnetic\ncore inductor at high frequencies with a 1 nH inductance and peak quality\nfactor of 4 at ~3 GHz. Such compact inductors show potential for efficiently\nmeeting the need of mobile electronics in the future.\n",
"title": "GHz-Band Integrated Magnetic Inductors"
} | null | null | [
"Physics"
]
| null | true | null | 20850 | null | Validated | null | null |
null | {
"abstract": " We simplify the construction of projection complexes due to\nBestvina-Bromberg-Fujiwara. To do so, we introduce a sharper version of the\nBehrstock inequality, and show that it can always be enforced. Furthermore, we\nuse the new setup to prove acylindricity results for the action on the\nprojection complexes. We also treat quasi-trees of metric spaces associated to\nprojection complexes, and prove an acylindricity criterion in that context as\nwell.\n",
"title": "Acylindrical actions on projection complexes"
} | null | null | [
"Mathematics"
]
| null | true | null | 20851 | null | Validated | null | null |
null | {
"abstract": " Designing software systems for Geometric Computing applications can be a\nchallenging task. Software engineers typically use software abstractions to\nhide and manage the high complexity of such systems. Without the presence of a\nunifying algebraic system to describe geometric models, the use of software\nabstractions alone can result in many design and maintenance problems.\nGeometric Algebra (GA) can be a universal abstract algebraic language for\nsoftware engineering geometric computing applications. Few sources, however,\nprovide enough information about GA-based software implementations targeting\nthe software engineering community. In particular, successfully introducing GA\nto software engineers requires quite different approaches from introducing GA\nto mathematicians or physicists. This article provides a high-level\nintroduction to the abstract concepts and algebraic representations behind the\nelegant GA mathematical structure. The article focuses on the conceptual and\nrepresentational abstraction levels behind GA mathematics with sufficient\nreferences for more details. In addition, the article strongly recommends\napplying the methods of Computational Thinking in both introducing GA to\nsoftware engineers, and in using GA as a mathematical language for developing\nGeometric Computing software systems.\n",
"title": "Introducing Geometric Algebra to Geometric Computing Software Developers: A Computational Thinking Approach"
} | null | null | null | null | true | null | 20852 | null | Default | null | null |
null | {
"abstract": " In the planted partition problem, the $n$ vertices of a random graph are\npartitioned into $k$ \"clusters,\" and edges between vertices in the same cluster\nand different clusters are included with constant probability $p$ and $q$,\nrespectively (where $0 \\le q < p \\le 1$). We give an efficient spectral\nalgorithm that recovers the clusters with high probability, provided that the\nsizes of any two clusters are either very close or separated by $\\geq\n\\Omega(\\sqrt n)$. We also discuss a generalization of planted partition in\nwhich the algorithm's input is not a random graph, but a random real symmetric\nmatrix with independent above-diagonal entries.\nOur algorithm is an adaptation of a previous algorithm for the uniform case,\ni.e., when all clusters are size $n / k \\geq \\Omega(\\sqrt n)$. The original\nalgorithm recovers the clusters one by one via iterated projection: it\nconstructs the orthogonal projection operator onto the dominant $k$-dimensional\neigenspace of the random graph's adjacency matrix, uses it to recover one of\nthe clusters, then deletes it and recurses on the remaining vertices. We show\nherein that a similar algorithm works in the nonuniform case.\n",
"title": "Recovering Nonuniform Planted Partitions via Iterated Projection"
} | null | null | null | null | true | null | 20853 | null | Default | null | null |
null | {
"abstract": " Business Architecture (BA) plays a significant role in helping organizations\nunderstand enterprise structures and processes, and align them with strategic\nobjectives. However, traditional BAs are represented in fixed structure with\nstatic model elements and fail to dynamically capture business insights based\non internal and external data. To solve this problem, this paper introduces the\ngraph theory into BAs with aim of building extensible data-driven analytics and\nautomatically generating business insights. We use IBM's Component Business\nModel (CBM) as an example to illustrate various ways in which graph theory can\nbe leveraged for data-driven analytics, including what and how business\ninsights can be obtained. Future directions for applying graph theory to\nbusiness architecture analytics are discussed.\n",
"title": "Data-driven Analytics for Business Architectures: Proposed Use of Graph Theory"
} | null | null | null | null | true | null | 20854 | null | Default | null | null |
null | {
"abstract": " The lack of interpretability remains a key barrier to the adoption of deep\nmodels in many applications. In this work, we explicitly regularize deep models\nso human users might step through the process behind their predictions in\nlittle time. Specifically, we train deep time-series models so their\nclass-probability predictions have high accuracy while being closely modeled by\ndecision trees with few nodes. Using intuitive toy examples as well as medical\ntasks for treating sepsis and HIV, we demonstrate that this new tree\nregularization yields models that are easier for humans to simulate than\nsimpler L1 or L2 penalties without sacrificing predictive power.\n",
"title": "Beyond Sparsity: Tree Regularization of Deep Models for Interpretability"
} | null | null | null | null | true | null | 20855 | null | Default | null | null |
null | {
"abstract": " Probability modelling for DNA sequence evolution is well established and\nprovides a rich framework for understanding genetic variation between samples\nof individuals from one or more populations. We show that both classical and\nmore recent models for coalescence (with or without recombination) can be\ndescribed in terms of the so-called phase-type theory, where complicated and\ntedious calculations are circumvented by the use of matrices. The application\nof phase-type theory consists of describing the stochastic model as a Markov\nmodel by appropriately setting up a state space and calculating the\ncorresponding intensity and reward matrices. Formulae of interest are then\nexpressed in terms of these aforementioned matrices. We illustrate this by a\nfew examples calculating the mean, variance and even higher order moments of\nthe site frequency spectrum in the multiple merger coalescent models, and by\nanalysing the mean and variance for the number of segregating sites for\nmultiple samples in the two-locus ancestral recombination graph. We believe\nthat phase-type theory has great potential as a tool for analysing probability\nmodels in population genetics. The compact matrix notation is useful for\nclarification of current models, in particular their formal manipulation\n(calculation), but also for further development or extensions.\n",
"title": "Phase-type distributions in population genetics"
} | null | null | [
"Statistics",
"Quantitative Biology"
]
| null | true | null | 20856 | null | Validated | null | null |
null | {
"abstract": " We consider the problem of robust inference under the important generalized\nlinear model (GLM) with stochastic covariates. We derive the properties of the\nminimum density power divergence estimator of the parameters in GLM with random\ndesign and used this estimator to propose a robust Wald-type test for testing\nany general composite null hypothesis about the GLM. The asymptotic and\nrobustness properties of the proposed test are also examined for the GLM with\nrandom design. Application of the proposed robust inference procedures to the\npopular Poisson regression model for analyzing count data is discussed in\ndetail both theoretically and numerically with some interesting real data\nexamples.\n",
"title": "Robust Wald-type test in GLM with random design based on minimum density power divergence estimators"
} | null | null | [
"Statistics"
]
| null | true | null | 20857 | null | Validated | null | null |
null | {
"abstract": " Convolutional Neural Networks (CNN) and the locally connected layer are\nlimited in capturing the importance and relations of different local receptive\nfields, which are often crucial for tasks such as face verification, visual\nquestion answering, and word sequence prediction. To tackle the issue, we\npropose a novel locally smoothed neural network (LSNN) in this paper. The main\nidea is to represent the weight matrix of the locally connected layer as the\nproduct of the kernel and the smoother, where the kernel is shared over\ndifferent local receptive fields, and the smoother is for determining the\nimportance and relations of different local receptive fields. Specifically, a\nmulti-variate Gaussian function is utilized to generate the smoother, for\nmodeling the location relations among different local receptive fields.\nFurthermore, the content information can also be leveraged by setting the mean\nand precision of the Gaussian function according to the content. Experiments on\nsome variant of MNIST clearly show our advantages over CNN and locally\nconnected layer.\n",
"title": "Locally Smoothed Neural Networks"
} | null | null | null | null | true | null | 20858 | null | Default | null | null |
null | {
"abstract": " This paper presents new results on prediction of linear processes in function\nspaces. The autoregressive Hilbertian process framework of order one (ARH(1)\nprocess framework) is adopted. A componentwise estimator of the autocorrelation\noperator is formulated, from the moment-based estimation of its diagonal\ncoefficients, with respect to the orthogonal eigenvectors of the\nauto-covariance operator, which are assumed to be known. Mean-square\nconvergence to the theoretical autocorrelation operator, in the space of\nHilbert-Schmidt operators, is proved. Consistency then follows in that space.\nFor the associated ARH(1) plug-in predictor, mean absolute convergence to the\ncorresponding conditional expectation, in the considered Hilbert space, is\nobtained. Hence, consistency in that space also holds. A simulation study is\nundertaken to illustrate the finite-large sample behavior of the formulated\ncomponentwise estimator and predictor. The performance of the presented\napproach is compared with alternative approaches in the previous and current\nARH(1) framework literature, including the case of unknown eigenvectors.\n",
"title": "Asymptotic properties of a componentwise ARH(1) plug-in predictor"
} | null | null | null | null | true | null | 20859 | null | Default | null | null |
null | {
"abstract": " Population control policies are proposed and in some places employed as a\nmeans towards curbing population growth. This paper is concerned with a\ndisturbing side-effect of such policies, namely, the potential risk of societal\nfragmentation due to changes in the distribution of family sizes. This effect\nis illustrated in some simple settings and demonstrated by simulation. In\naddition, the dependence of societal fragmentation on family size distribution\nis analyzed. In particular, it is shown that under the studied model, any\npopulation control policy that disallows families of 3 or more children incurs\nthe possible risk of societal fragmentation.\n",
"title": "The Effect of Population Control Policies on Societal Fragmentation"
} | null | null | null | null | true | null | 20860 | null | Default | null | null |
null | {
"abstract": " The Neutralized Drift Compression Experiment-II (NDCX-II) is an induction\nlinac that generates intense pulses of 1.2 MeV helium ions for heating matter\nto extreme conditions. Here, we present recent results on optimizing beam\ntransport. The NDCX-II beamline includes a 1-meter-long drift section\ndownstream of the last transport solenoid, which is filled with\ncharge-neutralizing plasma that enables rapid longitudinal compression of an\nintense ion beam against space-charge forces. The transport section on NDCX-II\nconsists of 28 solenoids. Finding optimal field settings for a group of\nsolenoids requires knowledge of the envelope parameters of the beam. Imaging\nthe beam on scintillator gives the radius of the beam, but the envelope angle\ndr/dz is not measured directly. We demonstrate how the parameters of the beam\nenvelope (r, dr/dz, and emittance) can be reconstructed from a series of images\ntaken at varying B-field strengths of a solenoid upstream of the scintillator.\nWe use this technique to evaluate emittance at several points in the NDCX-II\nbeamline and for optimizing the trajectory of the beam at the entry of the\nplasma-filled drift section.\n",
"title": "Optimizing Beam Transport in Rapidly Compressing Beams on the Neutralized Drift Compression Experiment - II"
} | null | null | null | null | true | null | 20861 | null | Default | null | null |
null | {
"abstract": " Many structured prediction problems (particularly in vision and language\ndomains) are ambiguous, with multiple outputs being correct for an input - e.g.\nthere are many ways of describing an image, multiple ways of translating a\nsentence; however, exhaustively annotating the applicability of all possible\noutputs is intractable due to exponentially large output spaces (e.g. all\nEnglish sentences). In practice, these problems are cast as multi-class\nprediction, with the likelihood of only a sparse set of annotations being\nmaximized - unfortunately penalizing for placing beliefs on plausible but\nunannotated outputs. We make and test the following hypothesis - for a given\ninput, the annotations of its neighbors may serve as an additional supervisory\nsignal. Specifically, we propose an objective that transfers supervision from\nneighboring examples. We first study the properties of our developed method in\na controlled toy setup before reporting results on multi-label classification\nand two image-grounded sequence modeling tasks - captioning and question\ngeneration. We evaluate using standard task-specific metrics and measures of\noutput diversity, finding consistent improvements over standard maximum\nlikelihood training and other baselines.\n",
"title": "Learn from Your Neighbor: Learning Multi-modal Mappings from Sparse Annotations"
} | null | null | null | null | true | null | 20862 | null | Default | null | null |
null | {
"abstract": " Quantitative methods are more familiar to most geophysicists with direct\ninversion or indirect inversion. We will discuss seismic inversion in a high\nlevel sense without getting into the actual algorithms. We will stay with\nmeta-equations and argue pros and cons based on certain mathematical theorems.\n",
"title": "Direct and indirect seismic inversion: interpretation of certain mathematical theorems"
} | null | null | [
"Physics"
]
| null | true | null | 20863 | null | Validated | null | null |
null | {
"abstract": " We define a right Cartan-Eilenberg structure on the category of Kan's\ncombinatorial spectra, and the category of sheaves of such spectra, assuming\nsome conditions. In both structures, we use the geometric concept of homotopy\nequivalence as the strong equivalence. In the case of sheaves, we use local\nequivalence as the weak equivalence. This paper is the first step in a\nlarger-scale program of investigating sheaves of spectra from a geometric\nviewpoint.\n",
"title": "Kan's combinatorial spectra and their sheaves revisited"
} | null | null | null | null | true | null | 20864 | null | Default | null | null |
null | {
"abstract": " Quantum computing for machine learning attracts increasing attention and\nrecent technological developments suggest that especially adiabatic quantum\ncomputing may soon be of practical interest. In this paper, we therefore\nconsider this paradigm and discuss how to adopt it to the problem of binary\nclustering. Numerical simulations demonstrate the feasibility of our approach\nand illustrate how systems of qubits adiabatically evolve towards a solution.\n",
"title": "Adiabatic Quantum Computing for Binary Clustering"
} | null | null | [
"Statistics"
]
| null | true | null | 20865 | null | Validated | null | null |
null | {
"abstract": " This paper provides a new similarity detection algorithm. Given an input set\nof multi-dimensional data points, where each data point is assumed to be\nmulti-dimensional, and an additional reference data point for similarity\nfinding, the algorithm uses kernel method that embeds the data points into a\nlow dimensional manifold. Unlike other kernel methods, which consider the\nentire data for the embedding, our method selects a specific set of kernel\neigenvectors. The eigenvectors are chosen to separate between the data points\nand the reference data point so that similar data points can be easily\nidentified as being distinct from most of the members in the dataset.\n",
"title": "Similarity Search Over Graphs Using Localized Spectral Analysis"
} | null | null | null | null | true | null | 20866 | null | Default | null | null |
null | {
"abstract": " These are reminiscences of my interactions with Julian Schwinger from 1968\nthrough 1981 and beyond.\n",
"title": "Reminiscences of Julian Schwinger: Late Harvard, Early UCLA Years (1968-1981)"
} | null | null | [
"Physics"
]
| null | true | null | 20867 | null | Validated | null | null |
null | {
"abstract": " We demonstrate site-resolved imaging of a strongly correlated quantum system\nwithout relying on laser-cooling techniques during fluorescence imaging. We\nobserved the formation of Mott shells in the insulating regime and realized\nthermometry on the atomic cloud. This work proves the feasibility of the\nnoncooled approach and opens the door to extending the detection technology to\nnew atomic species.\n",
"title": "Site-resolved imaging of a bosonic Mott insulator using ytterbium atoms"
} | null | null | null | null | true | null | 20868 | null | Default | null | null |
null | {
"abstract": " We develop a unified continuum modeling framework for viscous fluids and\nhyperelastic solids using the Gibbs free energy as the thermodynamic potential.\nThis framework naturally leads to a pressure primitive variable formulation for\nthe continuum body, which is well-behaved in both compressible and\nincompressible regimes. Our derivation also provides a rational justification\nof the isochoric-volumetric additive split of free energies in nonlinear\ncontinuum mechanics. The variational multiscale analysis is performed for the\ncontinuum model to construct a foundation for numerical discretization. We\nfirst consider the continuum body instantiated as a hyperelastic material and\ndevelop a variational multiscale formulation for the hyper-elastodynamic\nproblem. The generalized-alpha method is applied for temporal discretization. A\nsegregated algorithm for the nonlinear solver is designed and carefully\nanalyzed. Second, we apply the new formulation to construct a novel unified\nformulation for fluid-solid coupled problems. The variational multiscale\nformulation is utilized for spatial discretization in both fluid and solid\nsubdomains. The generalized-alpha method is applied for the whole continuum\nbody, and optimal high-frequency dissipation is achieved in both fluid and\nsolid subproblems. A new predictor multi-corrector algorithm is developed based\non the segregated algorithm to attain a good balance between robustness and\nefficiency. The efficacy of the new formulations is examined in several\nbenchmark problems. The results indicate that the proposed modeling and\nnumerical methodologies constitute a promising technology for biomedical and\nengineering applications, particularly those necessitating incompressible\nmodels.\n",
"title": "A unified continuum and variational multiscale formulation for fluids, solids, and fluid-structure interaction"
} | null | null | null | null | true | null | 20869 | null | Default | null | null |
null | {
"abstract": " Fast timing capability in X-ray observation of astrophysical objects is one\nof the key properties for the ASTRO-H (Hitomi) mission. Absolute timing\naccuracies of 350 micro second or 35 micro second are required to achieve\nnominal scientific goals or to study fast variabilities of specific sources.\nThe satellite carries a GPS receiver to obtain accurate time information, which\nis distributed from the central onboard computer through the large and complex\nSpaceWire network. The details on the time system on the hardware and software\ndesign are described. In the distribution of the time information, the\npropagation delays and jitters affect the timing accuracy. Six other items\nidentified within the timing system will also contribute to absolute time\nerror. These error items have been measured and checked on ground to ensure the\ntime error budgets meet the mission requirements. The overall timing\nperformance in combination with hardware performance, software algorithm, and\nthe orbital determination accuracies, etc, under nominal conditions satisfies\nthe mission requirements of 35 micro second. This work demonstrates key points\nfor space-use instruments in hardware and software designs and calibration\nmeasurements for fine timing accuracy on the order of microseconds for\nmid-sized satellites using the SpaceWire (IEEE1355) network.\n",
"title": "Time Assignment System and Its Performance aboard the Hitomi Satellite"
} | null | null | [
"Physics"
]
| null | true | null | 20870 | null | Validated | null | null |
null | {
"abstract": " In this article, we use $\\lambda$-sequences to derive common fixed points for\na family of self-mappings defined on a complete $G$-metric space. We imitate\nsome existing techniques in our proofs and show that the tools emlyed can be\nused at a larger scale. These results generalise well known results in the\nliterature.\n",
"title": "Common fixed points via $λ$-sequences in $G$-metric spaces"
} | null | null | null | null | true | null | 20871 | null | Default | null | null |
null | {
"abstract": " Nested weighted automata (NWA) present a robust and convenient\nautomata-theoretic formalism for quantitative specifications. Previous works\nhave considered NWA that processed input words only in the forward direction.\nIt is natural to allow the automata to process input words backwards as well,\nfor example, to measure the maximal or average time between a response and the\npreceding request. We therefore introduce and study bidirectional NWA that can\nprocess input words in both directions. First, we show that bidirectional NWA\ncan express interesting quantitative properties that are not expressible by\nforward-only NWA. Second, for the fundamental decision problems of emptiness\nand universality, we establish decidability and complexity results for the new\nframework which match the best-known results for the special case of\nforward-only NWA. Thus, for NWA, the increased expressiveness of\nbidirectionality is achieved at no additional computational complexity. This is\nin stark contrast to the unweighted case, where bidirectional finite automata\nare no more expressive but exponentially more succinct than their forward-only\ncounterparts.\n",
"title": "Bidirectional Nested Weighted Automata"
} | null | null | null | null | true | null | 20872 | null | Default | null | null |
null | {
"abstract": " We study the regularity of the solutions of second order boundary value\nproblems on manifolds with boundary and bounded geometry. We first show that\nthe regularity property of a given boundary value problem $(P, C)$ is\nequivalent to the uniform regularity of the natural family $(P_x, C_x)$ of\nassociated boundary value problems in local coordinates. We verify that this\nproperty is satisfied for the Dirichlet boundary conditions and strongly\nelliptic operators via a compactness argument. We then introduce a uniform\nShapiro-Lopatinski regularity condition, which is a modification of the\nclassical one, and we prove that it characterizes the boundary value problems\nthat satisfy the usual regularity property. We also show that the natural Robin\nboundary conditions always satisfy the uniform Shapiro-Lopatinski regularity\ncondition, provided that our operator satisfies the strong Legendre condition.\nThis is achieved by proving that \"well-posedness implies regularity\" via a\nmodification of the classical \"Nirenberg trick\". When combining our regularity\nresults with the Poincaré inequality of (Ammann-Grosse-Nistor, preprint\n2015), one obtains the usual well-posedness results for the classical boundary\nvalue problems in the usual scale of Sobolev spaces, thus extending these\nimportant, well-known theorems from smooth, bounded domains, to manifolds with\nboundary and bounded geometry. As we show in several examples, these results do\nnot hold true anymore if one drops the bounded geometry assumption. We also\nintroduce a uniform Agmon condition and show that it is equivalent to the\ncoerciveness. Consequently, we prove a well-posedness result for parabolic\nequations whose elliptic generator satisfies the uniform Agmon condition.\n",
"title": "Uniform Shapiro-Lopatinski conditions and boundary value problems on manifolds with bounded geometry"
} | null | null | null | null | true | null | 20873 | null | Default | null | null |
null | {
"abstract": " A Fourier-Chebyshev spectral method is proposed in this paper for solving the\ncavitation problem in nonlinear elasticity. The interpolation error for the\ncavitation solution is analyzed, the elastic energy error estimate for the\ndiscrete cavitation solution is obtained, and the convergence of the method is\nproved. An algorithm combined a gradient type method with a damped quasi-Newton\nmethod is applied to solve the discretized nonlinear equilibrium equations.\nNumerical experiments show that the Fourier-Chebyshev spectral method is\nefficient and capable of producing accurate numerical cavitation solutions.\n",
"title": "A Fourier-Chebyshev Spectral Method for Cavitation Computation in Nonlinear Elasticity"
} | null | null | null | null | true | null | 20874 | null | Default | null | null |
null | {
"abstract": " Many biological and cognitive systems do not operate deep into one or other\nregime of activity. Instead, they exploit critical surfaces poised at\ntransitions in their parameter space. The pervasiveness of criticality in\nnatural systems suggests that there may be general principles inducing this\nbehaviour. However, there is a lack of conceptual models explaining how\nembodied agents propel themselves towards these critical points. In this paper,\nwe present a learning model driving an embodied Boltzmann Machine towards\ncritical behaviour by maximizing the heat capacity of the network. We test and\ncorroborate the model implementing an embodied agent in the mountain car\nbenchmark, controlled by a Boltzmann Machine that adjust its weights according\nto the model. We find that the neural controller reaches a point of\ncriticality, which coincides with a transition point of the behaviour of the\nagent between two regimes of behaviour, maximizing the synergistic information\nbetween its sensors and the hidden and motor neurons. Finally, we discuss the\npotential of our learning model to study the contribution of criticality to the\nbehaviour of embodied living systems in scenarios not necessarily constrained\nby biological restrictions of the examples of criticality we find in nature.\n",
"title": "Learning Criticality in an Embodied Boltzmann Machine"
} | null | null | [
"Computer Science",
"Physics"
]
| null | true | null | 20875 | null | Validated | null | null |
null | {
"abstract": " The present paper is a companion to the paper by Villone and Rampf (2017),\ntitled \"Hermann Hankel's On the general theory of motion of fluids, an essay\nincluding an English translation of the complete Preisschrift from 1861\"\ntogether with connected documents. Here we give a critical assessment of\nHankel's work, which covers many important aspects of fluid dynamics considered\nfrom a Lagrangian-coordinates point of view: variational formulation in the\nspirit of Hamilton for elastic (barotropic) fluids, transport (we would now say\nLie transport) of vorticity, the Lagrangian significance of Clebsch variables,\netc. Hankel's work is also put in the perspective of previous and future work.\nHence, the action spans about two centuries: from Lagrange's 1760-1761 Turin\npaper on variational approaches to mechanics and fluid mechanics problems to\nArnold's 1966 founding paper on the geometrical/variational formulation of\nincompressible flow. The 22-year old Hankel - who was to die 12 years later -\nemerges as a highly innovative master of mathematical fluid dynamics, fully\ndeserving Riemann's assessment that his Preisschrift contains \"all manner of\ngood things.\"\n",
"title": "A contemporary look at Hermann Hankel's 1861 pioneering work on Lagrangian fluid dynamics"
} | null | null | null | null | true | null | 20876 | null | Default | null | null |
null | {
"abstract": " We propose a dynamic edge exchangeable network model that can capture sparse\nconnections observed in real temporal networks, in contrast to existing models\nwhich are dense. The model achieved superior link prediction accuracy on\nmultiple data sets when compared to a dynamic variant of the blockmodel, and is\nable to extract interpretable time-varying community structures from the data.\nIn addition to sparsity, the model accounts for the effect of social influence\non vertices' future behaviours. Compared to the dynamic blockmodels, our model\nhas a smaller latent space. The compact latent space requires a smaller number\nof parameters to be estimated in variational inference and results in a\ncomputationally friendly inference algorithm.\n",
"title": "A Dynamic Edge Exchangeable Model for Sparse Temporal Networks"
} | null | null | null | null | true | null | 20877 | null | Default | null | null |
null | {
"abstract": " Motivated by advantages of current-mode design, this brief contribution\nexplores the implementation of weight matrices in neuromemristive systems via\ncurrent-mode memristor crossbar circuits. After deriving theoretical results\nfor the range and distribution of weights in the current-mode design, it is\nshown that any weight matrix based on voltage-mode crossbars can be mapped to a\ncurrent-mode crossbar if the voltage-mode weights are carefully bounded. Then,\na modified gradient descent rule is derived for the current-mode design that\ncan be used to perform backpropagation training. Behavioral simulations on the\nMNIST dataset indicate that both voltage and current-mode designs are able to\nachieve similar accuracy and have similar defect tolerance. However, analysis\nof trained weight distributions reveals that current-mode and voltage-mode\ndesigns may use different feature representations.\n",
"title": "Current-mode Memristor Crossbars for Neuromemristive Systems"
} | null | null | null | null | true | null | 20878 | null | Default | null | null |
null | {
"abstract": " We obtain a new bound for incomplete Gauss sums modulo primes. Our argument\nfalls under the framework of Vinogradov's method which we use to reduce the\nproblem under consideration to bounding the number of solutions to two distinct\nsystems of congruences. The first is related to Vinogradov's mean value\ntheorem, although the second does not appear to have been considered before.\nOur bound improves on current results in the range $N\\ge\nq^{2k^{-1/2}+O(k^{-3/2})}$.\n",
"title": "Incomplete Gauss sums modulo primes"
} | null | null | null | null | true | null | 20879 | null | Default | null | null |
null | {
"abstract": " Functional Magnetic Resonance Imaging is a noninvasive tool used to study\nbrain function. Detecting activation is challenged by many factors, and even\nmore so in low-signal scenarios that arise in the performance of high-level\ncognitive tasks. We provide a fully automated and fast adaptive smoothing and\nthresholding (FAST) algorithm that uses smoothing and extreme value theory on\ncorrelated statistical parametric maps for thresholding. Performance on\nexperiments spanning a range of low-signal settings is very encouraging. The\nmethodology also performs well in a study to identify the cerebral regions that\nperceive only-auditory-reliable and only-visual-reliable speech stimuli as well\nas those that perceive one but not the other.\n",
"title": "FAST Adaptive Smoothing and Thresholding for Improved Activation Detection in Low-Signal fMRI"
} | null | null | null | null | true | null | 20880 | null | Default | null | null |
null | {
"abstract": " The layered cuprate Bi$_{2}$CuO$_{4}$ is investigated using magnetic,\ndielectric and pyroelectric measurements. This system is observed to be an\nimproper multiferroic, with a robust ferroelectric state being established near\nthe magnetic transition. Magnetic and dielectric measurements indicate the\npresence of a region above the antiferromagnetic Neel temperature with\nconcomitant polar and magnetic short range order. Bi$_{2}$CuO$_{4}$ is also\nseen to exhibit colossal dielectric constants at higher temperatures with\nclearly distinguishable grain and grain boundary contributions, both of which\nexhibit non-Debye relaxation.\n",
"title": "Improper multiferroicity and colossal dielectric constants in Bi$_{2}$CuO$_{4}$"
} | null | null | null | null | true | null | 20881 | null | Default | null | null |
null | {
"abstract": " In this paper, we study Landau damping in the weakly collisional limit of a\nVlasov-Fokker-Planck equation with nonlinear collisions in the phase-space\n$(x,v) \\in \\mathbb T_x^n \\times \\mathbb R^n_v$. The goal is four-fold: (A) to\nunderstand how collisions suppress plasma echoes and enable Landau damping in\nagreement with linearized theory in Sobolev spaces, (B) to understand how phase\nmixing accelerates collisional relaxation, (C) to understand better how the\nplasma returns to global equilibrium during Landau damping, and (D) to rule out\nthat collision-driven nonlinear instabilities dominate. We give an estimate for\nthe scaling law between Knudsen number and the maximal size of the perturbation\nnecessary for linear theory to be accurate in Sobolev regularity. We conjecture\nthis scaling to be sharp (up to logarithmic corrections) due to potential\nnonlinear echoes in the collisionless model.\n",
"title": "Suppression of plasma echoes and Landau damping in Sobolev spaces by weak collisions in a Vlasov-Fokker-Planck equation"
} | null | null | null | null | true | null | 20882 | null | Default | null | null |
null | {
"abstract": " Deep learning has revolutionised many fields, but it is still challenging to\ntransfer its success to small mobile robots with minimal hardware.\nSpecifically, some work has been done to this effect in the RoboCup humanoid\nfootball domain, but results that are performant and efficient and still\ngenerally applicable outside of this domain are lacking. We propose an approach\nconceptually different from those taken previously. It is based on semantic\nsegmentation and does achieve these desired properties. In detail, it is being\nable to process full VGA images in real-time on a low-power mobile processor.\nIt can further handle multiple image dimensions without retraining, it does not\nrequire specific domain knowledge for achieving a high frame rate and it is\napplicable on a minimal mobile hardware.\n",
"title": "Deep Learning for Semantic Segmentation on Minimal Hardware"
} | null | null | null | null | true | null | 20883 | null | Default | null | null |
null | {
"abstract": " We prove that the generating function of overpartition $M2$-rank differences\nis, up to coefficient signs, a component of the vector-valued mock Eisenstein\nseries attached to a certain quadratic form. We use this to compute analogs of\nthe class number relations for $M2$-rank differences. As applications we split\nthe Kronecker-Hurwitz relation into its \"even\" and \"odd\" parts and calculate\nsums over Hurwitz class numbers of the form $\\sum_{r \\in \\mathbb{Z}} H(n -\n2r^2)$.\n",
"title": "Overpartition $M2$-rank differences, class number relations, and vector-valued mock Eisenstein series"
} | null | null | null | null | true | null | 20884 | null | Default | null | null |
null | {
"abstract": " We study the decentralized machine learning scenario where many users\ncollaborate to learn personalized models based on (i) their local datasets and\n(ii) a similarity graph over the users' learning tasks. Our approach trains\nnonlinear classifiers in a multi-task boosting manner without exchanging\npersonal data and with low communication costs. When background knowledge about\ntask similarities is not available, we propose to jointly learn the\npersonalized models and a sparse collaboration graph through an alternating\noptimization procedure. We analyze the convergence rate, memory consumption and\ncommunication complexity of our decentralized algorithms, and demonstrate the\nbenefits of our approach compared to competing techniques on synthetic and real\ndatasets.\n",
"title": "Communication-Efficient and Decentralized Multi-Task Boosting while Learning the Collaboration Graph"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20885 | null | Validated | null | null |
null | {
"abstract": " We prove existence and uniqueness of strong solutions for a class of\nsecond-order stochastic PDEs with multiplicative Wiener noise and drift of the\nform $\\operatorname{div} \\gamma(\\nabla \\cdot)$, where $\\gamma$ is a maximal\nmonotone graph in $\\mathbb{R}^n \\times \\mathbb{R}^n$ obtained as the\nsubdifferential of a convex function satisfying very mild assumptions on its\nbehavior at infinity. The well-posedness result complements the corresponding\none in our recent work arXiv:1612.08260 where, under the additional assumption\nthat $\\gamma$ is single-valued, a solution with better integrability and\nregularity properties is constructed. The proof given here, however, is\nself-contained.\n",
"title": "On the well-posedness of SPDEs with singular drift in divergence form"
} | null | null | null | null | true | null | 20886 | null | Default | null | null |
null | {
"abstract": " We propose and evaluate new techniques for compressing and speeding up dense\nmatrix multiplications as found in the fully connected and recurrent layers of\nneural networks for embedded large vocabulary continuous speech recognition\n(LVCSR). For compression, we introduce and study a trace norm regularization\ntechnique for training low rank factored versions of matrix multiplications.\nCompared to standard low rank training, we show that our method leads to good\naccuracy versus number of parameter trade-offs and can be used to speed up\ntraining of large models. For speedup, we enable faster inference on ARM\nprocessors through new open sourced kernels optimized for small batch sizes,\nresulting in 3x to 7x speed ups over the widely used gemmlowp library. Beyond\nLVCSR, we expect our techniques and kernels to be more generally applicable to\nembedded neural networks with large fully connected or recurrent layers.\n",
"title": "Trace norm regularization and faster inference for embedded speech recognition RNNs"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20887 | null | Validated | null | null |
null | {
"abstract": " Sparse subspace clustering (SSC) is one of the current state-of-the-art\nmethods for partitioning data points into the union of subspaces, with strong\ntheoretical guarantees. However, it is not practical for large data sets as it\nrequires solving a LASSO problem for each data point, where the number of\nvariables in each LASSO problem is the number of data points. To improve the\nscalability of SSC, we propose to select a few sets of anchor points using a\nrandomized hierarchical clustering method, and, for each set of anchor points,\nsolve the LASSO problems for each data point allowing only anchor points to\nhave a non-zero weight (this reduces drastically the number of variables). This\ngenerates a multilayer graph where each layer corresponds to a different set of\nanchor points. Using the Grassmann manifold of orthogonal matrices, the shared\nconnectivity among the layers is summarized within a single subspace. Finally,\nwe use $k$-means clustering within that subspace to cluster the data points,\nsimilarly as done by spectral clustering in SSC. We show on both synthetic and\nreal-world data sets that the proposed method not only allows SSC to scale to\nlarge-scale data sets, but that it is also much more robust as it performs\nsignificantly better on noisy data and on data with close susbspaces and\noutliers, while it is not prone to oversegmentation.\n",
"title": "Scalable and Robust Sparse Subspace Clustering Using Randomized Clustering and Multilayer Graphs"
} | null | null | [
"Statistics"
]
| null | true | null | 20888 | null | Validated | null | null |
null | {
"abstract": " Health related social media mining is a valuable apparatus for the early\nrecognition of the diverse antagonistic medicinal conditions. Mostly, the\nexisting methods are based on machine learning with knowledge-based learning.\nThis working note presents the Recurrent neural network (RNN) and Long\nshort-term memory (LSTM) based embedding for automatic health text\nclassification in the social media mining. For each task, two systems are built\nand that classify the tweet at the tweet level. RNN and LSTM are used for\nextracting features and non-linear activation function at the last layer\nfacilitates to distinguish the tweets of different categories. The experiments\nare conducted on 2nd Social Media Mining for Health Applications Shared Task at\nAMIA 2017. The experiment results are considerable; however the proposed method\nis appropriate for the health text classification. This is primarily due to the\nreason that, it doesn't rely on any feature engineering mechanisms.\n",
"title": "Deep Health Care Text Classification"
} | null | null | null | null | true | null | 20889 | null | Default | null | null |
null | {
"abstract": " Recently, efforts have been made to improve ptychography phase retrieval\nalgorithms so that they are more robust against noise. Often the algorithm is\nadapted by changing the cost functional that needs to be minimized. In\nparticular, it has been suggested that the cost functional should be obtained\nusing a maximum-likelihood approach that takes the noise statistics into\naccount. Here, we consider the different choices of cost functional, and to how\nthey affect the reconstruction results. We find that seemingly the only\nconsistently reliable way to improve reconstruction results in the presence of\nnoise is to reduce the step size of the update function. In addition, a\nnoise-robust ptychographic reconstruction method has been proposed that relies\non adapting the intensity constraints\n",
"title": "Study of cost functionals for ptychographic phase retrieval to improve the robustness against noise, and a proposal for another noise-robust ptychographic phase retrieval scheme"
} | null | null | null | null | true | null | 20890 | null | Default | null | null |
null | {
"abstract": " A fundamental question in language learning concerns the role of a speaker's\nfirst language in second language acquisition. We present a novel methodology\nfor studying this question: analysis of eye-movement patterns in second\nlanguage reading of free-form text. Using this methodology, we demonstrate for\nthe first time that the native language of English learners can be predicted\nfrom their gaze fixations when reading English. We provide analysis of\nclassifier uncertainty and learned features, which indicates that differences\nin English reading are likely to be rooted in linguistic divergences across\nnative languages. The presented framework complements production studies and\noffers new ground for advancing research on multilingualism.\n",
"title": "Predicting Native Language from Gaze"
} | null | null | [
"Computer Science"
]
| null | true | null | 20891 | null | Validated | null | null |
null | {
"abstract": " We present a form of Schwarz's lemma for holomorphic maps between convex\ndomains $D_1$ and $D_2$. This result provides a lower bound on the distance\nbetween the images of relatively compact subsets of $D_1$ and the boundary of\n$D_2$. This is a natural improvement of an old estimate by Bernal-González\nthat takes into account the geometry of $\\partial{D_1}$. We also provide a new\nestimate for the Kobayashi metric on bounded convex domains.\n",
"title": "A form of Schwarz's lemma and a bound for the Kobayashi metric on convex domains"
} | null | null | [
"Mathematics"
]
| null | true | null | 20892 | null | Validated | null | null |
null | {
"abstract": " In this paper, we focus on the COM-type negative binomial distribution with\nthree parameters, which belongs to COM-type $(a,b,0)$ class distributions and\nfamily of equilibrium distributions of arbitrary birth-death process. Besides,\nwe show abundant distributional properties such as overdispersion and\nunderdispersion, log-concavity, log-convexity (infinite divisibility), pseudo\ncompound Poisson, stochastic ordering and asymptotic approximation. Some\ncharacterizations including sum of equicorrelated geometrically distributed\nrandom variables, conditional distribution, limit distribution of COM-negative\nhypergeometric distribution, and Stein's identity are given for theoretical\nproperties. COM-negative binomial distribution was applied to overdispersion\nand ultrahigh zero-inflated data sets. With the aid of ratio regression, we\nemploy maximum likelihood method to estimate the parameters and the\ngoodness-of-fit are evaluated by the discrete Kolmogorov-Smirnov test.\n",
"title": "The COM-negative binomial distribution: modeling overdispersion and ultrahigh zero-inflated count data"
} | null | null | [
"Mathematics",
"Statistics"
]
| null | true | null | 20893 | null | Validated | null | null |
null | {
"abstract": " Most contextual bandit algorithms minimize regret against the best fixed\npolicy, a questionable benchmark for non-stationary environments that are\nubiquitous in applications. In this work, we develop several efficient\ncontextual bandit algorithms for non-stationary environments by equipping\nexisting methods for i.i.d. problems with sophisticated statistical tests so as\nto dynamically adapt to a change in distribution.\nWe analyze various standard notions of regret suited to non-stationary\nenvironments for these algorithms, including interval regret, switching regret,\nand dynamic regret. When competing with the best policy at each time, one of\nour algorithms achieves regret $\\mathcal{O}(\\sqrt{ST})$ if there are $T$ rounds\nwith $S$ stationary periods, or more generally\n$\\mathcal{O}(\\Delta^{1/3}T^{2/3})$ where $\\Delta$ is some non-stationarity\nmeasure. These results almost match the optimal guarantees achieved by an\ninefficient baseline that is a variant of the classic Exp4 algorithm. The\ndynamic regret result is also the first one for efficient and fully adversarial\ncontextual bandit.\nFurthermore, while the results above require tuning a parameter based on the\nunknown quantity $S$ or $\\Delta$, we also develop a parameter free algorithm\nachieving regret $\\min\\{S^{1/4}T^{3/4}, \\Delta^{1/5}T^{4/5}\\}$. This improves\nand generalizes the best existing result $\\Delta^{0.18}T^{0.82}$ by Karnin and\nAnava (2016) which only holds for the two-armed bandit problem.\n",
"title": "Efficient Contextual Bandits in Non-stationary Worlds"
} | null | null | null | null | true | null | 20894 | null | Default | null | null |
null | {
"abstract": " The search engine is tightly coupled with social networks and is primarily\ndesigned for users to acquire interested information. Specifically, the search\nengine assists the information dissemination for social networks, i.e.,\nenabling users to access interested contents with keywords-searching and\npromoting the process of contents-transferring from the source users directly\nto potential interested users. Accompanying such processes, the social network\nevolves as new links emerge between users with common interests. However, there\nis no clear understanding of such a \"chicken-and-egg\" problem, namely, new\nlinks encourage more social interactions, and vice versa. In this paper, we aim\nto quantitatively characterize the social network evolution phenomenon driven\nby a search engine. First, we propose a search network model for social network\nevolution. Second, we adopt two performance metrics, namely, degree\ndistribution and network diameter. Theoretically, we prove that the degree\ndistribution follows an intensified power-law, and the network diameter\nshrinks. Third, we quantitatively show that the search engine accelerates the\nrumor propagation in social networks. Finally, based on four real-world data\nsets (i.e., CDBLP, Facebook, Weibo Tweets, P2P), we verify our theoretical\nfindings. Furthermore, we find that the search engine dramatically increases\nthe speed of rumor propagation.\n",
"title": "Search Engine Drives the Evolution of Social Networks"
} | null | null | null | null | true | null | 20895 | null | Default | null | null |
null | {
"abstract": " We consider the task of semantic robotic grasping, in which a robot picks up\nan object of a user-specified class using only monocular images. Inspired by\nthe two-stream hypothesis of visual reasoning, we present a semantic grasping\nframework that learns object detection, classification, and grasp planning in\nan end-to-end fashion. A \"ventral stream\" recognizes object class while a\n\"dorsal stream\" simultaneously interprets the geometric relationships necessary\nto execute successful grasps. We leverage the autonomous data collection\ncapabilities of robots to obtain a large self-supervised dataset for training\nthe dorsal stream, and use semi-supervised label propagation to train the\nventral stream with only a modest amount of human supervision. We\nexperimentally show that our approach improves upon grasping systems whose\ncomponents are not learned end-to-end, including a baseline method that uses\nbounding box detection. Furthermore, we show that jointly training our model\nwith auxiliary data consisting of non-semantic grasping data, as well as\nsemantically labeled images without grasp actions, has the potential to\nsubstantially improve semantic grasping performance.\n",
"title": "End-to-End Learning of Semantic Grasping"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 20896 | null | Validated | null | null |
null | {
"abstract": " For any group $G$ and any set $A$, a cellular automaton (CA) is a\ntransformation of the configuration space $A^G$ defined via a finite memory set\nand a local function. Let $\\text{CA}(G;A)$ be the monoid of all CA over $A^G$.\nIn this paper, we investigate a generalisation of the inverse of a CA from the\nsemigroup-theoretic perspective. An element $\\tau \\in \\text{CA}(G;A)$ is von\nNeumann regular (or simply regular) if there exists $\\sigma \\in \\text{CA}(G;A)$\nsuch that $\\tau \\circ \\sigma \\circ \\tau = \\tau$ and $\\sigma \\circ \\tau \\circ\n\\sigma = \\sigma$, where $\\circ$ is the composition of functions. Such an\nelement $\\sigma$ is called a generalised inverse of $\\tau$. The monoid\n$\\text{CA}(G;A)$ itself is regular if all its elements are regular. We\nestablish that $\\text{CA}(G;A)$ is regular if and only if $\\vert G \\vert = 1$\nor $\\vert A \\vert = 1$, and we characterise all regular elements in\n$\\text{CA}(G;A)$ when $G$ and $A$ are both finite. Furthermore, we study\nregular linear CA when $A= V$ is a vector space over a field $\\mathbb{F}$; in\nparticular, we show that every regular linear CA is invertible when $G$ is\ntorsion-free elementary amenable (e.g. when $G=\\mathbb{Z}^d, \\ d \\in\n\\mathbb{N}$) and $V=\\mathbb{F}$, and that every linear CA is regular when $V$\nis finite-dimensional and $G$ is locally finite with $\\text{Char}(\\mathbb{F})\n\\nmid o(g)$ for all $g \\in G$.\n",
"title": "Von Neumann Regular Cellular Automata"
} | null | null | null | null | true | null | 20897 | null | Default | null | null |
null | {
"abstract": " We present a simplified treatment of stability of filtrations on finite\nspaces. Interestingly, we can lift the stability result for combinatorial\nfiltrations from [CSEM06] to the case when two filtrations live on different\nspaces without directly invoking the concept of interleaving. We then prove\nthat this distance is intrinsic by constructing explicit geodesics between any\npair of filtered spaces. Finally we use this construction to obtain a\nstrengthening of the stability result.\n",
"title": "A Distance Between Filtered Spaces Via Tripods"
} | null | null | null | null | true | null | 20898 | null | Default | null | null |
null | {
"abstract": " Existing shape estimation methods for deformable object manipulation suffer\nfrom the drawbacks of being off-line, model dependent, noise-sensitive or\nocclusion-sensitive, and thus are not appropriate for manipulation tasks\nrequiring high precision. In this paper, we present a real-time shape\nestimation approach for autonomous robotic manipulation of 3D deformable\nobjects. Our method fulfills all the requirements necessary for the\nhigh-quality deformable object manipulation in terms of being real-time,\nmodel-free and robust to noise and occlusion. These advantages are accomplished\nusing a joint tracking and reconstruction framework, in which we track the\nobject deformation by aligning a reference shape model with the stream input\nfrom the RGB-D camera, and simultaneously upgrade the reference shape model\naccording to the newly captured RGB-D data. We have evaluated the quality and\nrobustness of our real-time shape estimation pipeline on a set of deformable\nmanipulation tasks implemented on physical robots. Videos are available at\nthis https URL\n",
"title": "Robust Shape Estimation for 3D Deformable Object Manipulation"
} | null | null | null | null | true | null | 20899 | null | Default | null | null |
null | {
"abstract": " In recent years, reinforcement learning (RL) methods have been applied to\nmodel gameplay with great success, achieving super-human performance in various\nenvironments, such as Atari, Go, and Poker. However, those studies mostly focus\non winning the game and have largely ignored the rich and complex human\nmotivations, which are essential for understanding different players' diverse\nbehaviors. In this paper, we present a novel method called Multi-Motivation\nBehavior Modeling (MMBM) that takes the multifaceted human motivations into\nconsideration and models the underlying value structure of the players using\ninverse RL. Our approach does not require the access to the dynamic of the\nsystem, making it feasible to model complex interactive environments such as\nmassively multiplayer online games. MMBM is tested on the World of Warcraft\nAvatar History dataset, which recorded over 70,000 users' gameplay spanning\nthree years period. Our model reveals the significant difference of value\nstructures among different player groups. Using the results of motivation\nmodeling, we also predict and explain their diverse gameplay behaviors and\nprovide a quantitative assessment of how the redesign of the game environment\nimpacts players' behaviors.\n",
"title": "Beyond Winning and Losing: Modeling Human Motivations and Behaviors Using Inverse Reinforcement Learning"
} | null | null | null | null | true | null | 20900 | null | Default | null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.