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": " We have used soft x-ray photoemission electron microscopy to image the\nmagnetization of single domain La$_{0.7}$Sr$_{0.3}$MnO$_{3}$ nano-islands\narranged in geometrically frustrated configurations such as square ice and\nkagome ice geometries. Upon thermal randomization, ensembles of nano-islands\nwith strong inter-island magnetic coupling relax towards low-energy\nconfigurations. Statistical analysis shows that the likelihood of ensembles\nfalling into low-energy configurations depends strongly on the annealing\ntemperature. Annealing to just below the Curie temperature of the ferromagnetic\nfilm (T$_{C}$ = 338 K) allows for a much greater probability of achieving low\nenergy configurations as compared to annealing above the Curie temperature. At\nthis thermally active temperature of 325 K, the ensemble of ferromagnetic\nnano-islands explore their energy landscape over time and eventually transition\nto lower energy states as compared to the frozen-in configurations obtained\nupon cooling from above the Curie temperature. Thus, this materials system\nallows for a facile method to systematically study thermal evolution of\nartificial spin ice arrays of nano-islands at temperatures modestly above room\ntemperature.\n",
"title": "Nanostructured complex oxides as a route towards thermal behavior in artificial spin ice systems"
} | null | null | null | null | true | null | 901 | null | Default | null | null |
null | {
"abstract": " A set function $f$ on a finite set $V$ is submodular if $f(X) + f(Y) \\geq f(X\n\\cup Y) + f(X \\cap Y)$ for any pair $X, Y \\subseteq V$. The symmetric\ndifference transformation (SD-transformation) of $f$ by a canonical set $S\n\\subseteq V$ is a set function $g$ given by $g(X) = f(X \\vartriangle S)$ for $X\n\\subseteq V$,where $X \\vartriangle S = (X \\setminus S) \\cup (S \\setminus X)$\ndenotes the symmetric difference between $X$ and $S$. Submodularity and\nSD-transformations are regarded as the counterparts of convexity and affine\ntransformations in a discrete space, respectively. However, submodularity is\nnot preserved under SD-transformations, in contrast to the fact that convexity\nis invariant under affine transformations. This paper presents a\ncharacterization of SD-stransformations preserving submodularity. Then, we are\nconcerned with the problem of discovering a canonical set $S$, given the\nSD-transformation $g$ of a submodular function $f$ by $S$, provided that $g(X)$\nis given by a function value oracle. A submodular function $f$ on $V$ is said\nto be strict if $f(X) + f(Y) > f(X \\cup Y) + f(X \\cap Y)$ holds whenever both\n$X \\setminus Y$ and $Y \\setminus X$ are nonempty. We show that the problem is\nsolved by using ${\\rm O}(|V|)$ oracle calls when $f$ is strictly submodular,\nalthough it requires exponentially many oracle calls in general.\n",
"title": "Finding Submodularity Hidden in Symmetric Difference"
} | null | null | null | null | true | null | 902 | null | Default | null | null |
null | {
"abstract": " Quick Shift is a popular mode-seeking and clustering algorithm. We present\nfinite sample statistical consistency guarantees for Quick Shift on mode and\ncluster recovery under mild distributional assumptions. We then apply our\nresults to construct a consistent modal regression algorithm.\n",
"title": "On the Consistency of Quick Shift"
} | null | null | null | null | true | null | 903 | null | Default | null | null |
null | {
"abstract": " In recent years Deep Neural Networks (DNNs) have been rapidly developed in\nvarious applications, together with increasingly complex architectures. The\nperformance gain of these DNNs generally comes with high computational costs\nand large memory consumption, which may not be affordable for mobile platforms.\nDeep model quantization can be used for reducing the computation and memory\ncosts of DNNs, and deploying complex DNNs on mobile equipment. In this work, we\npropose an optimization framework for deep model quantization. First, we\npropose a measurement to estimate the effect of parameter quantization errors\nin individual layers on the overall model prediction accuracy. Then, we propose\nan optimization process based on this measurement for finding optimal\nquantization bit-width for each layer. This is the first work that\ntheoretically analyse the relationship between parameter quantization errors of\nindividual layers and model accuracy. Our new quantization algorithm\noutperforms previous quantization optimization methods, and achieves 20-40%\nhigher compression rate compared to equal bit-width quantization at the same\nmodel prediction accuracy.\n",
"title": "Adaptive Quantization for Deep Neural Network"
} | null | null | null | null | true | null | 904 | null | Default | null | null |
null | {
"abstract": " Convolutional Neural Networks (CNNs) have been recently introduced in the\ndomain of session-based next item recommendation. An ordered collection of past\nitems the user has interacted with in a session (or sequence) are embedded into\na 2-dimensional latent matrix, and treated as an image. The convolution and\npooling operations are then applied to the mapped item embeddings. In this\npaper, we first examine the typical session-based CNN recommender and show that\nboth the generative model and network architecture are suboptimal when modeling\nlong-range dependencies in the item sequence. To address the issues, we\nintroduce a simple, but very effective generative model that is capable of\nlearning high-level representation from both short- and long-range item\ndependencies. The network architecture of the proposed model is formed of a\nstack of \\emph{holed} convolutional layers, which can efficiently increase the\nreceptive fields without relying on the pooling operation. Another contribution\nis the effective use of residual block structure in recommender systems, which\ncan ease the optimization for much deeper networks. The proposed generative\nmodel attains state-of-the-art accuracy with less training time in the next\nitem recommendation task. It accordingly can be used as a powerful\nrecommendation baseline to beat in future, especially when there are long\nsequences of user feedback.\n",
"title": "A Simple Convolutional Generative Network for Next Item Recommendation"
} | null | null | null | null | true | null | 905 | null | Default | null | null |
null | {
"abstract": " We present a terahertz spectroscopic study of polar ferrimagnet\nFeZnMo$_3$O$_8$. Our main finding is a giant high-temperature optical diode\neffect, or nonreciprocal directional dichroism, where the transmitted light\nintensity in one direction is over 100 times lower than intensity transmitted\nin the opposite direction. The effect takes place in the paramagnetic phase\nwith no long-range magnetic order in the crystal, which contrasts sharply with\nall existing reports of the terahertz optical diode effect in other\nmagnetoelectric materials, where the long-range magnetic ordering is a\nnecessary prerequisite. In \\fzmo, the effect occurs resonantly with a strong\nmagnetic dipole active transition centered at 1.27 THz and assigned as electron\nspin resonance between the eigenstates of the single-ion anisotropy\nHamiltonian. We propose that the optical diode effect in paramagnetic\nFeZnMo$_3$O$_8$ is driven by signle-ion terms in magnetoelectric free energy.\n",
"title": "High-temperature terahertz optical diode effect without magnetic order in polar FeZnMo$_3$O$_8$"
} | null | null | null | null | true | null | 906 | null | Default | null | null |
null | {
"abstract": " We compare the social character networks of biographical, legendary and\nfictional texts, in search for marks of genre differentiation. We examine the\ndegree distribution of character appearance and find a power law that does not\ndepend on the literary genre or historical content. We also analyze local and\nglobal complex networks measures, in particular, correlation plots between the\nrecently introduced Lobby (or Hirsh $H(1)$) index and Degree, Betweenness and\nCloseness centralities. Assortativity plots, which previous literature claims\nto separate fictional from real social networks, were also studied. We've found\nno relevant differences in the books for these network measures and we give a\nplausible explanation why the previous assortativity result is not correct.\n",
"title": "Character Networks and Book Genre Classification"
} | null | null | [
"Computer Science",
"Physics"
]
| null | true | null | 907 | null | Validated | null | null |
null | {
"abstract": " Motivated by the recent experimental realization of the Haldane model by\nultracold fermions in an optical lattice, we investigate phase diagrams of the\nhard-core Bose-Hubbard model on a honeycomb lattice. This model is closely\nrelated with a spin-1/2 antiferromagnetic (AF) quantum spin model.\nNearest-neighbor (NN) hopping amplitude is positive and it prefers an AF\nconfigurations of phases of Bose-Einstein condensates. On the other hand, an\namplitude of the next-NN hopping depends on an angle variable as in the Haldane\nmodel. Phase diagrams are obtained by means of an extended path-integral\nMonte-Carlo simulations. Besides the AF state, a 120$^o$-order state, there\nappear other phases including a Bose metal in which no long-range orders exist.\n",
"title": "Phase diagrams of Bose-Hubbard model and antiferromagnetic spin-1/2 models on a honeycomb lattice"
} | null | null | null | null | true | null | 908 | null | Default | null | null |
null | {
"abstract": " Exoplanet host star activity, in the form of unocculted star spots or\nfaculae, alters the observed transmission and emission spectra of the\nexoplanet. This effect can be exacerbated when combining data from different\nepochs if the stellar photosphere varies between observations due to activity.\nredHere we present a method to characterize and correct for relative changes\ndue to stellar activity by exploiting multi-epoch ($\\ge$2 visits/transits)\nobservations to place them in a consistent reference frame. Using measurements\nfrom portions of the planet's orbit where negligible planet transmission or\nemission can be assumed, we determine changes to the stellar spectral\namplitude. With the analytical methods described here, we predict the impact of\nstellar variability on transit observations. Supplementing these forecasts with\nKepler-measured stellar variabilities for F-, G-, K-, and M-dwarfs, and\npredicted transit precisions by JWST's NIRISS, NIRCam, and MIRI, we conclude\nthat stellar activity does not impact infrared transiting exoplanet\nobservations of most presently-known or predicted TESS targets by current or\nnear-future platforms, such as JWST.\n",
"title": "Forecasting the Impact of Stellar Activity on Transiting Exoplanet Spectra"
} | null | null | null | null | true | null | 909 | null | Default | null | null |
null | {
"abstract": " Developers of Molecular Dynamics (MD) codes face significant challenges when\nadapting existing simulation packages to new hardware. In a continuously\ndiversifying hardware landscape it becomes increasingly difficult for\nscientists to be experts both in their own domain (physics/chemistry/biology)\nand specialists in the low level parallelisation and optimisation of their\ncodes. To address this challenge, we describe a \"Separation of Concerns\"\napproach for the development of parallel and optimised MD codes: the science\nspecialist writes code at a high abstraction level in a domain specific\nlanguage (DSL), which is then translated into efficient computer code by a\nscientific programmer. In a related context, an abstraction for the solution of\npartial differential equations with grid based methods has recently been\nimplemented in the (Py)OP2 library. Inspired by this approach, we develop a\nPython code generation system for molecular dynamics simulations on different\nparallel architectures, including massively parallel distributed memory systems\nand GPUs. We demonstrate the efficiency of the auto-generated code by studying\nits performance and scalability on different hardware and compare it to other\nstate-of-the-art simulation packages. With growing data volumes the extraction\nof physically meaningful information from the simulation becomes increasingly\nchallenging and requires equally efficient implementations. A particular\nadvantage of our approach is the easy expression of such analysis algorithms.\nWe consider two popular methods for deducing the crystalline structure of a\nmaterial from the local environment of each atom, show how they can be\nexpressed in our abstraction and implement them in the code generation\nframework.\n",
"title": "A Domain Specific Language for Performance Portable Molecular Dynamics Algorithms"
} | null | null | null | null | true | null | 910 | null | Default | null | null |
null | {
"abstract": " We obtain the non-linear generalization of the Sachs-Wolfe + integrated\nSachs-Wolfe (ISW) formula describing the CMB temperature anisotropies. Our\nformula is valid at all orders in perturbation theory, is also valid in all\ngauges and includes scalar, vector and tensor modes. A direct consequence of\nour results is that the maps of the logarithmic temperature anisotropies are\nmuch cleaner than the usual CMB maps, because they automatically remove many\nsecondary anisotropies. This can for instance, facilitate the search for\nprimordial non-Gaussianity in future works. It also disentangles the non-linear\nISW from other effects. Finally, we provide a method which can iteratively be\nused to obtain the lensing solution at the desired order.\n",
"title": "CMB anisotropies at all orders: the non-linear Sachs-Wolfe formula"
} | null | null | [
"Physics"
]
| null | true | null | 911 | null | Validated | null | null |
null | {
"abstract": " We prove sharp decoupling inequalities for a class of two dimensional\nnon-degenerate surfaces in R^5, introduced by Prendiville. As a consequence, we\nobtain sharp bounds on the number of integer solutions of the Diophantine\nsystems associated with these surfaces.\n",
"title": "On a binary system of Prendiville: The cubic case"
} | null | null | null | null | true | null | 912 | null | Default | null | null |
null | {
"abstract": " In this paper, we introduce a simple, yet powerful pipeline for medical image\nsegmentation that combines Fully Convolutional Networks (FCNs) with Fully\nConvolutional Residual Networks (FC-ResNets). We propose and examine a design\nthat takes particular advantage of recent advances in the understanding of both\nConvolutional Neural Networks as well as ResNets. Our approach focuses upon the\nimportance of a trainable pre-processing when using FC-ResNets and we show that\na low-capacity FCN model can serve as a pre-processor to normalize medical\ninput data. In our image segmentation pipeline, we use FCNs to obtain\nnormalized images, which are then iteratively refined by means of a FC-ResNet\nto generate a segmentation prediction. As in other fully convolutional\napproaches, our pipeline can be used off-the-shelf on different image\nmodalities. We show that using this pipeline, we exhibit state-of-the-art\nperformance on the challenging Electron Microscopy benchmark, when compared to\nother 2D methods. We improve segmentation results on CT images of liver\nlesions, when contrasting with standard FCN methods. Moreover, when applying\nour 2D pipeline on a challenging 3D MRI prostate segmentation challenge we\nreach results that are competitive even when compared to 3D methods. The\nobtained results illustrate the strong potential and versatility of the\npipeline by achieving highly accurate results on multi-modality images from\ndifferent anatomical regions and organs.\n",
"title": "Learning Normalized Inputs for Iterative Estimation in Medical Image Segmentation"
} | null | null | null | null | true | null | 913 | null | Default | null | null |
null | {
"abstract": " Our purpose is to focus attention on a new criterion for quantum schemes by\nbringing together the notions of quantum game and game isomorphism. A quantum\ngame scheme is required to generate the classical game as a special case. Now,\ngiven a quantum game scheme and two isomorphic classical games, we additionally\nrequire the resulting quantum games to be isomorphic as well. We show how this\nisomorphism condition influences the players' strategy sets. We are concerned\nwith the Marinatto-Weber type quantum game scheme and the strong isomorphism\nbetween games in strategic form.\n",
"title": "Strong isomorphism in Marinatto-Weber type quantum games"
} | null | null | null | null | true | null | 914 | null | Default | null | null |
null | {
"abstract": " Stochastic variance reduction algorithms have recently become popular for\nminimizing the average of a large but finite number of loss functions. In this\npaper, we propose a novel Riemannian extension of the Euclidean stochastic\nvariance reduced gradient algorithm (R-SVRG) to a manifold search space. The\nkey challenges of averaging, adding, and subtracting multiple gradients are\naddressed with retraction and vector transport. We present a global convergence\nanalysis of the proposed algorithm with a decay step size and a local\nconvergence rate analysis under a fixed step size under some natural\nassumptions. The proposed algorithm is applied to problems on the Grassmann\nmanifold, such as principal component analysis, low-rank matrix completion, and\ncomputation of the Karcher mean of subspaces, and outperforms the standard\nRiemannian stochastic gradient descent algorithm in each case.\n",
"title": "Riemannian stochastic variance reduced gradient"
} | null | null | null | null | true | null | 915 | null | Default | null | null |
null | {
"abstract": " The algorithmic Markov condition states that the most likely causal direction\nbetween two random variables X and Y can be identified as that direction with\nthe lowest Kolmogorov complexity. Due to the halting problem, however, this\nnotion is not computable.\nWe hence propose to do causal inference by stochastic complexity. That is, we\npropose to approximate Kolmogorov complexity via the Minimum Description Length\n(MDL) principle, using a score that is mini-max optimal with regard to the\nmodel class under consideration. This means that even in an adversarial\nsetting, such as when the true distribution is not in this class, we still\nobtain the optimal encoding for the data relative to the class.\nWe instantiate this framework, which we call CISC, for pairs of univariate\ndiscrete variables, using the class of multinomial distributions. Experiments\nshow that CISC is highly accurate on synthetic, benchmark, as well as\nreal-world data, outperforming the state of the art by a margin, and scales\nextremely well with regard to sample and domain sizes.\n",
"title": "Causal Inference by Stochastic Complexity"
} | null | null | null | null | true | null | 916 | null | Default | null | null |
null | {
"abstract": " The recently proposed Temporal Ensembling has achieved state-of-the-art\nresults in several semi-supervised learning benchmarks. It maintains an\nexponential moving average of label predictions on each training example, and\npenalizes predictions that are inconsistent with this target. However, because\nthe targets change only once per epoch, Temporal Ensembling becomes unwieldy\nwhen learning large datasets. To overcome this problem, we propose Mean\nTeacher, a method that averages model weights instead of label predictions. As\nan additional benefit, Mean Teacher improves test accuracy and enables training\nwith fewer labels than Temporal Ensembling. Without changing the network\narchitecture, Mean Teacher achieves an error rate of 4.35% on SVHN with 250\nlabels, outperforming Temporal Ensembling trained with 1000 labels. We also\nshow that a good network architecture is crucial to performance. Combining Mean\nTeacher and Residual Networks, we improve the state of the art on CIFAR-10 with\n4000 labels from 10.55% to 6.28%, and on ImageNet 2012 with 10% of the labels\nfrom 35.24% to 9.11%.\n",
"title": "Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results"
} | null | null | null | null | true | null | 917 | null | Default | null | null |
null | {
"abstract": " We present a new code for astrophysical magneto-hydrodynamics specifically\ndesigned and optimized for high performance and scaling on modern and future\nsupercomputers. We describe a novel hybrid OpenMP/MPI programming model that\nemerged from a collaboration between Cray, Inc. and the University of\nMinnesota. This design utilizes MPI-RMA optimized for thread scaling, which\nallows the code to run extremely efficiently at very high thread counts ideal\nfor the latest generation of the multi-core and many-core architectures. Such\nperformance characteristics are needed in the era of \"exascale\" computing. We\ndescribe and demonstrate our high-performance design in detail with the intent\nthat it may be used as a model for other, future astrophysical codes intended\nfor applications demanding exceptional performance.\n",
"title": "WOMBAT: A Scalable and High Performance Astrophysical MHD Code"
} | null | null | null | null | true | null | 918 | null | Default | null | null |
null | {
"abstract": " Virtual Network Functions as a Service (VNFaaS) is currently under attentive\nstudy by telecommunications and cloud stakeholders as a promising business and\ntechnical direction consisting of providing network functions as a service on a\ncloud (NFV Infrastructure), instead of delivering standalone network\nappliances, in order to provide higher scalability and reduce maintenance\ncosts. However, the functioning of such NFVI hosting the VNFs is fundamental\nfor all the services and applications running on top of it, forcing to\nguarantee a high availability level. Indeed the availability of an VNFaaS\nrelies on the failure rate of its single components, namely the servers, the\nvirtualization software, and the communication network. The proper assignment\nof the virtual machines implementing network functions to NFVI servers and\ntheir protection is essential to guarantee high availability. We model the High\nAvailability Virtual Network Function Placement (HA-VNFP) as the problem of\nfinding the best assignment of virtual machines to servers guaranteeing\nprotection by replication. We propose a probabilistic approach to measure the\nreal availability of a system and design both efficient and effective\nalgorithms that can be used by stakeholders for both online and offline\nplanning.\n",
"title": "Securing Virtual Network Function Placement with High Availability Guarantees"
} | null | null | null | null | true | null | 919 | null | Default | null | null |
null | {
"abstract": " The Zika virus has been found in individual cases but has not been confirmed\nas the cause of in the large number of cases of microcephaly in Brazil in\n2015-6. Indeed, disparities between the incidence of Zika and microcephaly\nacross geographic locations has led to questions about the virus's role. Here\nwe consider whether the insecticide pyriproxyfen used in Brazilian drinking\nwater might be the primary cause or a cofactor. Pyriproxifen is a juvenile\nhormone analog which has been shown to correspond in mammals to a number of fat\nsoluble regulatory molecules including retinoic acid, a metabolite of vitamin\nA, with which it has cross-reactivity and whose application during development\nhas been shown to cause microcephaly. Methoprene, another juvenile hormone\nanalog approved as an insecticide in the 1970s has been shown to cause\ndevelopmental disorders in mammals. Isotretinoin is another retinoid causing\nmicrocephaly via activation of the retinoid X receptor in developing fetuses.\nWe review tests of pyriproxyfen by the manufacturer Sumitomo, which actually\nfound some evidence for this effect, including low brain mass and\narhinencephaly in exposed rat pups. Pyriproxyfen use in Brazil is\nunprecedented, never having been applied to a water supply on a large scale.\nClaims that its geographical pattern of use rule it out as a cause have not\nbeen documented or confirmed. On the other hand, the very few microcephaly\ncases reported in Colombia and the wide discrepancies of incidence in different\nstates across Brazil despite large numbers of Zika cases undermine the claim\nthat Zika is the cause. Given this combination of potential molecular\nmechanism, toxicological and epidemiological evidence we strongly recommend\nthat the use of pyriproxyfen in Brazil be suspended until the potential causal\nlink to microcephaly is investigated further.\n",
"title": "The Case for Pyriproxyfen as a Potential Cause for Microcephaly; From Biology to Epidemiology"
} | null | null | null | null | true | null | 920 | null | Default | null | null |
null | {
"abstract": " This paper considers the problem of decentralized optimization with a\ncomposite objective containing smooth and non-smooth terms. To solve the\nproblem, a proximal-gradient scheme is studied. Specifically, the smooth and\nnonsmooth terms are dealt with by gradient update and proximal update,\nrespectively. The studied algorithm is closely related to a previous\ndecentralized optimization algorithm, PG-EXTRA [37], but has a few advantages.\nFirst of all, in our new scheme, agents use uncoordinated step-sizes and the\nstable upper bounds on step-sizes are independent from network topology. The\nstep-sizes depend on local objective functions, and they can be as large as\nthat of the gradient descent. Secondly, for the special case without non-smooth\nterms, linear convergence can be achieved under the strong convexity\nassumption. The dependence of the convergence rate on the objective functions\nand the network are separated, and the convergence rate of our new scheme is as\ngood as one of the two convergence rates that match the typical rates for the\ngeneral gradient descent and the consensus averaging. We also provide some\nnumerical experiments to demonstrate the efficacy of the introduced algorithms\nand validate our theoretical discoveries.\n",
"title": "A decentralized proximal-gradient method with network independent step-sizes and separated convergence rates"
} | null | null | [
"Mathematics",
"Statistics"
]
| null | true | null | 921 | null | Validated | null | null |
null | {
"abstract": " Recent studies have demonstrated that near-data processing (NDP) is an\neffective technique for improving performance and energy efficiency of\ndata-intensive workloads. However, leveraging NDP in realistic systems with\nmultiple memory modules introduces a new challenge. In today's systems, where\nno computation occurs in memory modules, the physical address space is\ninterleaved at a fine granularity among all memory modules to help improve the\nutilization of processor-memory interfaces by distributing the memory traffic.\nHowever, this is at odds with efficient use of NDP, which requires careful\nplacement of data in memory modules such that near-data computations and their\nexclusively used data can be localized in individual memory modules, while\ndistributing shared data among memory modules to reduce hotspots. In order to\naddress this new challenge, we propose a set of techniques that (1) enable\ncollections of OS pages to either be fine-grain interleaved among memory\nmodules (as is done today) or to be placed contiguously on individual memory\nmodules (as is desirable for NDP private data), and (2) decide whether to\nlocalize or distribute each memory object based on its anticipated access\npattern and steer computations to the memory where the data they access is\nlocated. Our evaluations across a wide range of workloads show that the\nproposed mechanism improves performance by 31% and reduces 38% remote data\naccesses over a baseline system that cannot exploit computate-data affinity\ncharacteristics.\n",
"title": "CODA: Enabling Co-location of Computation and Data for Near-Data Processing"
} | null | null | null | null | true | null | 922 | null | Default | null | null |
null | {
"abstract": " Historically, machine learning in computer security has prioritized defense:\nthink intrusion detection systems, malware classification, and botnet traffic\nidentification. Offense can benefit from data just as well. Social networks,\nwith their access to extensive personal data, bot-friendly APIs, colloquial\nsyntax, and prevalence of shortened links, are the perfect venues for spreading\nmachine-generated malicious content. We aim to discover what capabilities an\nadversary might utilize in such a domain. We present a long short-term memory\n(LSTM) neural network that learns to socially engineer specific users into\nclicking on deceptive URLs. The model is trained with word vector\nrepresentations of social media posts, and in order to make a click-through\nmore likely, it is dynamically seeded with topics extracted from the target's\ntimeline. We augment the model with clustering to triage high value targets\nbased on their level of social engagement, and measure success of the LSTM's\nphishing expedition using click-rates of IP-tracked links. We achieve state of\nthe art success rates, tripling those of historic email attack campaigns, and\noutperform humans manually performing the same task.\n",
"title": "Generative Models for Spear Phishing Posts on Social Media"
} | null | null | null | null | true | null | 923 | null | Default | null | null |
null | {
"abstract": " Deep learning has demonstrated tremendous potential for Automatic Text\nScoring (ATS) tasks. In this paper, we describe a new neural architecture that\nenhances vanilla neural network models with auxiliary neural coherence\nfeatures. Our new method proposes a new \\textsc{SkipFlow} mechanism that models\nrelationships between snapshots of the hidden representations of a long\nshort-term memory (LSTM) network as it reads. Subsequently, the semantic\nrelationships between multiple snapshots are used as auxiliary features for\nprediction. This has two main benefits. Firstly, essays are typically long\nsequences and therefore the memorization capability of the LSTM network may be\ninsufficient. Implicit access to multiple snapshots can alleviate this problem\nby acting as a protection against vanishing gradients. The parameters of the\n\\textsc{SkipFlow} mechanism also acts as an auxiliary memory. Secondly,\nmodeling relationships between multiple positions allows our model to learn\nfeatures that represent and approximate textual coherence. In our model, we\ncall this \\textit{neural coherence} features. Overall, we present a unified\ndeep learning architecture that generates neural coherence features as it reads\nin an end-to-end fashion. Our approach demonstrates state-of-the-art\nperformance on the benchmark ASAP dataset, outperforming not only feature\nengineering baselines but also other deep learning models.\n",
"title": "SkipFlow: Incorporating Neural Coherence Features for End-to-End Automatic Text Scoring"
} | null | null | null | null | true | null | 924 | null | Default | null | null |
null | {
"abstract": " This paper introduces a method, based on deep reinforcement learning, for\nautomatically generating a general purpose decision making function. A Deep\nQ-Network agent was trained in a simulated environment to handle speed and lane\nchange decisions for a truck-trailer combination. In a highway driving case, it\nis shown that the method produced an agent that matched or surpassed the\nperformance of a commonly used reference model. To demonstrate the generality\nof the method, the exact same algorithm was also tested by training it for an\novertaking case on a road with oncoming traffic. Furthermore, a novel way of\napplying a convolutional neural network to high level input that represents\ninterchangeable objects is also introduced.\n",
"title": "Automated Speed and Lane Change Decision Making using Deep Reinforcement Learning"
} | null | null | null | null | true | null | 925 | null | Default | null | null |
null | {
"abstract": " One of the most challenging problems in correlated topological systems is a\nrealization of the reduction of topological classification, but very few\nexperimental platforms have been proposed so far. We here demonstrate that\nultracold dipolar fermions (e.g., $^{167}$Er, $^{161}$Dy, and $^{53}$Cr) loaded\nin an optical lattice of two-leg ladder geometry can be the first promising\ntestbed for the reduction $\\mathbb{Z}\\to\\mathbb{Z}_4$, where solid evidence for\nthe reduction is available thanks to their high controllability. We further\ngive a detailed account of how to experimentally access this phenomenon; around\nthe edges, the destruction of one-particle gapless excitations can be observed\nby the local radio frequency spectroscopy, while that of gapless spin\nexcitations can be observed by a time-dependent spin expectation value of a\nsuperposed state of the ground state and the first excited state. We clarify\nthat even when the reduction occurs, a gapless edge mode is recovered around a\ndislocation, which can be another piece of evidence for the reduction.\n",
"title": "Reduction of topological $\\mathbb{Z}$ classification in cold atomic systems"
} | null | null | null | null | true | null | 926 | null | Default | null | null |
null | {
"abstract": " Cosmological parameter constraints from observations of time-delay lenses are\nbecoming increasingly precise. However, there may be significant bias and\nscatter in these measurements due to, among other things, the so-called\nmass-sheet degeneracy. To estimate these uncertainties, we analyze strong\nlenses from the largest EAGLE hydrodynamical simulation. We apply a mass-sheet\ntransformation to the radial density profiles of lenses, and by selecting\nlenses near isothermality, we find that the bias on H0 can be reduced to 5%\nwith an intrinsic scatter of 10%, confirming previous results performed on a\ndifferent simulation data set. We further investigate whether combining lensing\nobservables with kinematic constraints helps to minimize this bias. We do not\ndetect any significant dependence of the bias on lens model parameters or\nobservational properties of the galaxy, but depending on the source--lens\nconfiguration, a bias may still exist. Cross lenses provide an accurate\nestimate of the Hubble constant, while fold (double) lenses tend to be biased\nlow (high). With kinematic constraints, double lenses show bias and intrinsic\nscatter of 6% and 10%, respectively, while quad lenses show bias and intrinsic\nscatter of 0.5% and 10%, respectively. For lenses with a reduced $\\chi^2 > 1$,\na power-law dependence of the $\\chi^2$ on the lens environment (number of\nnearby galaxies) is seen. Lastly, we model, in greater detail, the cases of two\ndouble lenses that are significantly biased. We are able to remove the bias,\nsuggesting that the remaining biases could also be reduced by carefully taking\ninto account additional sources of systematic uncertainty.\n",
"title": "Reducing biases on $H_0$ measurements using strong lensing and galaxy dynamics: results from the EAGLE simulation"
} | null | null | [
"Physics"
]
| null | true | null | 927 | null | Validated | null | null |
null | {
"abstract": " We characterize the response of the quiet time (no substorms or storms)\nlarge-scale ionospheric transient equivalent currents to north-south and\nsouth-north IMF turnings by using a dynamical network of ground-based\nmagnetometers. Canonical correlation between all pairs of SuperMAG magnetometer\nstations in the Northern Hemisphere (magnetic latitude (MLAT) 50-82$^{\\circ}$)\nis used to establish the extent of near-simultaneous magnetic response between\nregions of magnetic local time-MLAT. Parameters and maps that describe\nspatial-temporal correlation are used to characterize the system and its\nresponse to the turnings aggregated over several hundred events. We find that\nregions that experience large increases in correlation post turning coincide\nwith typical locations of a two-cell convection system and are influenced by\nthe interplanetary magnetic field $\\mathit{B}_{y}$. The time between the\nturnings reaching the magnetopause and a network response is found to be\n$\\sim$8-10 min and correlation in the dayside occurs 2-8 min before that in the\nnightside.\n",
"title": "Characterizing the ionospheric current pattern response to southward and northward IMF turnings with dynamical SuperMAG correlation networks"
} | null | null | [
"Physics"
]
| null | true | null | 928 | null | Validated | null | null |
null | {
"abstract": " Let R be a local ring of dimension d. Buchweitz asks if the rank of the d-th\nsyzygy of a module of finite lengh is greater than or equal to the rank of the\nd-th syzygy of the residue field, unless the module has finite projective\ndimension. Assuming that R is Gorenstein, we prove that if the question is\naffrmative, then R is a hypersurface. If moreover R has dimension two, then we\nshow that the converse also holds true.\n",
"title": "On a question of Buchweitz about ranks of syzygies of modules of finite length"
} | null | null | null | null | true | null | 929 | null | Default | null | null |
null | {
"abstract": " Learning to make decisions from observed data in dynamic environments remains\na problem of fundamental importance in a number of fields, from artificial\nintelligence and robotics, to medicine and finance. This paper concerns the\nproblem of learning control policies for unknown linear dynamical systems so as\nto maximize a quadratic reward function. We present a method to optimize the\nexpected value of the reward over the posterior distribution of the unknown\nsystem parameters, given data. The algorithm involves sequential convex\nprograming, and enjoys reliable local convergence and robust stability\nguarantees. Numerical simulations and stabilization of a real-world inverted\npendulum are used to demonstrate the approach, with strong performance and\nrobustness properties observed in both.\n",
"title": "Learning convex bounds for linear quadratic control policy synthesis"
} | null | null | [
"Statistics"
]
| null | true | null | 930 | null | Validated | null | null |
null | {
"abstract": " Based on the KP hierarchy reduction method, the general bright-dark mixed\nmulti-soliton solution of the multi-component Maccari system is constructed.\nThe multi-component Maccari system considered comprised of multiple (say $M$)\nshort-wave components and one long-wave component with all possible\ncombinations of nonlinearities including all-focusing, all-defocusing and mixed\ntypes. We firstly derive the two-bright-one-dark (2-b-1-d) and\none-bright-two-dark (1-b-2-d) mixed multi-soliton solutions to the\nthree-component Maccari system in detail. For the interaction between two\nsolitons, the asymptotic analysis shows that inelastic collision can take place\nin a $M$-component Maccari system with $M \\geq 3$ only if the bright parts of\nthe mixed solitons appear at least in two short-wave components. The\nenergy-exchanging inelastic collision characterized by an intensity\nredistribution among the bright parts of the mixed solitons. While the dark\nparts of the mixed solitons and the solitons in the long-wave component always\nundergo elastic collision which just accompanied by a position shift. In the\nend, we extend the corresponding analysis to the $M$-component Maccari system\nto obtain its mixed multi-soliton solution. The formula obtained unifies the\nall-bright, all-dark and mixed multi-soliton solutions.\n",
"title": "General mixed multi-soliton solution to the multi-component Maccari system"
} | null | null | null | null | true | null | 931 | null | Default | null | null |
null | {
"abstract": " Deep learning has been successfully applied to various tasks, but its\nunderlying mechanism remains unclear. Neural networks associate similar inputs\nin the visible layer to the same state of hidden variables in deep layers. The\nfraction of inputs that are associated to the same state is a natural measure\nof similarity and is simply related to the cost in bits required to represent\nthese inputs. The degeneracy of states with the same information cost provides\ninstead a natural measure of noise and is simply related the entropy of the\nfrequency of states, that we call relevance. Representations with minimal\nnoise, at a given level of similarity (resolution), are those that maximise the\nrelevance. A signature of such efficient representations is that frequency\ndistributions follow power laws. We show, in extensive numerical experiments,\nthat deep neural networks extract a hierarchy of efficient representations from\ndata, because they i) achieve low levels of noise (i.e. high relevance) and ii)\nexhibit power law distributions. We also find that the layer that is most\nefficient to reliably generate patterns of training data is the one for which\nrelevance and resolution are traded at the same price, which implies that\nfrequency distribution follows Zipf's law.\n",
"title": "Resolution and Relevance Trade-offs in Deep Learning"
} | null | null | null | null | true | null | 932 | null | Default | null | null |
null | {
"abstract": " The modular Gromov-Hausdorff propinquity is a distance on classes of modules\nendowed with quantum metric information, in the form of a metric form of a\nconnection and a left Hilbert module structure. This paper proves that the\nfamily of Heisenberg modules over quantum two tori, when endowed with their\ncanonical connections, form a family of metrized quantum vector bundles, as a\nfirst step in proving that Heisenberg modules form a continuous family for the\nmodular Gromov-Hausdorff propinquity.\n",
"title": "Heisenberg Modules over Quantum 2-tori are metrized quantum vector bundles"
} | null | null | null | null | true | null | 933 | null | Default | null | null |
null | {
"abstract": " We present a prototype of a software tool for exploration of multiple\ncombinatorial optimisation problems in large real-world and synthetic complex\nnetworks. Our tool, called GraphCombEx (an acronym of Graph Combinatorial\nExplorer), provides a unified framework for scalable computation and\npresentation of high-quality suboptimal solutions and bounds for a number of\nwidely studied combinatorial optimisation problems. Efficient representation\nand applicability to large-scale graphs and complex networks are particularly\nconsidered in its design. The problems currently supported include maximum\nclique, graph colouring, maximum independent set, minimum vertex clique\ncovering, minimum dominating set, as well as the longest simple cycle problem.\nSuboptimal solutions and intervals for optimal objective values are estimated\nusing scalable heuristics. The tool is designed with extensibility in mind,\nwith the view of further problems and both new fast and high-performance\nheuristics to be added in the future. GraphCombEx has already been successfully\nused as a support tool in a number of recent research studies using\ncombinatorial optimisation to analyse complex networks, indicating its promise\nas a research software tool.\n",
"title": "GraphCombEx: A Software Tool for Exploration of Combinatorial Optimisation Properties of Large Graphs"
} | null | null | null | null | true | null | 934 | null | Default | null | null |
null | {
"abstract": " Word sense disambiguation (WSD) improves many Natural Language Processing\n(NLP) applications such as Information Retrieval, Machine Translation or\nLexical Simplification. WSD is the ability of determining a word sense among\ndifferent ones within a polysemic lexical unit taking into account the context.\nThe most straightforward approach uses a semantic proximity measure between the\nword sense candidates of the target word and those of its context. Such a\nmethod very easily entails a combinatorial explosion. In this paper, we propose\ntwo methods based on distributional analysis which enable to reduce the\nexponential complexity without losing the coherence. We present a comparison\nbetween the selection of distributional neighbors and the linearly nearest\nneighbors. The figures obtained show that selecting distributional neighbors\nleads to better results.\n",
"title": "Approches d'analyse distributionnelle pour améliorer la désambiguïsation sémantique"
} | null | null | null | null | true | null | 935 | null | Default | null | null |
null | {
"abstract": " We develop a theory for non-degenerate parametric resonance in a tunable\nsuperconducting cavity. We focus on nonlinear effects that are caused by\nnonlinear Josephson elements connected to the cavity. We analyze parametric\namplification in a strong nonlinear regime at the parametric instability\nthreshold, and calculate maximum gain values. Above the threshold, in the\nparametric oscillator regime the linear cavity response diverges at the\noscillator frequency at all pump strengths. We show that this divergence is\nrelated to the continuous degeneracy of the free oscillator state with respect\nto the phase. Applying on-resonance input lifts the degeneracy and removes the\ndivergence. We also investigate the quantum noise squeezing. It is shown that\nin the strong amplification regime the noise undergoes four-mode squeezing, and\nthat in this regime the output signal to noise ratio can significantly exceed\nthe input value. We also analyze the intermode frequency conversion and\nidentify parameters at which full conversion is achieved.\n",
"title": "Non-degenerate parametric resonance in tunable superconducting cavity"
} | null | null | [
"Physics"
]
| null | true | null | 936 | null | Validated | null | null |
null | {
"abstract": " Generic generation and manipulation of text is challenging and has limited\nsuccess compared to recent deep generative modeling in visual domain. This\npaper aims at generating plausible natural language sentences, whose attributes\nare dynamically controlled by learning disentangled latent representations with\ndesignated semantics. We propose a new neural generative model which combines\nvariational auto-encoders and holistic attribute discriminators for effective\nimposition of semantic structures. With differentiable approximation to\ndiscrete text samples, explicit constraints on independent attribute controls,\nand efficient collaborative learning of generator and discriminators, our model\nlearns highly interpretable representations from even only word annotations,\nand produces realistic sentences with desired attributes. Quantitative\nevaluation validates the accuracy of sentence and attribute generation.\n",
"title": "Toward Controlled Generation of Text"
} | null | null | null | null | true | null | 937 | null | Default | null | null |
null | {
"abstract": " Our eyes sample a disproportionately large amount of information at the\ncentre of gaze with increasingly sparse sampling into the periphery. This\nsampling scheme is widely believed to be a wiring constraint whereby high\nresolution at the centre is achieved by sacrificing spatial acuity in the\nperiphery. Here we propose that this sampling scheme may be optimal for object\nrecognition because the relevant spatial content is dense near an object and\nsparse in the surrounding vicinity. We tested this hypothesis by training deep\nconvolutional neural networks on full-resolution and foveated images. Our main\nfinding is that networks trained on images with foveated sampling show better\nobject classification compared to networks trained on full resolution images.\nImportantly, blurring images according to the human blur function yielded the\nbest performance compared to images with shallower or steeper blurring. Taken\ntogether our results suggest that, peripheral blurring in our eyes may have\nevolved for optimal object recognition, rather than merely to satisfy wiring\nconstraints.\n",
"title": "Human peripheral blur is optimal for object recognition"
} | null | null | null | null | true | null | 938 | null | Default | null | null |
null | {
"abstract": " Molecular dynamics simulates the~movements of atoms. Due to its high cost,\nmany methods have been developed to \"push the~simulation forward\". One of them,\nmetadynamics, can hasten the~molecular dynamics with the~help of variables\ndescribing the~simulated process. However, the~evaluation of these variables\ncan include numerous mean square distance calculations that introduce\nsubstantial computational demands, thus jeopardize the~benefit of the~approach.\nRecently, we proposed an~approximative method that significantly reduces\nthe~number of these distance calculations. Here we evaluate the~performance and\nthe~scalability on two molecular systems. We assess the~maximal theoretical\nspeed-up based on the reduction of distance computations and Ahmdal's law and\ncompare it to the~practical speed-up achieved with our implementation.\n",
"title": "Acceleration of Mean Square Distance Calculations with Floating Close Structure in Metadynamics Simulations"
} | null | null | null | null | true | null | 939 | null | Default | null | null |
null | {
"abstract": " Developers increasingly rely on text matching tools to analyze the relation\nbetween natural language words and APIs. However, semantic gaps, namely textual\nmismatches between words and APIs, negatively affect these tools. Previous\nstudies have transformed words or APIs into low-dimensional vectors for\nmatching; however, inaccurate results were obtained due to the failure of\nmodeling words and APIs simultaneously. To resolve this problem, two main\nchallenges are to be addressed: the acquisition of massive words and APIs for\nmining and the alignment of words and APIs for modeling. Therefore, this study\nproposes Word2API to effectively estimate relatedness of words and APIs.\nWord2API collects millions of commonly used words and APIs from code\nrepositories to address the acquisition challenge. Then, a shuffling strategy\nis used to transform related words and APIs into tuples to address the\nalignment challenge. Using these tuples, Word2API models words and APIs\nsimultaneously. Word2API outperforms baselines by 10%-49.6% of relatedness\nestimation in terms of precision and NDCG. Word2API is also effective on\nsolving typical software tasks, e.g., query expansion and API documents\nlinking. A simple system with Word2API-expanded queries recommends up to 21.4%\nmore related APIs for developers. Meanwhile, Word2API improves comparison\nalgorithms by 7.9%-17.4% in linking questions in Question&Answer communities to\nAPI documents.\n",
"title": "Bridging Semantic Gaps between Natural Languages and APIs with Word Embedding"
} | null | null | null | null | true | null | 940 | null | Default | null | null |
null | {
"abstract": " We prove the existence of an optimal feedback controller for a stochastic\noptimization problem constituted by a variation of the Heston model, where a\nstochastic input process is added in order to minimize a given performance\ncriterion. The stochastic feedback controller is searched by solving a\nnonlinear backward parabolic equation for which one proves the existence of a\nmartingale solution.\n",
"title": "Feedback optimal controllers for the Heston model"
} | null | null | null | null | true | null | 941 | null | Default | null | null |
null | {
"abstract": " How can we enable novice users to create effective task plans for\ncollaborative robots? Must there be a tradeoff between generalizability and\nease of use? To answer these questions, we conducted a user study with the\nCoSTAR system, which integrates perception and reasoning into a Behavior\nTree-based task plan editor. In our study, we ask novice users to perform\nsimple pick-and-place assembly tasks under varying perception and planning\ncapabilities. Our study shows that users found Behavior Trees to be an\neffective way of specifying task plans. Furthermore, users were also able to\nmore quickly, effectively, and generally author task plans with the addition of\nCoSTAR's planning, perception, and reasoning capabilities. Despite these\nimprovements, concepts associated with these capabilities were rated by users\nas less usable, and our results suggest a direction for further refinement.\n",
"title": "User Experience of the CoSTAR System for Instruction of Collaborative Robots"
} | null | null | null | null | true | null | 942 | null | Default | null | null |
null | {
"abstract": " The (torsion) complexity of a finite edge-weighted graph is defined to be the\norder of the torsion subgroup of the abelian group presented by its Laplacian\nmatrix. When G is d-periodic (i.e., G has a free action of the rank-d free\nabelian group by graph automorphisms, with finite quotient) the Mahler measure\nof its Laplacian determinant polynomial is the growth rate of the complexity of\nfinite quotients of G. Lehmer's question, an open question about the roots of\nmonic integral polynomials, is equivalent to a question about the complexity\ngrowth of edge-weighted 1-periodic graphs.\n",
"title": "Graph complexity and Mahler measure"
} | null | null | null | null | true | null | 943 | null | Default | null | null |
null | {
"abstract": " Neutronic performance is investigated for a potential accident tolerant fuel\n(ATF),which consists of U$_3$Si$_2$ fuel and FeCrAl cladding. In comparison\nwith current UO$_2$-Zr system, FeCrAl has a better oxidation resistance but a\nlarger thermal neutron absorption cross section. U$_3$Si$_2$ has a higher\nthermal conductivity and a higher uranium density, which can compensate the\nreactivity suppressed by FeCrAl. Based on neutronic investigations, a possible\nU$_3$Si$_2$-FeCrAl fuel-cladding systemis taken into consideration. Fundamental\nproperties of the suggested fuel-cladding combination are investigated in a\nfuel assembly.These properties include moderator and fuel temperature\ncoefficients, control rods worth, radial power distribution (in a fuel rod),\nand different void reactivity coefficients. The present work proves that the\nnew combination has less reactivity variation during its service lifetime.\nAlthough, compared with the current system, it has a little larger deviation on\npower distribution and a little less negative temperature coefficient and void\nreactivity coefficient and its control rods worth is less important, variations\nof these parameters are less important during the service lifetime of fuel.\nHence, U$_3$Si$_2$-FeCrAl system is a potential ATF candidate from a neutronic\nview.\n",
"title": "Neutronic Analysis on Potential Accident Tolerant Fuel-Cladding Combination U$_3$Si$_2$-FeCrAl"
} | null | null | null | null | true | null | 944 | null | Default | null | null |
null | {
"abstract": " We prove that the only entrywise transforms of rectangular matrices which\npreserve total positivity or total non-negativity are either constant or\nlinear. This follows from an extended classification of preservers of these two\nproperties for matrices of fixed dimension. We also prove that the same\nassertions hold upon working only with symmetric matrices; for total-positivity\npreservers our proofs proceed through solving two totally positive completion\nproblems.\n",
"title": "Total-positivity preservers"
} | null | null | null | null | true | null | 945 | null | Default | null | null |
null | {
"abstract": " The most popular and widely used subtract-with-borrow generator, also known\nas RANLUX, is reimplemented as a linear congruential generator using large\ninteger arithmetic with the modulus size of 576 bits. Modern computers, as well\nas the specific structure of the modulus inferred from RANLUX, allow for the\ndevelopment of a fast modular multiplication -- the core of the procedure. This\nwas previously believed to be slow and have too high cost in terms of computing\nresources. Our tests show a significant gain in generation speed which is\ncomparable with other fast, high quality random number generators. An\nadditional feature is the fast skipping of generator states leading to a\nseeding scheme which guarantees the uniqueness of random number sequences.\n",
"title": "A revision of the subtract-with-borrow random number generators"
} | null | null | null | null | true | null | 946 | null | Default | null | null |
null | {
"abstract": " We consider the refined topological vertex of Iqbal et al, as a function of\ntwo parameters (x, y), and deform it by introducing Macdonald parameters (q,\nt), as in the work of Vuletic on plane partitions, to obtain 'a Macdonald\nrefined topological vertex'. In the limit q -> t, we recover the refined\ntopological vertex of Iqbal et al. In the limit x -> y, we obtain a\nqt-deformation of the topological vertex of Aganagic et al. Copies of the\nvertex can be glued to obtain qt-deformed 5D instanton partition functions that\nhave well-defined 4D limits and, for generic values of (q, t), contain\ninfinite-towers of poles for every pole in the limit q -> t.\n",
"title": "A Macdonald refined topological vertex"
} | null | null | null | null | true | null | 947 | null | Default | null | null |
null | {
"abstract": " We investigate bias voltage effects on the spin-dependent transport\nproperties of Fe/MgAl${}_2$O${}_4$/Fe(001) magnetic tunneling junctions (MTJs)\nby comparing them with those of Fe/MgO/Fe(001) MTJs. By means of the\nnonequilibrium Green's function method and the density functional theory, we\ncalculate bias voltage dependences of magnetoresistance (MR) ratios in both the\nMTJs. We find that in both the MTJs, the MR ratio decreases as the bias voltage\nincreases and finally vanishes at a critical bias voltage $V_{\\rm c}$. We also\nfind that the critical bias voltage $V_{\\rm c}$ of the MgAl${}_2$O${}_4$-based\nMTJ is clearly larger than that of the MgO-based MTJ. Since the in-plane\nlattice constant of the Fe/MgAl${}_2$O${}_4$/Fe(001) supercell is twice that of\nthe Fe/MgO/Fe(001) one, the Fe electrodes in the MgAl${}_2$O${}_4$-based MTJs\nhave an identical band structure to that obtained by folding the Fe band\nstructure of the MgO-based MTJs in the Brillouin zone of the in-plane wave\nvector. We show that such a difference in the Fe band structure is the origin\nof the difference in the critical bias voltage $V_{\\rm c}$ between the\nMgAl${}_2$O${}_4$- and MgO-based MTJs.\n",
"title": "Bias voltage effects on tunneling magnetoresistance in Fe/MgAl${}_2$O${}_4$/Fe(001) junctions: Comparative study with Fe/MgO/Fe(001) junctions"
} | null | null | null | null | true | null | 948 | null | Default | null | null |
null | {
"abstract": " Extensive efforts have been devoted to recognizing facial action units (AUs).\nHowever, it is still challenging to recognize AUs from spontaneous facial\ndisplays especially when they are accompanied with speech. Different from all\nprior work that utilized visual observations for facial AU recognition, this\npaper presents a novel approach that recognizes speech-related AUs exclusively\nfrom audio signals based on the fact that facial activities are highly\ncorrelated with voice during speech. Specifically, dynamic and physiological\nrelationships between AUs and phonemes are modeled through a continuous time\nBayesian network (CTBN); then AU recognition is performed by probabilistic\ninference via the CTBN model.\nA pilot audiovisual AU-coded database has been constructed to evaluate the\nproposed audio-based AU recognition framework. The database consists of a\n\"clean\" subset with frontal and neutral faces and a challenging subset\ncollected with large head movements and occlusions. Experimental results on\nthis database show that the proposed CTBN model achieves promising recognition\nperformance for 7 speech-related AUs and outperforms the state-of-the-art\nvisual-based methods especially for those AUs that are activated at low\nintensities or \"hardly visible\" in the visual channel. Furthermore, the CTBN\nmodel yields more impressive recognition performance on the challenging subset,\nwhere the visual-based approaches suffer significantly.\n",
"title": "Listen to Your Face: Inferring Facial Action Units from Audio Channel"
} | null | null | null | null | true | null | 949 | null | Default | null | null |
null | {
"abstract": " We found an easy and quick post-learning method named \"Icing on the Cake\" to\nenhance a classification performance in deep learning. The method is that we\ntrain only the final classifier again after an ordinary training is done.\n",
"title": "Icing on the Cake: An Easy and Quick Post-Learnig Method You Can Try After Deep Learning"
} | null | null | null | null | true | null | 950 | null | Default | null | null |
null | {
"abstract": " As traditional neural network consumes a significant amount of computing\nresources during back propagation, \\citet{Sun2017mePropSB} propose a simple yet\neffective technique to alleviate this problem. In this technique, only a small\nsubset of the full gradients are computed to update the model parameters. In\nthis paper we extend this technique into the Convolutional Neural Network(CNN)\nto reduce calculation in back propagation, and the surprising results verify\nits validity in CNN: only 5\\% of the gradients are passed back but the model\nstill achieves the same effect as the traditional CNN, or even better. We also\nshow that the top-$k$ selection of gradients leads to a sparse calculation in\nback propagation, which may bring significant computational benefits for high\ncomputational complexity of convolution operation in CNN.\n",
"title": "Minimal Effort Back Propagation for Convolutional Neural Networks"
} | null | null | null | null | true | null | 951 | null | Default | null | null |
null | {
"abstract": " Mobile robots are increasingly being used to assist with active pursuit and\nlaw enforcement. One major limitation for such missions is the resource\n(battery) allocated to the robot. Factors like nature and agility of evader,\nterrain over which pursuit is being carried out, plausible traversal velocity\nand the amount of necessary data to be collected all influence how long the\nrobot can last in the field and how far it can travel. In this paper, we\ndevelop an analytical model that analyzes the energy utilization for a variety\nof components mounted on a robot to estimate the maximum operational range\nachievable by the robot operating on a single battery discharge. We categorize\nthe major consumers of energy as: 1.) ancillary robotic functions such as\ncomputation, communication, sensing etc., and 2.) maneuvering which involves\npropulsion, steering etc. Both these consumers draw power from the common power\nsource but the achievable range is largely affected by the proportion of power\navailable for maneuvering. For this case study, we performed experiments with\nreal robots on planar and graded surfaces and evaluated the estimation error\nfor each case.\n",
"title": "Estimating Achievable Range of Ground Robots Operating on Single Battery Discharge for Operational Efficacy Amelioration"
} | null | null | null | null | true | null | 952 | null | Default | null | null |
null | {
"abstract": " Mixed effects models are widely used to describe heterogeneity in a\npopulation. A crucial issue when adjusting such a model to data consists in\nidentifying fixed and random effects. From a statistical point of view, it\nremains to test the nullity of the variances of a given subset of random\neffects. Some authors have proposed to use the likelihood ratio test and have\nestablished its asymptotic distribution in some particular cases. Nevertheless,\nto the best of our knowledge, no general variance components testing procedure\nhas been fully investigated yet. In this paper, we study the likelihood ratio\ntest properties to test that the variances of a general subset of the random\neffects are equal to zero in both linear and nonlinear mixed effects model,\nextending the existing results. We prove that the asymptotic distribution of\nthe test is a chi-bar-square distribution, that is to say a mixture of\nchi-square distributions, and we identify the corresponding weights. We\nhighlight in particular that the limiting distribution depends on the presence\nof correlations between the random effects but not on the linear or nonlinear\nstructure of the mixed effects model. We illustrate the finite sample size\nproperties of the test procedure through simulation studies and apply the test\nprocedure to two real datasets of dental growth and of coucal growth.\n",
"title": "Likelihood ratio test for variance components in nonlinear mixed effects models"
} | null | null | null | null | true | null | 953 | null | Default | null | null |
null | {
"abstract": " We show that a positive Borel measure of positive finite total mass, on\ncompact Hermitian manifolds, admits a Holder continuous quasi-plurisubharmonic\nsolution to the Monge-Ampere equation if and only if it is dominated locally by\nMonge-Ampere measures of Holder continuous plurisubharmonic functions.\n",
"title": "Hölder continuous solutions of the Monge-Ampère equation on compact Hermitian manifolds"
} | null | null | null | null | true | null | 954 | null | Default | null | null |
null | {
"abstract": " The Large European Array for Pulsars combines Europe's largest radio\ntelescopes to form a tied-array telescope that provides high signal-to-noise\nobservations of millisecond pulsars (MSPs) with the objective to increase the\nsensitivity of detecting low-frequency gravitational waves. As part of this\nendeavor we have developed a software correlator and beamformer which enables\nthe formation of a tied-array beam from the raw voltages from each of\ntelescopes. We explain the concepts and techniques involved in the process of\nadding the raw voltages coherently. We further present the software processing\npipeline that is specifically designed to deal with data from widely spaced,\ninhomogeneous radio telescopes and describe the steps involved in preparing,\ncorrelating and creating the tied-array beam. This includes polarization\ncalibration, bandpass correction, frequency dependent phase correction,\ninterference mitigation and pulsar gating. A link is provided where the\nsoftware can be obtained.\n",
"title": "The beamformer and correlator for the Large European Array for Pulsars"
} | null | null | null | null | true | null | 955 | null | Default | null | null |
null | {
"abstract": " In this paper, we discuss how a suitable family of tensor kernels can be used\nto efficiently solve nonparametric extensions of $\\ell^p$ regularized learning\nmethods. Our main contribution is proposing a fast dual algorithm, and showing\nthat it allows to solve the problem efficiently. Our results contrast recent\nfindings suggesting kernel methods cannot be extended beyond Hilbert setting.\nNumerical experiments confirm the effectiveness of the method.\n",
"title": "Solving $\\ell^p\\!$-norm regularization with tensor kernels"
} | null | null | null | null | true | null | 956 | null | Default | null | null |
null | {
"abstract": " Deep Learning models are vulnerable to adversarial examples, i.e.\\ images\nobtained via deliberate imperceptible perturbations, such that the model\nmisclassifies them with high confidence. However, class confidence by itself is\nan incomplete picture of uncertainty. We therefore use principled Bayesian\nmethods to capture model uncertainty in prediction for observing adversarial\nmisclassification. We provide an extensive study with different Bayesian neural\nnetworks attacked in both white-box and black-box setups. The behaviour of the\nnetworks for noise, attacks and clean test data is compared. We observe that\nBayesian neural networks are uncertain in their predictions for adversarial\nperturbations, a behaviour similar to the one observed for random Gaussian\nperturbations. Thus, we conclude that Bayesian neural networks can be\nconsidered for detecting adversarial examples.\n",
"title": "Adversarial Phenomenon in the Eyes of Bayesian Deep Learning"
} | null | null | null | null | true | null | 957 | null | Default | null | null |
null | {
"abstract": " The recent discovery of the planetary system hosted by the ultracool dwarf\nstar TRAPPIST-1 could open new perspectives into the investigation of planetary\nclimates of Earth-sized exoplanets, their atmospheres and their possible\nhabitability. In this paper, we use a simple climate-vegetation energy-balance\nmodel to study the climate of the seven TRAPPIST-1 planets and the climate\ndependence on the global albedo, on the fraction of vegetation that could cover\ntheir surfaces and on the different greenhouse conditions. The model allows us\nto investigate whether liquid water could be maintained on the planetary\nsurfaces (i.e., by defining a \"surface water zone\") in different planetary\nconditions, with or without the presence of greenhouse effect.\nIt is shown that planet TRAPPIST-1d seems to be the most stable from an\nEarth-like perspective, since it resides in the surface water zone for a wide\nrange of reasonable values of the model parameters. Moreover, according to the\nmodel outer planets (f, g and h) cannot host liquid water on their surfaces,\neven for Earth-like conditions, entering a snowball state. Although very\nsimple, the model allows to extract the main features of the TRAPPIST-1\nplanetary climates.\n",
"title": "Comparative Climates of TRAPPIST-1 planetary system: results from a simple climate-vegetation model"
} | null | null | null | null | true | null | 958 | null | Default | null | null |
null | {
"abstract": " We show that $\\mathbb{Q}$-Fano varieties of fixed dimension with\nanti-canonical degrees and alpha-invariants bounded from below form a bounded\nfamily. As a corollary, K-semistable $\\mathbb{Q}$-Fano varieties of fixed\ndimension with anti-canonical degrees bounded from below form a bounded family.\n",
"title": "Boundedness of $\\mathbb{Q}$-Fano varieties with degrees and alpha-invariants bounded from below"
} | null | null | null | null | true | null | 959 | null | Default | null | null |
null | {
"abstract": " We consider quantum, nondterministic and probabilistic versions of known\ncomputational model Ordered Read-$k$-times Branching Programs or Ordered Binary\nDecision Diagrams with repeated test ($k$-QOBDD, $k$-NOBDD and $k$-POBDD). We\nshow width hierarchy for complexity classes of Boolean function computed by\nthese models and discuss relation between different variants of $k$-OBDD.\n",
"title": "Width Hierarchies for Quantum and Classical Ordered Binary Decision Diagrams with Repeated Test"
} | null | null | [
"Computer Science"
]
| null | true | null | 960 | null | Validated | null | null |
null | {
"abstract": " We study the relation between the microscopic properties of a many-body\nsystem and the electron spectra, experimentally accessible by photoemission. In\na recent paper [Phys. Rev. Lett. 114, 236402 (2015)], we introduced the\n\"fluctuation diagnostics\" approach, to extract the dominant wave vector\ndependent bosonic fluctuations from the electronic self-energy. Here, we first\nreformulate the theory in terms of fermionic modes, to render its connection\nwith resonance valence bond (RVB) fluctuations more transparent. Secondly, by\nusing a large-U expansion, where U is the Coulomb interaction, we relate the\nfluctuations to real space correlations. Therefore, it becomes possible to\nstudy how electron spectra are related to charge, spin, superconductivity and\nRVB-like real space correlations, broadening the analysis of an earlier work\n[Phys. Rev. B 89, 245130 (2014)]. This formalism is applied to the pseudogap\nphysics of the two-dimensional Hubbard model, studied in the dynamical cluster\napproximation. We perform calculations for embedded clusters with up to 32\nsites, having three inequivalent K-points at the Fermi surface. We find that as\nU is increased, correlation functions gradually attain values consistent with\nan RVB state. This first happens for correlation functions involving the\nantinodal point and gradually spreads to the nodal point along the Fermi\nsurface. Simultaneously a pseudogap opens up along the Fermi surface. We relate\nthis to a crossover from a Kondo-like state to an RVB-like localized cluster\nstate and to the presence of RVB and spin fluctuations. These changes are\ncaused by a strong momentum dependence in the cluster bath-couplings along the\nFermi surface. We also show, from a more algorithmic perspective, how the\ntime-consuming calculations in fluctuation diagnostics can be drastically\nsimplified.\n",
"title": "Complementary views on electron spectra: From Fluctuation Diagnostics to real space correlations"
} | null | null | null | null | true | null | 961 | null | Default | null | null |
null | {
"abstract": " We study a generic one-dimensional model for an intracellular cargo driven by\nN motor proteins against an external applied force. The model includes\nmotor-cargo and motor-motor interactions. The cargo motion is described by an\nover-damped Langevin equation, while motor dynamics is specified by hopping\nrates which follow a local detailed balance condition with respect to change in\nenergy per hopping event. Based on this model, we show that the stall force,\nthe mean external force corresponding to zero mean cargo velocity, is\ncompletely independent of the details of the interactions and is, therefore,\nalways equal to the sum of the stall forces of the individual motors. This\nexact result is arrived on the basis of a simple assumption: the (macroscopic)\nstate of stall of the cargo is analogous to a state of thermodynamic\nequilibrium, and is characterized by vanishing net probability current between\nany two microstates, with the latter specified by motor positions relative to\nthe cargo. The corresponding probability distribution of the microstates under\nstall is also determined. These predictions are in complete agreement with\nnumerical simulations, carried out using specific forms of interaction\npotentials.\n",
"title": "Stall force of a cargo driven by N interacting motor proteins"
} | null | null | [
"Physics"
]
| null | true | null | 962 | null | Validated | null | null |
null | {
"abstract": " We report experiments on an agarose gel tablet loaded with camphoric acid\n(c-boat) set into self-motion by interfacial tension gradients at the air-water\ninterface. We observe three distinct modes of c-boat motion: harmonic mode\nwhere the c-boat speed oscillates sinusoidally in time, a steady mode where the\nc-boat maintains constant speed, and a relaxation oscillation mode where the\nc-boat maintains near-zero speed between sudden jumps in speed and position at\nregular time intervals. Whereas all three modes have been separately reported\nbefore in different systems, we show they belong to a common description.\nThrough control of the air-water surface tension with Sodium Dodecyl Sulfate\n(SDS), we experimentally deduce the three self-propulsive modes result from\nsurface tension difference between Camphoric Acid (CA) and the ambient\nsurroundings.\n",
"title": "Dynamics of a Camphoric Acid boat at the air-water interface"
} | null | null | [
"Physics"
]
| null | true | null | 963 | null | Validated | null | null |
null | {
"abstract": " Artificial intelligence is revolutionizing our lives at an ever increasing\npace. At the heart of this revolution is the recent advancements in deep neural\nnetworks (DNN), learning to perform sophisticated, high-level tasks. However,\ntraining DNNs requires massive amounts of data and is very computationally\nintensive. Gaining analytical understanding of the solutions found by DNNs can\nhelp us devise more efficient training algorithms, replacing the commonly used\nmthod of stochastic gradient descent (SGD). We analyze the dynamics of SGD and\nshow that, indeed, direct computation of the solutions is possible in many\ncases. We show that a high performing setup used in DNNs introduces a\nseparation of time-scales in the training dynamics, allowing SGD to train\nlayers from the lowest (closest to input) to the highest. We then show that for\neach layer, the distribution of solutions found by SGD can be estimated using a\nclass-based principal component analysis (PCA) of the layer's input. This\nfinding allows us to forgo SGD entirely and directly derive the DNN parameters\nusing this class-based PCA, which can be well estimated using significantly\nless data than SGD. We implement these results on image datasets MNIST, CIFAR10\nand CIFAR100 and find that, in fact, layers derived using our class-based PCA\nperform comparable or superior to neural networks of the same size and\narchitecture trained using SGD. We also confirm that the class-based PCA often\nconverges using a fraction of the data required for SGD. Thus, using our method\ntraining time can be reduced both by requiring less training data than SGD, and\nby eliminating layers in the costly backpropagation step of the training.\n",
"title": "Separation of time scales and direct computation of weights in deep neural networks"
} | null | null | null | null | true | null | 964 | null | Default | null | null |
null | {
"abstract": " In this letter, we define the homodyne $q$-deformed quadrature operator.\nAnalytic expression for the wavefunctions of $q$-deformed oscillator in the\nquadrature basis are found. Furthermore, we compute the explicit analytical\nexpression for the tomogram of the $q$-deformed coherent states by finding the\neigenstates of the $q$-deformed quadrature operator.\n",
"title": "$q$-deformed quadrature operator and optical tomogram"
} | null | null | null | null | true | null | 965 | null | Default | null | null |
null | {
"abstract": " Graph edit distance (GED) is an important similarity measure adopted in a\nsimilarity-based analysis between two graphs, and computing GED is a primitive\noperator in graph database analysis. Partially due to the NP-hardness, the\nexisting techniques for computing GED are only able to process very small\ngraphs with less than 30 vertices. Motivated by this, in this paper we\nsystematically study the problems of both GED computation, and GED verification\n(i.e., verify whether the GED between two graphs is no larger than a user-given\nthreshold). Firstly, we develop a unified framework that can be instantiated\ninto either a best-first search approach AStar+ or a depth-first search\napproach DFS+. Secondly, we design anchor-aware lower bound estimation\ntechniques to compute tighter lower bounds for intermediate search states,\nwhich significantly reduce the search spaces of both AStar+ and DFS+. We also\npropose efficient techniques to compute the lower bounds. Thirdly, based on our\nunified framework, we contrast AStar+ with DFS+ regarding their time and space\ncomplexities, and recommend that AStar+ is better than DFS+ by having a much\nsmaller search space. Extensive empirical studies validate that AStar+ performs\nbetter than DFS+, and show that our AStar+-BMa approach outperforms the\nstate-of-the-art technique by more than four orders of magnitude.\n",
"title": "Efficient Graph Edit Distance Computation and Verification via Anchor-aware Lower Bound Estimation"
} | null | null | null | null | true | null | 966 | null | Default | null | null |
null | {
"abstract": " We introduce canonical measures on a locally finite simplicial complex $K$\nand study their asymptotic behavior under infinitely many barycentric\nsubdivisions. We also compute the face polynomial of the asymptotic link and\ndual block of a simplex in the $d^{th}$ barycentric subdivision $Sd^d(K)$ of\n$K$, $d\\gg0$. It is almost everywhere constant. When $K$ is finite, we study\nthe limit face polynomial of $Sd^d(K)$ after F.Brenti-V.Welker and\nE.Delucchi-A.Pixton-L.Sabalka.\n",
"title": "Asymptotic measures and links in simplicial complexes"
} | null | null | [
"Mathematics"
]
| null | true | null | 967 | null | Validated | null | null |
null | {
"abstract": " Reusing passwords across multiple websites is a common practice that\ncompromises security. Recently, Blum and Vempala have proposed password\nstrategies to help people calculate, in their heads, passwords for different\nsites without dependence on third-party tools or external devices. Thus far,\nthe security and efficiency of these \"mental algorithms\" has been analyzed only\ntheoretically. But are such methods usable? We present the first usability\nstudy of humanly computable password strategies, involving a learning phase (to\nlearn a password strategy), then a rehearsal phase (to login to a few\nwebsites), and multiple follow-up tests. In our user study, with training,\nparticipants were able to calculate a deterministic eight-character password\nfor an arbitrary new website in under 20 seconds.\n",
"title": "Usability of Humanly Computable Passwords"
} | null | null | null | null | true | null | 968 | null | Default | null | null |
null | {
"abstract": " As affordability pressures and tight rental markets in global cities mount,\nonline shared accommodation sites proliferate. Home sharing arrangements\npresent dilemmas for planning that aims to improve health and safety standards,\nwhile supporting positives such as the usage of dormant stock and the relieving\nof rental pressures on middle/lower income earners. Currently, no formal data\nexists on this internationally growing trend. Here, we present a first\nquantitative glance on shared accommodation practices across all major urban\ncenters of Australia enabled via collection and analysis of thousands of online\nlistings. We examine, countrywide, the spatial and short time scale temporal\ncharacteristics of this market, along with preliminary analysis on rents,\ndwelling types and other characteristics. Findings have implications for\nhousing policy makers and planning practitioners seeking to monitor and respond\nto housing policy and affordability pressures in formal and informal housing\nmarkets.\n",
"title": "Shared urbanism: Big data on accommodation sharing in urban Australia"
} | null | null | null | null | true | null | 969 | null | Default | null | null |
null | {
"abstract": " Interbank markets are often characterised in terms of a core-periphery\nnetwork structure, with a highly interconnected core of banks holding the\nmarket together, and a periphery of banks connected mostly to the core but not\ninternally. This paradigm has recently been challenged for short time scales,\nwhere interbank markets seem better characterised by a bipartite structure with\nmore core-periphery connections than inside the core. Using a novel\ncore-periphery detection method on the eMID interbank market, we enrich this\npicture by showing that the network is actually characterised by multiple\ncore-periphery pairs. Moreover, a transition from core-periphery to bipartite\nstructures occurs by shortening the temporal scale of data aggregation. We\nfurther show how the global financial crisis transformed the market, in terms\nof composition, multiplicity and internal organisation of core-periphery pairs.\nBy unveiling such a fine-grained organisation and transformation of the\ninterbank market, our method can find important applications in the\nunderstanding of how distress can propagate over financial networks.\n",
"title": "Structural changes in the interbank market across the financial crisis from multiple core-periphery analysis"
} | null | null | [
"Quantitative Finance"
]
| null | true | null | 970 | null | Validated | null | null |
null | {
"abstract": " We present a novel algorithm that uses exact learning and abstraction to\nextract a deterministic finite automaton describing the state dynamics of a\ngiven trained RNN. We do this using Angluin's L* algorithm as a learner and the\ntrained RNN as an oracle. Our technique efficiently extracts accurate automata\nfrom trained RNNs, even when the state vectors are large and require fine\ndifferentiation.\n",
"title": "Extracting Automata from Recurrent Neural Networks Using Queries and Counterexamples"
} | null | null | [
"Computer Science"
]
| null | true | null | 971 | null | Validated | null | null |
null | {
"abstract": " Everything in the world is being connected, and things are becoming\ninteractive. The future of the interactive world depends on the future Internet\nof Things (IoT). Software-defined networking (SDN) technology, a new paradigm\nin the networking area, can be useful in creating an IoT because it can handle\ninteractivity by controlling physical devices, transmission of data among them,\nand data acquisition. However, digital signage can be one of the promising\ntechnologies in this era of technology that is progressing toward the\ninteractive world, connecting users to the IoT network through device-to-device\ncommunication technology. This article illustrates a novel prototype that is\nmainly focused on a smart digital signage system comprised of software-defined\nIoT (SD-IoT) and invisible image sensor communication technology. We have\nproposed an SDN scheme with a view to initiating its flexibility and\ncompatibility for an IoT network-based smart digital signage system. The idea\nof invisible communication can make the users of the technology trendier to it,\nand the usage of unused resources such as images and videos can be ensured. In\naddition, this communication has paved the way for interactivity between the\nuser and digital signage, where the digital signage and the camera of a\nsmartphone can be operated as a transmitter and a receiver, respectively. The\nproposed scheme might be applicable to real-world applications because SDN has\nthe flexibility to adapt with the alteration of network status without any\nhardware modifications while displays and smartphones are available everywhere.\nA performance analysis of this system showed the advantages of an SD-IoT\nnetwork over an Internet protocol-based IoT network considering a queuing\nanalysis for a dynamic link allocation process in the case of user access to\nthe IoT network.\n",
"title": "Performance analysis of smart digital signage system based on software-defined IoT and invisible image sensor communication"
} | null | null | null | null | true | null | 972 | null | Default | null | null |
null | {
"abstract": " Recently, an Atacama Large Millimeter/submillimeter Array (ALMA) observation\nof the water snow line in the protoplanetary disk around the FU Orionis star\nV883 Ori was reported. The radial variation of the spectral index at\nmm-wavelengths around the snow line was interpreted as being due to a pileup of\nparticles interior to the snow line. However, radial transport of solids in the\nouter disk operates on timescales much longer than the typical timescale of an\nFU Ori outburst ($10^{1}$--$10^{2}$ yr). Consequently, a steady-state pileup is\nunlikely. We argue that it is only necessary to consider water evaporation and\nre-coagulation of silicates to explain the recent ALMA observation of V883 Ori\nbecause these processes are short enough to have had their impact since the\noutburst. Our model requires the inner disk to have already been optically\nthick before the outburst, and our results suggest that the carbon content of\npebbles is low.\n",
"title": "What pebbles are made of: Interpretation of the V883 Ori disk"
} | null | null | null | null | true | null | 973 | null | Default | null | null |
null | {
"abstract": " The existence of weak solutions to the stationary Navier-Stokes equations in\nthe whole plane $\\mathbb{R}^2$ is proven. This particular geometry was the only\ncase left open since the work of Leray in 1933. The reason is that due to the\nabsence of boundaries the local behavior of the solutions cannot be controlled\nby the enstrophy in two dimensions. We overcome this difficulty by constructing\napproximate weak solutions having a prescribed mean velocity on some given\nbounded set. As a corollary, we obtain infinitely many weak solutions in\n$\\mathbb{R}^2$ parameterized by this mean velocity, which is reminiscent of the\nexpected convergence of the velocity field at large distances to any prescribed\nconstant vector field. This explicit parameterization of the weak solutions\nallows us to prove a weak-strong uniqueness theorem for small data. The\nquestion of the asymptotic behavior of the weak solutions remains however open,\nwhen the uniqueness theorem doesn't apply.\n",
"title": "Existence and uniqueness of steady weak solutions to the Navier-Stokes equations in $\\mathbb{R}^2$"
} | null | null | null | null | true | null | 974 | null | Default | null | null |
null | {
"abstract": " New results on the Baire product problem are presented. It is shown that an\narbitrary product of almost locally ccc Baire spaces is Baire; moreover, the\nproduct of a Baire space and a 1st countable space which is $\\beta$-unfavorable\nin the strong Choquet game is Baire.\n",
"title": "More on products of Baire spaces"
} | null | null | null | null | true | null | 975 | null | Default | null | null |
null | {
"abstract": " In the Ultimatum Game (UG) one player, named \"proposer\", has to decide how to\nallocate a certain amount of money between herself and a \"responder\". If the\noffer is greater than or equal to the responder's minimum acceptable offer\n(MAO), then the money is split as proposed, otherwise, neither the proposer nor\nthe responder get anything. The UG has intrigued generations of behavioral\nscientists because people in experiments blatantly violate the equilibrium\npredictions that self-interested proposers offer the minimum available non-zero\namount, and self-interested responders accept. Why are these predictions\nviolated? Previous research has mainly focused on the role of social\npreferences. Little is known about the role of general moral preferences for\ndoing the right thing, preferences that have been shown to play a major role in\nother social interactions (e.g., Dictator Game and Prisoner's Dilemma). Here I\ndevelop a theoretical model and an experiment designed to pit social\npreferences against moral preferences. I find that, although people recognize\nthat offering half and rejecting low offers are the morally right things to do,\nmoral preferences have no causal impact on UG behavior. The experimental data\nare indeed well fit by a model according to which: (i) high UG offers are\nmotivated by inequity aversion and, to a lesser extent, self-interest; (ii)\nhigh MAOs are motivated by inequity aversion.\n",
"title": "Social versus Moral preferences in the Ultimatum Game: A theoretical model and an experiment"
} | null | null | null | null | true | null | 976 | null | Default | null | null |
null | {
"abstract": " In this work, we study the tradeoffs between the error probabilities of\nclassical-quantum channels and the blocklength $n$ when the transmission rates\napproach the channel capacity at a rate slower than $1/\\sqrt{n}$, a research\ntopic known as moderate deviation analysis. We show that the optimal error\nprobability vanishes under this rate convergence. Our main technical\ncontributions are a tight quantum sphere-packing bound, obtained via Chaganty\nand Sethuraman's concentration inequality in strong large deviation theory, and\nasymptotic expansions of error-exponent functions. Moderate deviation analysis\nfor quantum hypothesis testing is also established. The converse directly\nfollows from our channel coding result, while the achievability relies on a\nmartingale inequality.\n",
"title": "Moderate Deviation Analysis for Classical-Quantum Channels and Quantum Hypothesis Testing"
} | null | null | [
"Computer Science"
]
| null | true | null | 977 | null | Validated | null | null |
null | {
"abstract": " Constructing tests or confidence regions that control over the error rates in\nthe long-run is probably one of the most important problem in statistics. Yet,\nthe theoretical justification for most methods in statistics is asymptotic. The\nbootstrap for example, despite its simplicity and its widespread usage, is an\nasymptotic method. There are in general no claim about the exactness of\ninferential procedures in finite sample. In this paper, we propose an\nalternative to the parametric bootstrap. We setup general conditions to\ndemonstrate theoretically that accurate inference can be claimed in finite\nsample.\n",
"title": "A simple recipe for making accurate parametric inference in finite sample"
} | null | null | null | null | true | null | 978 | null | Default | null | null |
null | {
"abstract": " Generalized Bäcklund-Darboux transformations (GBDTs) of discrete\nskew-selfadjoint Dirac systems have been successfully used for explicit solving\nof direct and inverse problems of Weyl-Titchmarsh theory. During explicit\nsolving of the direct and inverse problems, we considered GBDTs of the trivial\ninitial systems. However, GBDTs of arbitrary discrete skew-selfadjoint Dirac\nsystems are important as well and we introduce these transformations in the\npresent paper. The obtained results are applied to the construction of explicit\nsolutions of the interesting related non-stationary systems.\n",
"title": "GBDT of discrete skew-selfadjoint Dirac systems and explicit solutions of the corresponding non-stationary problems"
} | null | null | null | null | true | null | 979 | null | Default | null | null |
null | {
"abstract": " We describe a procedure called panel collapse for replacing a CAT(0) cube\ncomplex $\\Psi$ by a \"lower complexity\" CAT(0) cube complex $\\Psi_\\bullet$\nwhenever $\\Psi$ contains a codimension-$2$ hyperplane that is extremal in one\nof the codimension-$1$ hyperplanes containing it. Although $\\Psi_\\bullet$ is\nnot in general a subcomplex of $\\Psi$, it is a subspace consisting of a\nsubcomplex together with some cubes that sit inside $\\Psi$ \"diagonally\". The\nhyperplanes of $\\Psi_\\bullet$ extend to hyperplanes of $\\Psi$. Applying this\nprocedure, we prove: if a group $G$ acts cocompactly on a CAT(0) cube complex\n$\\Psi$, then there is a CAT(0) cube complex $\\Omega$ so that $G$ acts\ncocompactly on $\\Omega$ and for each hyperplane $H$ of $\\Omega$, the stabiliser\nin $G$ of $H$ acts on $H$ essentially.\nUsing panel collapse, we obtain a new proof of Stallings's theorem on groups\nwith more than one end. As another illustrative example, we show that panel\ncollapse applies to the exotic cubulations of free groups constructed by Wise.\nNext, we show that the CAT(0) cube complexes constructed by Cashen-Macura can\nbe collapsed to trees while preserving all of the necessary group actions. (It\nalso illustrates that our result applies to actions of some non-discrete\ngroups.) We also discuss possible applications to quasi-isometric rigidity for\ncertain classes of graphs of free groups with cyclic edge groups. Panel\ncollapse is also used in forthcoming work of the first-named author and Wilton\nto study fixed-point sets of finite subgroups of $\\mathrm{Out}(F_n)$ on the\nfree splitting complex.\n",
"title": "Panel collapse and its applications"
} | null | null | [
"Mathematics"
]
| null | true | null | 980 | null | Validated | null | null |
null | {
"abstract": " The recently proposed self-ensembling methods have achieved promising results\nin deep semi-supervised learning, which penalize inconsistent predictions of\nunlabeled data under different perturbations. However, they only consider\nadding perturbations to each single data point, while ignoring the connections\nbetween data samples. In this paper, we propose a novel method, called Smooth\nNeighbors on Teacher Graphs (SNTG). In SNTG, a graph is constructed based on\nthe predictions of the teacher model, i.e., the implicit self-ensemble of\nmodels. Then the graph serves as a similarity measure with respect to which the\nrepresentations of \"similar\" neighboring points are learned to be smooth on the\nlow-dimensional manifold. We achieve state-of-the-art results on\nsemi-supervised learning benchmarks. The error rates are 9.89%, 3.99% for\nCIFAR-10 with 4000 labels, SVHN with 500 labels, respectively. In particular,\nthe improvements are significant when the labels are fewer. For the\nnon-augmented MNIST with only 20 labels, the error rate is reduced from\nprevious 4.81% to 1.36%. Our method also shows robustness to noisy labels.\n",
"title": "Smooth Neighbors on Teacher Graphs for Semi-supervised Learning"
} | null | null | null | null | true | null | 981 | null | Default | null | null |
null | {
"abstract": " Efficient algorithms and techniques to detect and identify large flows in a\nhigh throughput traffic stream in the SDN match-and-action model are presented.\nThis is in contrast to previous work that either deviated from the match and\naction model by requiring additional switch level capabilities or did not\nexploit the SDN data plane. Our construction has two parts; (a) how to sample\nin an SDN match and action model, (b) how to detect large flows efficiently and\nin a scalable way, in the SDN model.\nOur large flow detection methods provide high accuracy and present a good and\npractical tradeoff between switch - controller traffic, and the number of\nentries required in the switch flow table. Based on different parameters, we\ndifferentiate between heavy flows, elephant flows and bulky flows and present\nefficient algorithms to detect flows of the different types.\nAdditionally, as part of our heavy flow detection scheme, we present sampling\nmethods to sample packets with arbitrary probability $p$ per packet or per byte\nthat traverses an SDN switch.\nFinally, we show how our algorithms can be adapted to a distributed\nmonitoring SDN setting with multiple switches, and easily scale with the number\nof monitoring switches.\n",
"title": "Detecting Heavy Flows in the SDN Match and Action Model"
} | null | null | null | null | true | null | 982 | null | Default | null | null |
null | {
"abstract": " With the tremendous increase of the Internet traffic, achieving the best\nperformance with limited resources is becoming an extremely urgent problem. In\norder to address this concern, in this paper, we build an optimization problem\nwhich aims to maximize the total utility of traffic flows with the capacity\nconstraint of nodes and links in the network. Based on Duality Theory, we\npropose an iterative algorithm which adjusts the rates of traffic flows and\ncapacity of nodes and links simultaneously to maximize the total utility.\nSimulation results show that our algorithm performs better than the NUP\nalgorithm on BA and ER network models, which has shown to get the best\nperformance so far. Since our research combines the topology information with\ncapacity constraint, it may give some insights for resource allocation in real\ncommunication networks.\n",
"title": "Optimal Resource Allocation with Node and Link Capacity Constraints in Complex Networks"
} | null | null | null | null | true | null | 983 | null | Default | null | null |
null | {
"abstract": " We consider a fundamental integer programming (IP) model for cost-benefit\nanalysis flood protection through dike building in the Netherlands, due to\nVerweij and Zwaneveld.\nExperimental analysis with data for the Ijsselmeer lead to integral optimal\nsolution of the linear programming relaxation of the IP model.\nThis naturally led to the question of integrality of the polytope associated\nwith the IP model.\nIn this paper we first give a negative answer to this question by\nestablishing non-integrality of the polytope.\nSecond, we establish natural conditions that guarantee the linear programming\nrelaxation of the IP model to be integral.\nWe then test the most recent data on flood probabilities, damage and\ninvestment costs of the IJsselmeer for these conditions.\nThird, we show that the IP model can be solved in polynomial time when the\nnumber of dike segments, or the number of feasible barrier heights, are\nconstant.\n",
"title": "On the complexity of solving a decision problem with flow-depending costs: the case of the IJsselmeer dikes"
} | null | null | null | null | true | null | 984 | null | Default | null | null |
null | {
"abstract": " In wireless communication, heterogeneous technologies such as WiFi, ZigBee\nand BlueTooth operate in the same ISM band.With the exponential growth in the\nnumber of wireless devices, the ISM band becomes more and more crowded. These\nheterogeneous devices have to compete with each other to access spectrum\nresources, generating cross-technology interference (CTI). Since CTI may\ndestroy wireless communication, this field is facing an urgent and challenging\nneed to investigate spectrum efficiency under CTI. In this paper, we introduce\na novel framework to address this problem from two aspects. On the one hand,\nfrom the perspective of each communication technology itself, we propose novel\nchannel/link models to capture the channel/link status under CTI. On the other\nhand, we investigate spectrum efficiency from the perspective by taking all\nheterogeneous technologies as a whole and building crosstechnology\ncommunication among them. The capability of direct communication among\nheterogeneous devices brings great opportunities to harmoniously sharing the\nspectrum with collaboration rather than competition.\n",
"title": "Achieving Spectrum Efficient Communication Under Cross-Technology Interference"
} | null | null | null | null | true | null | 985 | null | Default | null | null |
null | {
"abstract": " We have derived background corrected intensities of 3-50 MeV galactic\nelectrons observed by Voyager 1 as it passes through the heliosheath from 95 to\n122 AU. The overall intensity change of the background corrected data from the\ninner to the outer boundary of the heliosheath is a maximum of a factor ~100 at\n15 MeV. At lower energies this fractional change becomes less and the corrected\nelectron spectra in the heliosheath becomes progressively steeper, reaching\nvalues ~ -2.5 for the spectral index just outside of the termination shock. At\nhigher energies the spectra of electrons has an exponent changing from the\nnegative LIS spectral index of -1.3 to values approaching zero in the\nheliosheath as a result of the solar modulation of the galactic electron\ncomponent. The large modulation effects observed below ~100 MV are possible\nevidence for enhanced diffusion as part of the modulation process for electrons\nin the heliosheath.\n",
"title": "A Galactic Cosmic Ray Electron Intensity Increase of a factor of up to 100 At Energies between 3 and 50 MeV in the Heliosheath between the Termination Shock and the Heliopause Due to Solar Modulation As Measured by Voyager 1"
} | null | null | null | null | true | null | 986 | null | Default | null | null |
null | {
"abstract": " The `beta' is one of the key quantities in the capital asset pricing model\n(CAPM). In statistical language, the beta can be viewed as the slope of the\nregression line fitted to financial returns on the market against the returns\non the asset under consideration. The insurance counterpart of CAPM, called the\nweighted insurance pricing model (WIPM), gives rise to the so-called\nweighted-Gini beta. The aforementioned two betas may or may not coincide,\ndepending on the form of the underlying regression function, and this has\nprofound implications when designing portfolios and allocating risk capital. To\nfacilitate these tasks, in this paper we develop large-sample statistical\ninference results that, in a straightforward fashion, imply confidence\nintervals for, and hypothesis tests about, the equality of the two betas.\n",
"title": "Statistical foundations for assessing the difference between the classical and weighted-Gini betas"
} | null | null | null | null | true | null | 987 | null | Default | null | null |
null | {
"abstract": " In partially observed environments, it can be useful for a human to provide\nthe robot with declarative information that represents probabilistic relational\nconstraints on properties of objects in the world, augmenting the robot's\nsensory observations. For instance, a robot tasked with a search-and-rescue\nmission may be informed by the human that two victims are probably in the same\nroom. An important question arises: how should we represent the robot's\ninternal knowledge so that this information is correctly processed and combined\nwith raw sensory information? In this paper, we provide an efficient belief\nstate representation that dynamically selects an appropriate factoring,\ncombining aspects of the belief when they are correlated through information\nand separating them when they are not. This strategy works in open domains, in\nwhich the set of possible objects is not known in advance, and provides\nsignificant improvements in inference time over a static factoring, leading to\nmore efficient planning for complex partially observed tasks. We validate our\napproach experimentally in two open-domain planning problems: a 2D discrete\ngridworld task and a 3D continuous cooking task. A supplementary video can be\nfound at this http URL.\n",
"title": "Integrating Human-Provided Information Into Belief State Representation Using Dynamic Factorization"
} | null | null | null | null | true | null | 988 | null | Default | null | null |
null | {
"abstract": " Open problems abound in the theory of complex networks, which has found\nsuccessful application to diverse fields of science. With the aim of further\nadvancing the understanding of the brain's functional connectivity, we propose\nto evaluate a network metric which we term the geodesic entropy. This entropy,\nin a way that can be made precise, quantifies the Shannon entropy of the\ndistance distribution to a specific node from all other nodes. Measurements of\ngeodesic entropy allow for the characterization of the structural information\nof a network that takes into account the distinct role of each node into the\nnetwork topology. The measurement and characterization of this structural\ninformation has the potential to greatly improve our understanding of sustained\nactivity and other emergent behaviors in networks, such as self-organized\ncriticality sometimes seen in such contexts. We apply these concepts and\nmethods to study the effects of how the psychedelic Ayahuasca affects the\nfunctional connectivity of the human brain. We show that the geodesic entropy\nis able to differentiate the functional networks of the human brain in two\ndifferent states of consciousness in the resting state: (i) the ordinary waking\nstate and (ii) a state altered by ingestion of the Ayahuasca. The entropy of\nthe nodes of brain networks from subjects under the influence of Ayahuasca\ndiverge significantly from those of the ordinary waking state. The functional\nbrain networks from subjects in the altered state have, on average, a larger\ngeodesic entropy compared to the ordinary state. We conclude that geodesic\nentropy is a useful tool for analyzing complex networks and discuss how and why\nit may bring even further valuable insights into the study of the human brain\nand other empirical networks.\n",
"title": "Characterizing complex networks using Entropy-degree diagrams: unveiling changes in functional brain connectivity induced by Ayahuasca"
} | null | null | null | null | true | null | 989 | null | Default | null | null |
null | {
"abstract": " The high-energy non-thermal universe is dominated by power law-like spectra.\nTherefore results in high-energy astronomy are often reported as parameters of\npower law fits, or, in the case of a non-detection, as an upper limit assuming\nthe underlying unseen spectrum behaves as a power law. In this paper I\ndemonstrate a simple and powerful one-to-one relation of the integral upper\nlimit in the two dimensional power law parameter space into the spectrum\nparameter space and use this method to unravel the so far convoluted question\nof the sensitivity of astroparticle telescopes.\n",
"title": "On Integral Upper Limits Assuming Power Law Spectra and the Sensitivity in High-Energy Astronomy"
} | null | null | null | null | true | null | 990 | null | Default | null | null |
null | {
"abstract": " We develop estimates for the solutions and derive existence and uniqueness\nresults of various local boundary value problems for Dirac equations that\nimprove all relevant results known in the literature. With these estimates at\nhand, we derive a general existence, uniqueness and regularity theorem for\nsolutions of Dirac equations with such boundary conditions. We also apply these\nestimates to a new nonlinear elliptic-parabolic problem, the Dirac-harmonic\nheat flow on Riemannian spin manifolds. This problem is motivated by the\nsupersymmetric nonlinear $\\sigma$-model and combines a harmonic heat flow type\nequation with a Dirac equation that depends nonlinearly on the flow.\n",
"title": "Estimates for solutions of Dirac equations and an application to a geometric elliptic-parabolic problem"
} | null | null | null | null | true | null | 991 | null | Default | null | null |
null | {
"abstract": " In modern election campaigns, political parties utilize social media to\nadvertise their policies and candidates and to communicate to the electorate.\nIn Japan's latest general election in 2017, the 48th general election for the\nLower House, social media, especially Twitter, was actively used. In this\npaper, we analyze the users who retweeted tweets of political parties on\nTwitter during the election. Our aim is to clarify what kinds of users are\ndiffusing (retweeting) tweets of political parties. The results indicate that\nthe characteristics of retweeters of the largest ruling party (Liberal\nDemocratic Party of Japan) and the largest opposition party (The Constitutional\nDemocratic Party of Japan) were similar, even though the retweeters did not\noverlap each other. We also found that a particular opposition party (Japanese\nCommunist Party) had quite different characteristics from other political\nparties.\n",
"title": "Analysis of Political Party Twitter Accounts' Retweeters During Japan's 2017 Election"
} | null | null | null | null | true | null | 992 | null | Default | null | null |
null | {
"abstract": " In 1840 Jacob Steiner on Christian Rudolf's request proved that a triangle\nwith two equal bisectors is isosceles. But what about changing the bisectors to\ncevians? Cevian is any line segment in a triangle with one endpoint on a vertex\nof the triangle and other endpoint on the opposite side. Not for any pairs of\nequal cevians the triangle is isosceles. Theorem. If for a triangle ABC there\nare equal cevians issuing from A and B, which intersect on the bisector or on\nthe median of the angle C, then AC=BC (so the triangle ABC is isosceles).\nProposition. Let ABC be an isosceles triangle. Define circle C to be the circle\nsymmetric relative to AB to the circumscribed circle of the triangle ABC. Then\nthe locus of intersection points of pairs of equal cevians is the union of the\nbase AB, the triangle's axis of symmetry, and the circle C.\n",
"title": "When a triangle is isosceles?"
} | null | null | [
"Mathematics"
]
| null | true | null | 993 | null | Validated | null | null |
null | {
"abstract": " Anomaly detecting as an important technical in cloud computing is applied to\nsupport smooth running of the cloud platform. Traditional detecting methods\nbased on statistic, analysis, etc. lead to the high false-alarm rate due to\nnon-adaptive and sensitive parameters setting. We presented an online model for\nanomaly detecting using machine learning theory. However, most existing methods\nbased on machine learning linked all features from difference sub-systems into\na long feature vector directly, which is difficult to both exploit the\ncomplement information between sub-systems and ignore multi-view features\nenhancing the classification performance. Aiming to this problem, the proposed\nmethod automatic fuses multi-view features and optimize the discriminative\nmodel to enhance the accuracy. This model takes advantage of extreme learning\nmachine (ELM) to improve detection efficiency. ELM is the single hidden layer\nneural network, which is transforming iterative solution the output weights to\nsolution of linear equations and avoiding the local optimal solution. Moreover,\nwe rank anomies according to the relationship between samples and the\nclassification boundary, and then assigning weights for ranked anomalies,\nretraining the classification model finally. Our method exploits the complement\ninformation between sub-systems sufficiently, and avoids the influence from\nimbalance dataset, therefore, deal with various challenges from the cloud\ncomputing platform. We deploy the privately cloud platform by Openstack,\nverifying the proposed model and comparing results to the state-of-the-art\nmethods with better efficiency and simplicity.\n",
"title": "Anomaly detecting and ranking of the cloud computing platform by multi-view learning"
} | null | null | null | null | true | null | 994 | null | Default | null | null |
null | {
"abstract": " We survey the dimension theory of self-affine sets for general mathematical\naudience. The article is in Finnish.\n",
"title": "On self-affine sets"
} | null | null | null | null | true | null | 995 | null | Default | null | null |
null | {
"abstract": " Buoyancy-thermocapillary convection in a layer of volatile liquid driven by a\nhorizontal temperature gradient arises in a variety of situations. Recent\nstudies have shown that the composition of the gas phase, which is typically a\nmixture of vapour and air, has a noticeable effect on the critical Marangoni\nnumber describing the onset of convection as well as on the observed convection\npattern. Specifically, as the total pressure or, equivalently, the average\nconcentration of air is decreased, the threshold of the instability leading to\nthe emergence of convective rolls is found to increase rather significantly. We\npresent a linear stability analysis of the problem which shows that this trend\ncan be readily understood by considering the transport of heat and vapour\nthrough the gas phase. In particular, we show that transport in the gas phase\nhas a noticeable effect even at atmospheric conditions, when phase change is\ngreatly suppressed.\n",
"title": "The effect of phase change on stability of convective flow in a layer of volatile liquid driven by a horizontal temperature gradient"
} | null | null | null | null | true | null | 996 | null | Default | null | null |
null | {
"abstract": " In the paper we consider a graph model of message passing processes and\npresent a method verification of message passing processes. The method is\nillustrated by an example of a verification of sliding window protocol.\n",
"title": "A graph model of message passing processes"
} | null | null | null | null | true | null | 997 | null | Default | null | null |
null | {
"abstract": " In this article, we derive a Bayesian model to learning the sparse and low\nrank PARAFAC decomposition for the observed tensor with missing values via the\nelastic net, with property to find the true rank and sparse factor matrix which\nis robust to the noise. We formulate efficient block coordinate descent\nalgorithm and admax stochastic block coordinate descent algorithm to solve it,\nwhich can be used to solve the large scale problem. To choose the appropriate\nrank and sparsity in PARAFAC decomposition, we will give a solution path by\ngradually increasing the regularization to increase the sparsity and decrease\nthe rank. When we find the sparse structure of the factor matrix, we can fixed\nthe sparse structure, using a small to regularization to decreasing the\nrecovery error, and one can choose the proper decomposition from the solution\npath with sufficient sparse factor matrix with low recovery error. We test the\npower of our algorithm on the simulation data and real data, which show it is\npowerful.\n",
"title": "Learning the Sparse and Low Rank PARAFAC Decomposition via the Elastic Net"
} | null | null | null | null | true | null | 998 | null | Default | null | null |
null | {
"abstract": " We introduce a new family of thermostat flows on the unit tangent bundle of\nan oriented Riemannian $2$-manifold. Suitably reparametrised, these flows\ninclude the geodesic flow of metrics of negative Gauss curvature and the\ngeodesic flow induced by the Hilbert metric on the quotient surface of\ndivisible convex sets. We show that the family of flows can be parametrised in\nterms of certain weighted holomorphic differentials and investigate their\nproperties. In particular, we prove that they admit a dominated splitting and\nwe identify special cases in which the flows are Anosov. In the latter case, we\nstudy when they admit an invariant measure in the Lebesgue class and the\nregularity of the weak foliations.\n",
"title": "Holomorphic differentials, thermostats and Anosov flows"
} | null | null | null | null | true | null | 999 | null | Default | null | null |
null | {
"abstract": " We introduce the shifted quantum affine algebras. They map homomorphically\ninto the quantized $K$-theoretic Coulomb branches of $3d\\ {\\mathcal N}=4$ SUSY\nquiver gauge theories. In type $A$, they are endowed with a coproduct, and they\nact on the equivariant $K$-theory of parabolic Laumon spaces. In type $A_1$,\nthey are closely related to the open relativistic quantum Toda lattice of type\n$A$.\n",
"title": "Multiplicative slices, relativistic Toda and shifted quantum affine algebras"
} | null | null | null | null | true | null | 1000 | null | Default | null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.