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": " In this paper we analyse the benefits of incorporating interval-valued fuzzy\nsets into the Bousi-Prolog system. A syntax, declarative semantics and im-\nplementation for this extension is presented and formalised. We show, by using\npotential applications, that fuzzy logic programming frameworks enhanced with\nthem can correctly work together with lexical resources and ontologies in order\nto improve their capabilities for knowledge representation and reasoning.\n",
"title": "On the incorporation of interval-valued fuzzy sets into the Bousi-Prolog system: declarative semantics, implementation and applications"
} | null | null | null | null | true | null | 2201 | null | Default | null | null |
null | {
"abstract": " Neutron beam monitors with high efficiency, low gamma sensitivity, high time\nand space resolution are required in neutron beam experiments to continuously\ndiagnose the delivered beam. In this work, commercially available neutron beam\nmonitors have been characterized using the R2D2 beamline at IFE (Norway) and\nusing a Be-based neutron source. For the gamma sensitivity measurements\ndifferent gamma sources have been used. The evaluation of the monitors\nincludes, the study of their efficiency, attenuation, scattering and\nsensitivity to gamma. In this work we report the results of this\ncharacterization.\n",
"title": "Characterization of Thermal Neutron Beam Monitors"
} | null | null | [
"Physics"
]
| null | true | null | 2202 | null | Validated | null | null |
null | {
"abstract": " Previous studies have shown the filamentary structures in the cosmic web\ninfluence the alignments of nearby galaxies. We study this effect in the LOWZ\nsample of the Sloan Digital Sky Survey using the \"Cosmic Web Reconstruction\"\nfilament catalogue of Chen et al. (2016). We find that LOWZ galaxies exhibit a\nsmall but statistically significant alignment in the direction parallel to the\norientation of nearby filaments. This effect is detectable even in the absence\nof nearby galaxy clusters, which suggests it is an effect from the matter\ndistribution in the filament. A nonparametric regression model suggests that\nthe alignment effect with filaments extends over separations of $30-40$ Mpc. We\nfind that galaxies that are bright and early-forming align more strongly with\nthe directions of nearby filaments than those that are faint and late-forming;\nhowever, trends with stellar mass are less statistically significant, within\nthe narrow range of stellar mass of this sample.\n",
"title": "Detecting Galaxy-Filament Alignments in the Sloan Digital Sky Survey III"
} | null | null | null | null | true | null | 2203 | null | Default | null | null |
null | {
"abstract": " Pandeia is the exposure time calculator (ETC) system developed for the James\nWebb Space Telescope (JWST) that will be used for creating JWST proposals. It\nincludes a simulation-hybrid Python engine that calculates the two-dimensional\npixel-by-pixel signal and noise properties of the JWST instruments. This allows\nfor appropriate handling of realistic point spread functions, MULTIACCUM\ndetector readouts, correlated detector readnoise, and multiple photometric and\nspectral extraction strategies. Pandeia includes support for all the JWST\nobserving modes, including imaging, slitted/slitless spectroscopy, integral\nfield spectroscopy, and coronagraphy. Its highly modular, data-driven design\nmakes it easily adaptable to other observatories. An implementation for use\nwith WFIRST is also available.\n",
"title": "Pandeia: A Multi-mission Exposure Time Calculator for JWST and WFIRST"
} | null | null | null | null | true | null | 2204 | null | Default | null | null |
null | {
"abstract": " A basic combinatorial invariant of a convex polytope $P$ is its $f$-vector\n$f(P)=(f_0,f_1,\\dots,f_{\\dim P-1})$, where $f_i$ is the number of\n$i$-dimensional faces of $P$. Steinitz characterized all possible $f$-vectors\nof $3$-polytopes and Grünbaum characterized the pairs given by the first two\nentries of the $f$-vectors of $4$-polytopes. In this paper, we characterize the\npairs given by the first two entries of the $f$-vectors of $5$-polytopes. The\nsame result was also proved by Pineda-Villavicencio, Ugon and Yost\nindependently.\n",
"title": "The numbers of edges of 5-polytopes with a given number of vertices"
} | null | null | null | null | true | null | 2205 | null | Default | null | null |
null | {
"abstract": " Recently, the integration of geographical coordinates into a picture has\nbecome more and more popular. Indeed almost all smartphones and many cameras\ntoday have a built-in GPS receiver that stores the location information in the\nExif header when a picture is taken. Although the automatic embedding of\ngeotags in pictures is often ignored by smart phone users as it can lead to\nendless discussions about privacy implications, these geotags could be really\nuseful for investigators in analysing criminal activity. Currently, there are\nmany free tools as well as commercial tools available in the market that can\nhelp computer forensics investigators to cover a wide range of geographic\ninformation related to criminal scenes or activities. However, there are not\nspecific forensic tools available to deal with the geolocation of pictures\ntaken by smart phones or cameras. In this paper, we propose and develop an\nimage scanning and mapping tool for investigators. This tool scans all the\nfiles in a given directory and then displays particular photos based on\noptional filters (date, time, device, localisation) on Google Map. The file\nscanning process is not based on the file extension but its header. This tool\ncan also show efficiently to users if there is more than one image on the map\nwith the same GPS coordinates, or even if there are images with no GPS\ncoordinates taken by the same device in the same timeline. Moreover, this new\ntool is portable; investigators can run it on any operating system without any\ninstallation. Another useful feature is to be able to work in a read-only\nenvironment, so that forensic results will not be modified. We also present and\nevaluate this tool real world application in this paper.\n",
"title": "MapExif: an image scanning and mapping tool for investigators"
} | null | null | null | null | true | null | 2206 | null | Default | null | null |
null | {
"abstract": " Fitch-style modal deduction, in which modalities are eliminated by opening a\nsubordinate proof, and introduced by shutting one, were investigated in the\n1990s as a basis for lambda calculi. We show that such calculi have good\ncomputational properties for a variety of intuitionistic modal logics.\nSemantics are given in cartesian closed categories equipped with an adjunction\nof endofunctors, with the necessity modality interpreted by the right adjoint.\nWhere this functor is an idempotent comonad, a coherence result on the\nsemantics allows us to present a calculus for intuitionistic S4 that is simpler\nthan others in the literature. We show the calculi can be extended à la\ntense logic with the left adjoint of necessity, and are then complete for the\ncategorical semantics.\n",
"title": "Fitch-Style Modal Lambda Calculi"
} | null | null | [
"Computer Science"
]
| null | true | null | 2207 | null | Validated | null | null |
null | {
"abstract": " Recently, the first installment of data from ESA's Gaia astrometric satellite\nmission (Gaia-DR1) was released, containing positions of more than 1 billion\nstars with unprecedented precision, as well as only proper motions and\nparallaxes, however only for a subset of 2 million objects. The second release,\ndue in late 2017 or early 2018, will include those quantities for most objects.\nIn order to provide a dataset that bridges the time gap between the Gaia-DR1\nand Gaia-DR2 releases and partly remedies the lack of proper motions in the\nformer, HSOY (\"Hot Stuff for One Year\") was created as a hybrid catalogue\nbetween Gaia-DR1 and ground-based astrometry, featuring proper motions (but no\nparallaxes) for a large fraction of the DR1 objects. While not attempting to\ncompete with future Gaia releases in terms of data quality or number of\nobjects, the aim of HSOY is to provide improved proper motions partly based on\nGaia data, allowing some studies to be carried out just now or as pilot studies\nfor later larger projects requiring higher-precision data. The HSOY catalogue\nwas compiled using the positions taken from Gaia-DR1 combined with the input\ndata from the PPMXL catalogue, employing the same weighted least-squares\ntechnique that was used to assemble the PPMXL catalogue itself. Results. This\neffort resulted in a four-parameter astrometric catalogue containing\n583,000,000 objects, with Gaia-DR1 quality positions and proper motions with\nprecisions from significantly less than 1 mas/yr to 5 mas/yr, depending on the\nobject's brightness and location on the sky.\n",
"title": "Hot Stuff for One Year (HSOY) - A 583 million star proper motion catalogue derived from Gaia DR1 and PPMXL"
} | null | null | null | null | true | null | 2208 | null | Default | null | null |
null | {
"abstract": " We consider randomly distributed mixtures of bonds of ferromagnetic and\nantiferromagnetic type in a two-dimensional square lattice with probability\n$1-p$ and $p$, respectively, according to an i.i.d. random variable. We study\nminimizers of the corresponding nearest-neighbour spin energy on large domains\nin ${\\mathbb Z}^2$. We prove that there exists $p_0$ such that for $p\\le p_0$\nsuch minimizers are characterized by a majority phase; i.e., they take\nidentically the value $1$ or $-1$ except for small disconnected sets. A\ndeterministic analogue is also proved.\n",
"title": "Asymptotic behaviour of ground states for mixtures of ferromagnetic and antiferromagnetic interactions in a dilute regime"
} | null | null | null | null | true | null | 2209 | null | Default | null | null |
null | {
"abstract": " In this paper we propose a general framework for modeling an insurance\nclaims' information flow in continuous time, by generalizing the reduced-form\nframework for credit risk and life insurance. In particular, we assume a\nnontrivial dependence structure between the reference filtration and the\ninsurance internal filtration. We apply these results for pricing non-life\ninsurance liabilities in hybrid financial and insurance markets, while taking\ninto account the role of inflation under the benchmark approach. This framework\noffers at the same time a general and flexible structure, and explicit and\ntreatable pricing formula.\n",
"title": "Extended Reduced-Form Framework for Non-Life Insurance"
} | null | null | null | null | true | null | 2210 | null | Default | null | null |
null | {
"abstract": " This paper will cover several studies and design changes that will eventually\nbe implemented to the Fermi National Accelerator Laboratory (FNAL) magnetron\nion source. The topics include tungsten cathode insert, solenoid gas valves,\ncurrent controlled arc pulser, cesium boiler redesign, gas mixtures of hydrogen\nand nitrogen, and duty factor reduction. The studies were performed on the FNAL\ntest stand, with the aim to improve source lifetime, stability, and reducing\nthe amount of tuning needed.\n",
"title": "Overview of Recent Studies and Design Changes for the FNAL Magnetron Ion Source"
} | null | null | null | null | true | null | 2211 | null | Default | null | null |
null | {
"abstract": " We consider multidimensional optimization problems that are formulated in the\nframework of tropical mathematics to minimize functions defined on vectors over\na tropical semifield (a semiring with idempotent addition and invertible\nmultiplication). The functions, given by a matrix and calculated through\nmultiplicative conjugate transposition, are nonlinear in the tropical\nmathematics sense. We start with known results on the solution of the problems\nwith irreducible matrices. To solve the problems in the case of arbitrary\n(reducible) matrices, we first derive the minimum value of the objective\nfunction, and find a set of solutions. We show that all solutions of the\nproblem satisfy a system of vector inequalities, and then use these\ninequalities to establish characteristic properties of the solution set.\nFurthermore, all solutions of the problem are represented as a family of\nsubsets, each defined by a matrix that is obtained by using a matrix\nsparsification technique. We describe a backtracking procedure that allows one\nto reduce the brute-force generation of sparsified matrices by skipping those,\nwhich cannot provide solutions, and thus offers an economical way to obtain all\nsubsets in the family. Finally, the characteristic properties of the solution\nset are used to provide complete solutions in a closed form. We illustrate the\nresults obtained with simple numerical examples.\n",
"title": "Complete algebraic solution of multidimensional optimization problems in tropical semifield"
} | null | null | null | null | true | null | 2212 | null | Default | null | null |
null | {
"abstract": " A graph is a powerful concept for representation of relations between pairs\nof entities. Data with underlying graph structure can be found across many\ndisciplines and there is a natural desire for understanding such data better.\nDeep learning (DL) has achieved significant breakthroughs in a variety of\nmachine learning tasks in recent years, especially where data is structured on\na grid, such as in text, speech, or image understanding. However, surprisingly\nlittle has been done to explore the applicability of DL on arbitrary\ngraph-structured data directly.\nThe goal of this thesis is to investigate architectures for DL on graphs and\nstudy how to transfer, adapt or generalize concepts that work well on\nsequential and image data to this domain. We concentrate on two important\nprimitives: embedding graphs or their nodes into a continuous vector space\nrepresentation (encoding) and, conversely, generating graphs from such vectors\nback (decoding). To that end, we make the following contributions.\nFirst, we introduce Edge-Conditioned Convolutions (ECC), a convolution-like\noperation on graphs performed in the spatial domain where filters are\ndynamically generated based on edge attributes. The method is used to encode\ngraphs with arbitrary and varying structure.\nSecond, we propose SuperPoint Graph, an intermediate point cloud\nrepresentation with rich edge attributes encoding the contextual relationship\nbetween object parts. Based on this representation, ECC is employed to segment\nlarge-scale point clouds without major sacrifice in fine details.\nThird, we present GraphVAE, a graph generator allowing us to decode graphs\nwith variable but upper-bounded number of nodes making use of approximate graph\nmatching for aligning the predictions of an autoencoder with its inputs. The\nmethod is applied to the task of molecule generation.\n",
"title": "Deep Learning on Attributed Graphs: A Journey from Graphs to Their Embeddings and Back"
} | null | null | null | null | true | null | 2213 | null | Default | null | null |
null | {
"abstract": " This paper considers the assignment of multiple mobile robots to goal\nlocations under uncertain travel time estimates. Our aim is to produce optimal\nassignments, such that the average waiting time at destinations is minimized.\nOur premise is that time is the most valuable asset in the system. Hence, we\nmake use of redundant robots to counter the effect of uncertainty. Since\nsolving the redundant assignment problem is strongly NP-hard, we exploit\nstructural properties of our problem to propose a polynomial-time, near-optimal\nsolution. We demonstrate that our problem can be reduced to minimizing a\nsupermodular cost function subject to a matroid constraint. This allows us to\ndevelop a greedy algorithm, for which we derive sub-optimality bounds. A\ncomparison with the baseline non-redundant assignment shows that redundant\nassignment reduces the waiting time at goals, and that this performance gap\nincreases as noise increases. Finally, we evaluate our method on a mobility\ndata set (specifying vehicle availability and passenger requests), recorded in\nthe area of Manhattan, New York. Our algorithm performs in real-time, and\nreduces passenger waiting times when travel times are uncertain.\n",
"title": "Supermodular Optimization for Redundant Robot Assignment under Travel-Time Uncertainty"
} | null | null | [
"Computer Science"
]
| null | true | null | 2214 | null | Validated | null | null |
null | {
"abstract": " Cyclic codes have efficient encoding and decoding algorithms over finite\nfields, so that they have practical applications in communication systems,\nconsumer electronics and data storage systems. The objective of this paper is\nto give eight new classes of optimal ternary cyclic codes with parameters\n$[3^m-1,3^m-1-2m,4]$, according to a result on the non-existence of solutions\nto a certain equation over $F_{3^m}$. It is worth noticing that some recent\nconclusions on such optimal ternary cyclic codes are some special cases of our\nwork. More importantly, three of the nine open problems proposed by Ding and\nHelleseth in [8] are solved completely. In addition, another one among the nine\nopen problems is also promoted.\n",
"title": "Several classes of optimal ternary cyclic codes"
} | null | null | null | null | true | null | 2215 | null | Default | null | null |
null | {
"abstract": " We prove that indecomposable $\\Sigma$-pure-injective modules for a string\nalgebra are string or band modules. The key step in our proof is a splitting\nresult for infinite-dimensional linear relations.\n",
"title": "$Σ$-pure-injective modules for string algebras and linear relations"
} | null | null | null | null | true | null | 2216 | null | Default | null | null |
null | {
"abstract": " We present a novel continuous optimization method to the discrete problem of\nquadtree optimization. The optimization aims at achieving a quadtree structure\nwith the highest mechanical stiffness, where the edges in the quadtree are\ninterpreted as structural elements carrying mechanical loads. We formulate\nquadtree optimization as a continuous material distribution problem. The\ndiscrete design variables (i.e., to refine or not to refine) are replaced by\ncontinuous variables on multiple levels in the quadtree hierarchy. In discrete\nquadtree optimization, a cell is only eligible for refinement if its parent\ncell has been refined. We propose a continuous analogue to this dependency for\ncontinuous multi-level design variables, and integrate it in the iterative\noptimization process. Our results show that the continuously optimized quadtree\nstructures perform much stiffer than uniform patterns and the heuristically\noptimized counterparts. We demonstrate the use of adaptive structures as\nlightweight infill for 3D printed parts, where uniform geometric patterns have\nbeen typically used in practice.\n",
"title": "Continuous Optimization of Adaptive Quadtree Structures"
} | null | null | null | null | true | null | 2217 | null | Default | null | null |
null | {
"abstract": " We provide a microeconomic framework for decision trees: a popular machine\nlearning method. Specifically, we show how decision trees represent a\nnon-compensatory decision protocol known as disjunctions-of-conjunctions and\nhow this protocol generalizes many of the non-compensatory rules used in the\ndiscrete choice literature so far. Additionally, we show how existing decision\ntree variants address many economic concerns that choice modelers might have.\nBeyond theoretical interpretations, we contribute to the existing literature of\ntwo-stage, semi-compensatory modeling and to the existing decision tree\nliterature. In particular, we formulate the first bayesian model tree, thereby\nallowing for uncertainty in the estimated non-compensatory rules as well as for\ncontext-dependent preference heterogeneity in one's second-stage choice model.\nUsing an application of bicycle mode choice in the San Francisco Bay Area, we\nestimate our bayesian model tree, and we find that it is over 1,000 times more\nlikely to be closer to the true data-generating process than a multinomial\nlogit model (MNL). Qualitatively, our bayesian model tree automatically finds\nthe effect of bicycle infrastructure investment to be moderated by travel\ndistance, socio-demographics and topography, and our model identifies\ndiminishing returns from bike lane investments. These qualitative differences\nlead to bayesian model tree forecasts that directly align with the observed\nbicycle mode shares in regions with abundant bicycle infrastructure such as\nDavis, CA and the Netherlands. In comparison, MNL's forecasts are overly\noptimistic.\n",
"title": "Machine Learning Meets Microeconomics: The Case of Decision Trees and Discrete Choice"
} | null | null | null | null | true | null | 2218 | null | Default | null | null |
null | {
"abstract": " In recent years, Deep Reinforcement Learning has made impressive advances in\nsolving several important benchmark problems for sequential decision making.\nMany control applications use a generic multilayer perceptron (MLP) for\nnon-vision parts of the policy network. In this work, we propose a new neural\nnetwork architecture for the policy network representation that is simple yet\neffective. The proposed Structured Control Net (SCN) splits the generic MLP\ninto two separate sub-modules: a nonlinear control module and a linear control\nmodule. Intuitively, the nonlinear control is for forward-looking and global\ncontrol, while the linear control stabilizes the local dynamics around the\nresidual of global control. We hypothesize that this will bring together the\nbenefits of both linear and nonlinear policies: improve training sample\nefficiency, final episodic reward, and generalization of learned policy, while\nrequiring a smaller network and being generally applicable to different\ntraining methods. We validated our hypothesis with competitive results on\nsimulations from OpenAI MuJoCo, Roboschool, Atari, and a custom 2D urban\ndriving environment, with various ablation and generalization tests, trained\nwith multiple black-box and policy gradient training methods. The proposed\narchitecture has the potential to improve upon broader control tasks by\nincorporating problem specific priors into the architecture. As a case study,\nwe demonstrate much improved performance for locomotion tasks by emulating the\nbiological central pattern generators (CPGs) as the nonlinear part of the\narchitecture.\n",
"title": "Structured Control Nets for Deep Reinforcement Learning"
} | null | null | null | null | true | null | 2219 | null | Default | null | null |
null | {
"abstract": " We consider the problem of identifying any $k$ out of the best $m$ arms in an\n$n$-armed stochastic multi-armed bandit. Framed in the PAC setting, this\nparticular problem generalises both the problem of `best subset selection' and\nthat of selecting `one out of the best m' arms [arcsk 2017]. In applications\nsuch as crowd-sourcing and drug-designing, identifying a single good solution\nis often not sufficient. Moreover, finding the best subset might be hard due to\nthe presence of many indistinguishably close solutions. Our generalisation of\nidentifying exactly $k$ arms out of the best $m$, where $1 \\leq k \\leq m$,\nserves as a more effective alternative. We present a lower bound on the\nworst-case sample complexity for general $k$, and a fully sequential PAC\nalgorithm, \\GLUCB, which is more sample-efficient on easy instances. Also,\nextending our analysis to infinite-armed bandits, we present a PAC algorithm\nthat is independent of $n$, which identifies an arm from the best $\\rho$\nfraction of arms using at most an additive poly-log number of samples than\ncompared to the lower bound, thereby improving over [arcsk 2017] and\n[Aziz+AKA:2018]. The problem of identifying $k > 1$ distinct arms from the best\n$\\rho$ fraction is not always well-defined; for a special class of this\nproblem, we present lower and upper bounds. Finally, through a reduction, we\nestablish a relation between upper bounds for the `one out of the best $\\rho$'\nproblem for infinite instances and the `one out of the best $m$' problem for\nfinite instances. We conjecture that it is more efficient to solve `small'\nfinite instances using the latter formulation, rather than going through the\nformer.\n",
"title": "PAC Identification of Many Good Arms in Stochastic Multi-Armed Bandits"
} | null | null | null | null | true | null | 2220 | null | Default | null | null |
null | {
"abstract": " Recommender systems (RS) help users navigate large sets of items in the\nsearch for \"interesting\" ones. One approach to RS is Collaborative Filtering\n(CF), which is based on the idea that similar users are interested in similar\nitems. Most model-based approaches to CF seek to train a\nmachine-learning/data-mining model based on sparse data; the model is then used\nto provide recommendations. While most of the proposed approaches are effective\nfor small-size situations, the combinatorial nature of the problem makes it\nimpractical for medium-to-large instances. In this work we present a novel\napproach to CF that works by training a Denoising Auto-Encoder (DAE) on\ncorrupted baskets, i.e., baskets from which one or more items have been\nremoved. The DAE is then forced to learn to reconstruct the original basket\ngiven its corrupted input. Due to recent advancements in optimization and other\ntechnologies for training neural-network models (such as DAE), the proposed\nmethod results in a scalable and practical approach to CF. The contribution of\nthis work is twofold: (1) to identify missing items in observed baskets and,\nthus, directly providing a CF model; and, (2) to construct a generative model\nof baskets which may be used, for instance, in simulation analysis or as part\nof a more complex analytical method.\n",
"title": "Collaborative Filtering using Denoising Auto-Encoders for Market Basket Data"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 2221 | null | Validated | null | null |
null | {
"abstract": " We develop a new technique, based on Stein's method, for comparing two\nstationary distributions of irreducible Markov Chains whose update rules are\n`close enough'. We apply this technique to compare Ising models on $d$-regular\nexpander graphs to the Curie-Weiss model (complete graph) in terms of pairwise\ncorrelations and more generally $k$th order moments. Concretely, we show that\n$d$-regular Ramanujan graphs approximate the $k$th order moments of the\nCurie-Weiss model to within average error $k/\\sqrt{d}$ (averaged over the size\n$k$ subsets). The result applies even in the low-temperature regime; we also\nderive some simpler approximation results for functionals of Ising models that\nhold only at high enough temperatures.\n",
"title": "Stein's Method for Stationary Distributions of Markov Chains and Application to Ising Models"
} | null | null | null | null | true | null | 2222 | null | Default | null | null |
null | {
"abstract": " In this paper, we propose a novel and elegant solution to \"Multi-Source\nNeural Machine Translation\" (MSNMT) which only relies on preprocessing a N-way\nmultilingual corpus without modifying the Neural Machine Translation (NMT)\narchitecture or training procedure. We simply concatenate the source sentences\nto form a single long multi-source input sentence while keeping the target side\nsentence as it is and train an NMT system using this preprocessed corpus. We\nevaluate our method in resource poor as well as resource rich settings and show\nits effectiveness (up to 4 BLEU using 2 source languages and up to 6 BLEU using\n5 source languages) by comparing against existing methods for MSNMT. We also\nprovide some insights on how the NMT system leverages multilingual information\nin such a scenario by visualizing attention.\n",
"title": "Enabling Multi-Source Neural Machine Translation By Concatenating Source Sentences In Multiple Languages"
} | null | null | null | null | true | null | 2223 | null | Default | null | null |
null | {
"abstract": " The literature on Inverse Reinforcement Learning (IRL) typically assumes that\nhumans take actions in order to minimize the expected value of a cost function,\ni.e., that humans are risk neutral. Yet, in practice, humans are often far from\nbeing risk neutral. To fill this gap, the objective of this paper is to devise\na framework for risk-sensitive IRL in order to explicitly account for a human's\nrisk sensitivity. To this end, we propose a flexible class of models based on\ncoherent risk measures, which allow us to capture an entire spectrum of risk\npreferences from risk-neutral to worst-case. We propose efficient\nnon-parametric algorithms based on linear programming and semi-parametric\nalgorithms based on maximum likelihood for inferring a human's underlying risk\nmeasure and cost function for a rich class of static and dynamic\ndecision-making settings. The resulting approach is demonstrated on a simulated\ndriving game with ten human participants. Our method is able to infer and mimic\na wide range of qualitatively different driving styles from highly risk-averse\nto risk-neutral in a data-efficient manner. Moreover, comparisons of the\nRisk-Sensitive (RS) IRL approach with a risk-neutral model show that the RS-IRL\nframework more accurately captures observed participant behavior both\nqualitatively and quantitatively, especially in scenarios where catastrophic\noutcomes such as collisions can occur.\n",
"title": "Risk-sensitive Inverse Reinforcement Learning via Semi- and Non-Parametric Methods"
} | null | null | null | null | true | null | 2224 | null | Default | null | null |
null | {
"abstract": " In recent work it was shown how recursive factorisation of certain QRT maps\nleads to Somos-4 and Somos-5 recurrences with periodic coefficients, and to a\nfifth-order recurrence with the Laurent property. Here we recursively factorise\nthe 12-parameter symmetric QRT map, given by a second-order recurrence, to\nobtain a system of three coupled recurrences which possesses the Laurent\nproperty. As degenerate special cases, we first derive systems of two coupled\nrecurrences corresponding to the 5-parameter multiplicative and additive\nsymmetric QRT maps. In all cases, the Laurent property is established using a\ngeneralisation of a result due to Hickerson, and exact formulae for degree\ngrowth are found from ultradiscrete (tropical) analogues of the recurrences.\nFor the general 18-parameter QRT map it is shown that the components of the\niterates can be written as a ratio of quantities that satisfy the same Somos-7\nrecurrence.\n",
"title": "QRT maps and related Laurent systems"
} | null | null | [
"Physics",
"Mathematics"
]
| null | true | null | 2225 | null | Validated | null | null |
null | {
"abstract": " This paper seeks to combine differential game theory with the\nactor-critic-identifier architecture to determine forward-in-time, approximate\noptimal controllers for formation tracking in multi-agent systems, where the\nagents have uncertain heterogeneous nonlinear dynamics. A continuous control\nstrategy is proposed, using communication feedback from extended neighbors on a\ncommunication topology that has a spanning tree. A model-based reinforcement\nlearning technique is developed to cooperatively control a group of agents to\ntrack a trajectory in a desired formation. Simulation results are presented to\ndemonstrate the performance of the developed technique.\n",
"title": "Model-based reinforcement learning in differential graphical games"
} | null | null | null | null | true | null | 2226 | null | Default | null | null |
null | {
"abstract": " We introduce the notion of the depth of a finite group $G$, defined as the\nminimal length of an unrefinable chain of subgroups from $G$ to the trivial\nsubgroup. In this paper we investigate the depth of (non-abelian) finite simple\ngroups. We determine the simple groups of minimal depth, and show, somewhat\nsurprisingly, that alternating groups have bounded depth. We also establish\ngeneral upper bounds on the depth of simple groups of Lie type, and study the\nrelation between the depth and the much studied notion of the length of simple\ngroups. The proofs of our main theorems depend (among other tools) on a deep\nnumber-theoretic result, namely, Helfgott's recent solution of the ternary\nGoldbach conjecture.\n",
"title": "The depth of a finite simple group"
} | null | null | null | null | true | null | 2227 | null | Default | null | null |
null | {
"abstract": " Cosmic rays originating from extraterrestrial sources are permanently\narriving at Earth atmosphere, where they produce up to billions of secondary\nparticles. The analysis of the secondary particles reaching to the surface of\nthe Earth may provide a very valuable information about the Sun activity,\nchanges in the geomagnetic field and the atmosphere, among others. In this\narticle, we present the first preliminary results of the analysis of the cosmic\nrays measured with a high resolution tracking detector, TRAGALDABAS, located at\nthe Univ. of Santiago de Compostela, in Spain.\n",
"title": "TRAGALDABAS. First results on cosmic ray studies and their relation with the solar activity, the Earth magnetic field and the atmospheric properties"
} | null | null | null | null | true | null | 2228 | null | Default | null | null |
null | {
"abstract": " Our aim in this paper is to establish some strong stability properties of a\nsolution of a stochastic differential equation driven by a fractional Brownian\nmotion for which the pathwise uniqueness holds. The results are obtained using\nSkorokhod's selection theorem.\n",
"title": "Approximation of solutions of SDEs driven by a fractional Brownian motion, under pathwise uniqueness"
} | null | null | [
"Mathematics"
]
| null | true | null | 2229 | null | Validated | null | null |
null | {
"abstract": " This paper develops theory for feasible estimators of finite-dimensional\nparameters identified by general conditional quantile restrictions, under much\nweaker assumptions than previously seen in the literature. This includes\ninstrumental variables nonlinear quantile regression as a special case. More\nspecifically, we consider a set of unconditional moments implied by the\nconditional quantile restrictions, providing conditions for local\nidentification. Since estimators based on the sample moments are generally\nimpossible to compute numerically in practice, we study feasible estimators\nbased on smoothed sample moments. We propose a method of moments estimator for\nexactly identified models, as well as a generalized method of moments estimator\nfor over-identified models. We establish consistency and asymptotic normality\nof both estimators under general conditions that allow for weakly dependent\ndata and nonlinear structural models. Simulations illustrate the finite-sample\nproperties of the methods. Our in-depth empirical application concerns the\nconsumption Euler equation derived from quantile utility maximization.\nAdvantages of the quantile Euler equation include robustness to fat tails,\ndecoupling of risk attitude from the elasticity of intertemporal substitution,\nand log-linearization without any approximation error. For the four countries\nwe examine, the quantile estimates of discount factor and elasticity of\nintertemporal substitution are economically reasonable for a range of quantiles\nabove the median, even when two-stage least squares estimates are not\nreasonable.\n",
"title": "Smoothed GMM for quantile models"
} | null | null | null | null | true | null | 2230 | null | Default | null | null |
null | {
"abstract": " Generating realistic artificial preference distributions is an important part\nof any simulation analysis of electoral systems. While this has been discussed\nin some detail in the context of a single electoral district, many electoral\nsystems of interest are based on multiple districts. Neither treating\npreferences between districts as independent nor ignoring the district\nstructure yields satisfactory results. We present a model based on an extension\nof the classic Eggenberger-Pólya urn, in which each district is represented\nby an urn and there is correlation between urns. We show in detail that this\nprocedure has a small number of tunable parameters, is computationally\nefficient, and produces \"realistic-looking\" distributions. We intend to use it\nin further studies of electoral systems.\n",
"title": "Multi-district preference modelling"
} | null | null | null | null | true | null | 2231 | null | Default | null | null |
null | {
"abstract": " Let $(M,\\Omega)$ be a closed $8$-dimensional manifold equipped with a\ngenerically non-integrable $\\mathrm{Spin}(7)$-structure $\\Omega$. We prove that\nif $\\mathrm{Hom}(H^{3}(M,\\mathbb{Z}), \\mathbb{Z}_{2}) = 0$ then the moduli\nspace of irreducible $\\mathrm{Spin}(7)$-instantons on $(M,\\Omega)$ with gauge\ngroup $\\mathrm{SU}(r)$, $r\\geq 2$, is orientable.\n",
"title": "Orientability of the moduli space of Spin(7)-instantons"
} | null | null | [
"Mathematics"
]
| null | true | null | 2232 | null | Validated | null | null |
null | {
"abstract": " Deep learning methods have recently achieved great empirical success on\nmachine translation, dialogue response generation, summarization, and other\ntext generation tasks. At a high level, the technique has been to train\nend-to-end neural network models consisting of an encoder model to produce a\nhidden representation of the source text, followed by a decoder model to\ngenerate the target. While such models have significantly fewer pieces than\nearlier systems, significant tuning is still required to achieve good\nperformance. For text generation models in particular, the decoder can behave\nin undesired ways, such as by generating truncated or repetitive outputs,\noutputting bland and generic responses, or in some cases producing\nungrammatical gibberish. This paper is intended as a practical guide for\nresolving such undesired behavior in text generation models, with the aim of\nhelping enable real-world applications.\n",
"title": "Neural Text Generation: A Practical Guide"
} | null | null | null | null | true | null | 2233 | null | Default | null | null |
null | {
"abstract": " Point clouds provide a flexible and natural representation usable in\ncountless applications such as robotics or self-driving cars. Recently, deep\nneural networks operating on raw point cloud data have shown promising results\non supervised learning tasks such as object classification and semantic\nsegmentation. While massive point cloud datasets can be captured using modern\nscanning technology, manually labelling such large 3D point clouds for\nsupervised learning tasks is a cumbersome process. This necessitates effective\nunsupervised learning methods that can produce representations such that\ndownstream tasks require significantly fewer annotated samples. We propose a\nnovel method for unsupervised learning on raw point cloud data in which a\nneural network is trained to predict the spatial relationship between two point\ncloud segments. While solving this task, representations that capture semantic\nproperties of the point cloud are learned. Our method outperforms previous\nunsupervised learning approaches in downstream object classification and\nsegmentation tasks and performs on par with fully supervised methods.\n",
"title": "Context Prediction for Unsupervised Deep Learning on Point Clouds"
} | null | null | null | null | true | null | 2234 | null | Default | null | null |
null | {
"abstract": " A new challenge for learning algorithms in cyber-physical network systems is\nthe distributed solution of big-data classification problems, i.e., problems in\nwhich both the number of training samples and their dimension is high.\nMotivated by several problem set-ups in Machine Learning, in this paper we\nconsider a special class of quadratic optimization problems involving a \"large\"\nnumber of input data, whose dimension is \"big\". To solve these quadratic\noptimization problems over peer-to-peer networks, we propose an asynchronous,\ndistributed algorithm that scales with both the number and the dimension of the\ninput data (training samples in the classification problem). The proposed\ndistributed optimization algorithm relies on the notion of \"core-set\" which is\nused in geometric optimization to approximate the value function associated to\na given set of points with a smaller subset of points. By computing local\ncore-sets on a smaller version of the global problem and exchanging them with\nneighbors, the nodes reach consensus on a set of active constraints\nrepresenting an approximate solution for the global quadratic program.\n",
"title": "A core-set approach for distributed quadratic programming in big-data classification"
} | null | null | [
"Computer Science",
"Mathematics"
]
| null | true | null | 2235 | null | Validated | null | null |
null | {
"abstract": " The study of the restricted isometry property (RIP) for corrupted random\nmatrices is particularly important in the field of compressed sensing (CS) with\ncorruptions. If a matrix still satisfy RIP after a certain portion of rows are\nerased, then we say that the matrix has the strong restricted isometry property\n(SRIP. In the field of compressed sensing, random matrices satisfies certain\nmoment conditions are of particular interest. Among these matrices, those with\nentries generated from i.i.d Gaussian or i.i.d $\\pm1$ random variables are\noften typically considered. Recent studies have shown that a matrix generated\nfrom i.i.d Gaussian random variables satisfies the strong restricted isometry\nproperty under arbitrary erasure of rows. In the first part of this paper we\nwill work on $\\pm 1$ random matrices. We study the erasure robustness of $\\pm\n1$ random matrices show that with overwhelming probability the SRIP will still\nhold. Moreover the analysis will also lead to the robust version of the\nJohnson-Lindenstrauss Lemma for $\\pm 1$ matrices. Then in the second part of\nthis paper we work on finite frames. The study of the stability of finite\nframes under corruptions shares a lot of similarity to CS with corruption. We\nwill focus on the Gaussian finite frames as a starter. We will improve existing\nresults and confirm that a Gaussian random frame is numerically stable under\narbitrary erasure of rows.\n",
"title": "On the Erasure Robustness Property of Random Matrices"
} | null | null | null | null | true | null | 2236 | null | Default | null | null |
null | {
"abstract": " We propose new positive definite kernels for permutations. First we introduce\na weighted version of the Kendall kernel, which allows to weight unequally the\ncontributions of different item pairs in the permutations depending on their\nranks. Like the Kendall kernel, we show that the weighted version is invariant\nto relabeling of items and can be computed efficiently in $O(n \\ln(n))$\noperations, where $n$ is the number of items in the permutation. Second, we\npropose a supervised approach to learn the weights by jointly optimizing them\nwith the function estimated by a kernel machine. Third, while the Kendall\nkernel considers pairwise comparison between items, we extend it by considering\nhigher-order comparisons among tuples of items and show that the supervised\napproach of learning the weights can be systematically generalized to\nhigher-order permutation kernels.\n",
"title": "The Weighted Kendall and High-order Kernels for Permutations"
} | null | null | null | null | true | null | 2237 | null | Default | null | null |
null | {
"abstract": " We study a connection between synchronizing automata and its set $M$ of\nminimal reset words, i.e., such that no proper factor is a reset word. We first\nshow that any synchronizing automaton having the set of minimal reset words\nwhose set of factors does not contain a word of length at most\n$\\frac{1}{4}\\min\\{|u|: u\\in I\\}+\\frac{1}{16}$ has a reset word of length at\nmost $(n-\\frac{1}{2})^{2}$ In the last part of the paper we focus on the\nexistence of synchronizing automata with a given ideal $I$ that serves as the\nset of reset words. To this end, we introduce the notion of the tail structure\nof the (not necessarily regular) ideal $I=\\Sigma^{*}M\\Sigma^{*}$. With this\ntool, we first show the existence of an infinite strongly connected\nsynchronizing automaton $\\mathcal{A}$ having $I$ as the set of reset words and\nsuch that every other strongly connected synchronizing automaton having $I$ as\nthe set of reset words is an homomorphic image of $\\mathcal{A}$. Finally, we\nshow that for any non-unary regular ideal $I$ there is a strongly connected\nsynchronizing automaton having $I$ as the set of reset words with at most\n$(km^{k})2^{km^{k}n}$ states, where $k=|\\Sigma|$, $m$ is the length of a\nshortest word in $M$, and $n$ is the dimension of the smallest automaton\nrecognizing $M$ (state complexity of $M$). This automaton is computable and we\nshow an algorithm to compute it in time $\\mathcal{O}((k^{2}m^{k})2^{km^{k}n})$.\n",
"title": "Synchronizing automata and the language of minimal reset words"
} | null | null | [
"Computer Science"
]
| null | true | null | 2238 | null | Validated | null | null |
null | {
"abstract": " The quest for performant networks has been a significant force that drives\nthe advancements of deep learning in recent years. While rewarding, improving\nnetwork design has never been an easy journey. The large design space combined\nwith the tremendous cost required for network training poses a major obstacle\nto this endeavor. In this work, we propose a new approach to this problem,\nnamely, predicting the performance of a network before training, based on its\narchitecture. Specifically, we develop a unified way to encode individual\nlayers into vectors and bring them together to form an integrated description\nvia LSTM. Taking advantage of the recurrent network's strong expressive power,\nthis method can reliably predict the performances of various network\narchitectures. Our empirical studies showed that it not only achieved accurate\npredictions but also produced consistent rankings across datasets -- a key\ndesideratum in performance prediction.\n",
"title": "Peephole: Predicting Network Performance Before Training"
} | null | null | null | null | true | null | 2239 | null | Default | null | null |
null | {
"abstract": " We analyze the loss landscape and expressiveness of practical deep\nconvolutional neural networks (CNNs) with shared weights and max pooling\nlayers. We show that such CNNs produce linearly independent features at a\n\"wide\" layer which has more neurons than the number of training samples. This\ncondition holds e.g. for the VGG network. Furthermore, we provide for such wide\nCNNs necessary and sufficient conditions for global minima with zero training\nerror. For the case where the wide layer is followed by a fully connected layer\nwe show that almost every critical point of the empirical loss is a global\nminimum with zero training error. Our analysis suggests that both depth and\nwidth are very important in deep learning. While depth brings more\nrepresentational power and allows the network to learn high level features,\nwidth smoothes the optimization landscape of the loss function in the sense\nthat a sufficiently wide network has a well-behaved loss surface with almost no\nbad local minima.\n",
"title": "Optimization Landscape and Expressivity of Deep CNNs"
} | null | null | null | null | true | null | 2240 | null | Default | null | null |
null | {
"abstract": " We formulated and implemented a procedure to generate aliasing-free\nexcitation source signals. It uses a new antialiasing filter in the continuous\ntime domain followed by an IIR digital filter for response equalization. We\nintroduced a cosine-series-based general design procedure for the new\nantialiasing function. We applied this new procedure to implement the\nantialiased Fujisaki-Ljungqvist model. We also applied it to revise our\nprevious implementation of the antialiased Fant-Liljencrants model. A\ncombination of these signals and a lattice implementation of the time varying\nvocal tract model provides a reliable and flexible basis to test fo extractors\nand source aperiodicity analysis methods. MATLAB implementations of these\nantialiased excitation source models are available as part of our open source\ntools for speech science.\n",
"title": "A new cosine series antialiasing function and its application to aliasing-free glottal source models for speech and singing synthesis"
} | null | null | null | null | true | null | 2241 | null | Default | null | null |
null | {
"abstract": " Let $G$ be a quasi-simple algebraic group defined over an algebraically\nclosed field $k$ and $B$ a Borel subgroup of $G$ acting on the nilradical\n$\\mathfrak{n}$ of its Lie algebra $\\mathfrak{b}$ via the Adjoint\nrepresentation. It is known that $B$ has only finitely many orbits in only five\ncases: when $G$ is of type $A_{n}$ for $n \\leq 4$, and when $G$ is type\n$B_{2}$. In this paper, we elaborate on this work in the case when $G =SL_{n\n+1}(k)$ (type $A_{n})$, for $n \\leq 4$, by finding the polynomial defining\nequations of each orbit. Consequences of these equations include the dimension\nof the orbits and the closure ordering on the set of orbits, although these\nfacts are already known. The other case, when $G$ is type $B_{2}$, can be\napproached the same way and is treated in a separate paper, where we believe\nthe determination of the closure order is new.\n",
"title": "Defining Equations of Nilpotent Orbits for Borel Subgroups of Modality Zero in Type $A_{n}$"
} | null | null | null | null | true | null | 2242 | null | Default | null | null |
null | {
"abstract": " Let $f:\\{0,1\\}^n \\rightarrow \\{0,1\\}$ be a Boolean function. The certificate\ncomplexity $C(f)$ is a complexity measure that is quadratically tight for the\nzero-error randomized query complexity $R_0(f)$: $C(f) \\leq R_0(f) \\leq\nC(f)^2$. In this paper we study a new complexity measure that we call\nexpectational certificate complexity $EC(f)$, which is also a quadratically\ntight bound on $R_0(f)$: $EC(f) \\leq R_0(f) = O(EC(f)^2)$. We prove that $EC(f)\n\\leq C(f) \\leq EC(f)^2$ and show that there is a quadratic separation between\nthe two, thus $EC(f)$ gives a tighter upper bound for $R_0(f)$. The measure is\nalso related to the fractional certificate complexity $FC(f)$ as follows:\n$FC(f) \\leq EC(f) = O(FC(f)^{3/2})$. This also connects to an open question by\nAaronson whether $FC(f)$ is a quadratically tight bound for $R_0(f)$, as\n$EC(f)$ is in fact a relaxation of $FC(f)$.\nIn the second part of the work, we upper bound the distributed query\ncomplexity $D^\\mu_\\epsilon(f)$ for product distributions $\\mu$ by the square of\nthe query corruption bound ($\\mathrm{corr}_\\epsilon(f)$) which improves upon a\nresult of Harsha, Jain and Radhakrishnan [2015]. A similar statement for\ncommunication complexity is open.\n",
"title": "Quadratically Tight Relations for Randomized Query Complexity"
} | null | null | null | null | true | null | 2243 | null | Default | null | null |
null | {
"abstract": " Oshima's Lemma describes the orbits of parabolic subgroups of irreducible\nfinite Weyl groups on crystallographic root systems. This note generalises that\nresult to all root systems of finite Coxeter groups, and provides a self\ncontained proof, independent of the representation theory of semisimple complex\nLie algebras.\n",
"title": "Parabolic subgroup orbits on finite root systems"
} | null | null | null | null | true | null | 2244 | null | Default | null | null |
null | {
"abstract": " In human-in-the-loop machine learning, the user provides information beyond\nthat in the training data. Many algorithms and user interfaces have been\ndesigned to optimize and facilitate this human--machine interaction; however,\nfewer studies have addressed the potential defects the designs can cause.\nEffective interaction often requires exposing the user to the training data or\nits statistics. The design of the system is then critical, as this can lead to\ndouble use of data and overfitting, if the user reinforces noisy patterns in\nthe data. We propose a user modelling methodology, by assuming simple rational\nbehaviour, to correct the problem. We show, in a user study with 48\nparticipants, that the method improves predictive performance in a sparse\nlinear regression sentiment analysis task, where graded user knowledge on\nfeature relevance is elicited. We believe that the key idea of inferring user\nknowledge with probabilistic user models has general applicability in guarding\nagainst overfitting and improving interactive machine learning.\n",
"title": "User Modelling for Avoiding Overfitting in Interactive Knowledge Elicitation for Prediction"
} | null | null | null | null | true | null | 2245 | null | Default | null | null |
null | {
"abstract": " We employ the exponentially improved asymptotic expansions of the confluent\nhypergeometric functions on the Stokes lines discussed by the author [Appl.\nMath. Sci. {\\bf 7} (2013) 6601--6609] to give the analogous expansions of the\nmodified Bessel functions $I_\\nu(z)$ and $K_\\nu(z)$ for large $z$ and finite\n$\\nu$ on $\\arg\\,z=\\pm\\pi$ (and, in the case of $I_\\nu(z)$, also on\n$\\arg\\,z=0$). Numerical results are presented to illustrate the accuracy of\nthese expansions.\n",
"title": "A note on the asymptotics of the modified Bessel functions on the Stokes lines"
} | null | null | null | null | true | null | 2246 | null | Default | null | null |
null | {
"abstract": " In data summarization we want to choose k prototypes in order to summarize a\ndata set. We study a setting where the data set comprises several demographic\ngroups and we are restricted to choose k_i prototypes belonging to group i. A\ncommon approach to the problem without the fairness constraint is to optimize a\ncentroid-based clustering objective such as k-center. A natural extension then\nis to incorporate the fairness constraint into the clustering objective.\nExisting algorithms for doing so run in time super-quadratic in the size of the\ndata set. This is in contrast to the standard k-center objective that can be\napproximately optimized in linear time. In this paper, we resolve this gap by\nproviding a simple approximation algorithm for the k-center problem under the\nfairness constraint with running time linear in the size of the data set and k.\nIf the number of demographic groups is small, the approximation guarantee of\nour algorithm only incurs a constant-factor overhead. We demonstrate the\napplicability of our algorithm on both synthetic and real data sets.\n",
"title": "Fair k-Center Clustering for Data Summarization"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 2247 | null | Validated | null | null |
null | {
"abstract": " The brms package allows R users to easily specify a wide range of Bayesian\nsingle-level and multilevel models, which are fitted with the probabilistic\nprogramming language Stan behind the scenes. Several response distributions are\nsupported, of which all parameters (e.g., location, scale, and shape) can be\npredicted at the same time thus allowing for distributional regression.\nNon-linear relationships may be specified using non-linear predictor terms or\nsemi-parametric approaches such as splines or Gaussian processes. To make all\nof these modeling options possible in a multilevel framework, brms provides an\nintuitive and powerful formula syntax, which extends the well known formula\nsyntax of lme4. The purpose of the present paper is to introduce this syntax in\ndetail and to demonstrate its usefulness with four examples, each showing other\nrelevant aspects of the syntax.\n",
"title": "Advanced Bayesian Multilevel Modeling with the R Package brms"
} | null | null | null | null | true | null | 2248 | null | Default | null | null |
null | {
"abstract": " In this paper, we prove the characterization of the $(K, \\infty)$-super\nPerelman Ricci flows by various functional inequalities and gradient estimate\nfor the heat semigroup generated by the Witten Laplacian on manifolds equipped\nwith time dependent metrics and potentials. As a byproduct, we derive the\nHamilton type dimension free Harnack inequality on manifolds with $(K,\n\\infty)$-super Perelman Ricci flows. Based on a new second order differential\ninequality on the Boltzmann-Shannon entropy for the heat equation of the Witten\nLaplacian, we introduce a new $W$-entropy quantity and prove its monotonicity\nfor the heat equation of the Witten Laplacian on complete Riemannian manifolds\nwith the $CD(K, \\infty)$-condition and on compact manifolds with $(K,\n\\infty)$-super Perelman Ricci flows. Our results characterize the $(K,\n\\infty)$-Ricci solitons and the $(K, \\infty)$-Perelman Ricci flows. We also\nprove a second order differential entropy inequality on $(K, m)$-super Ricci\nflows, which can be used to characterize the $(K, m)$-Ricci solitons and the\n$(K, m)$-Ricci flows. Finally, we give a probabilistic interpretation of the\n$W$-entropy for the heat equation of the Witten Laplacian on manifolds with the\n$CD(K, m)$-condition.\n",
"title": "$W$-entropy, super Perelman Ricci flows and $(K, m)$-Ricci solitons"
} | null | null | null | null | true | null | 2249 | null | Default | null | null |
null | {
"abstract": " Anomalies in the abundance measurements of short lived radionuclides in\nmeteorites indicate that the protosolar nebulae was irradiated by a high amount\nof energetic particles (E$\\gtrsim$10 MeV). The particle flux of the\ncontemporary Sun cannot explain these anomalies. However, similar to T Tauri\nstars the young Sun was more active and probably produced enough high energy\nparticles to explain those anomalies. We want to study the interaction of\nstellar energetic particles with the gas component of the disk and identify\npossible observational tracers of this interaction. We use a 2D radiation\nthermo-chemical protoplanetary disk code to model a disk representative for T\nTauri stars. We use a particle energy distribution derived from solar flare\nobservations and an enhanced stellar particle flux proposed for T Tauri stars.\nFor this particle spectrum we calculate the stellar particle ionization rate\nthroughout the disk with an accurate particle transport model. We study the\nimpact of stellar particles for models with varying X-ray and cosmic-ray\nionization rates. We find that stellar particle ionization has a significant\nimpact on the abundances of the common disk ionization tracers HCO$^+$ and\nN$_2$H$^+$, especially in models with low cosmic-ray ionization rates. In\ncontrast to cosmic rays and X-rays, stellar particles cannot reach the midplane\nof the disk. Therefore molecular ions residing in the disk surface layers are\nmore affected by stellar particle ionization than molecular ions tracing the\ncold layers/midplane of the disk. Spatially resolved observations of molecular\nions tracing different vertical layers of the disk allow to disentangle the\ncontribution of stellar particle ionization from other competing ionization\nsources. Modeling such observations with a model like the one presented here\nallows to constrain the stellar particle flux in disks around T Tauri stars.\n",
"title": "Stellar energetic particle ionization in protoplanetary disks around T Tauri stars"
} | null | null | [
"Physics"
]
| null | true | null | 2250 | null | Validated | null | null |
null | {
"abstract": " We establish a geometric condition guaranteeing exact copositive relaxation\nfor the nonconvex quadratic optimization problem under two quadratic and\nseveral linear constraints, and present sufficient conditions for global\noptimality in terms of generalized Karush-Kuhn-Tucker multipliers. The\ncopositive relaxation is tighter than the usual Lagrangian relaxation. We\nillustrate this by providing a whole class of quadratic optimization problems\nthat enjoys exactness of copositive relaxation while the usual Lagrangian\nduality gap is infinite. Finally, we also provide verifiable conditions under\nwhich both the usual Lagrangian relaxation and the copositive relaxation are\nexact for an extended CDT (two-ball trust-region) problem. Importantly, the\nsufficient conditions can be verified by solving linear optimization problems.\n",
"title": "Extended Trust-Region Problems with One or Two Balls: Exact Copositive and Lagrangian Relaxations"
} | null | null | null | null | true | null | 2251 | null | Default | null | null |
null | {
"abstract": " This paper proposes a family of weighted batch means variance estimators,\nwhich are computationally efficient and can be conveniently applied in\npractice. The focus is on Markov chain Monte Carlo simulations and estimation\nof the asymptotic covariance matrix in the Markov chain central limit theorem,\nwhere conditions ensuring strong consistency are provided. Finite sample\nperformance is evaluated through auto-regressive, Bayesian spatial-temporal,\nand Bayesian logistic regression examples, where the new estimators show\nsignificant computational gains with a minor sacrifice in variance compared\nwith existing methods.\n",
"title": "Weighted batch means estimators in Markov chain Monte Carlo"
} | null | null | [
"Statistics"
]
| null | true | null | 2252 | null | Validated | null | null |
null | {
"abstract": " Gradient-based optimization is the foundation of deep learning and\nreinforcement learning. Even when the mechanism being optimized is unknown or\nnot differentiable, optimization using high-variance or biased gradient\nestimates is still often the best strategy. We introduce a general framework\nfor learning low-variance, unbiased gradient estimators for black-box functions\nof random variables. Our method uses gradients of a neural network trained\njointly with model parameters or policies, and is applicable in both discrete\nand continuous settings. We demonstrate this framework for training discrete\nlatent-variable models. We also give an unbiased, action-conditional extension\nof the advantage actor-critic reinforcement learning algorithm.\n",
"title": "Backpropagation through the Void: Optimizing control variates for black-box gradient estimation"
} | null | null | null | null | true | null | 2253 | null | Default | null | null |
null | {
"abstract": " Dirichlet process mixture models (DPMM) are a cornerstone of Bayesian\nnon-parametrics. While these models free from choosing the number of components\na-priori, computationally attractive variational inference often reintroduces\nthe need to do so, via a truncation on the variational distribution. In this\npaper we present a truncation-free hybrid inference for DPMM, combining the\nadvantages of sampling-based MCMC and variational methods. The proposed\nhybridization enables more efficient variational updates, while increasing\nmodel complexity only if needed. We evaluate the properties of the hybrid\nupdates and their empirical performance in single- as well as mixed-membership\nmodels. Our method is easy to implement and performs favorably compared to\nexisting schemas.\n",
"title": "Truncation-free Hybrid Inference for DPMM"
} | null | null | null | null | true | null | 2254 | null | Default | null | null |
null | {
"abstract": " We study the existence and uniqueness of minimal right determiners in various\ncategories. Particularly in a Hom-finite hereditary abelian category with\nenough projectives, we prove that the Auslander-Reiten-Smal{\\o}-Ringel formula\nof the minimal right determiner still holds. As an application, we give a\nformula of minimal right determiners in the category of finitely presented\nrepresentations of strongly locally finite quivers.\n",
"title": "Functors and morphisms determined by subcategories"
} | null | null | [
"Mathematics"
]
| null | true | null | 2255 | null | Validated | null | null |
null | {
"abstract": " We study the stationary photon output and statistics of small lasers. Our\nclosed-form expressions clarify the contribution of collective effects due to\nthe interaction between quantum emitters. We generalize laser rate equations\nand explain photon trapping: a decrease of the photon number output below the\nlasing threshold, derive an expression for the stationary cavity mode\nautocorrelation function $g_2$, which implies that collective effects may\nstrongly influence the photon statistics. We identify conditions for coherent,\nthermal and superthermal radiation, the latter being a unique fingerprint for\ncollective emission in lasers. These generic analytical results agree with\nrecent experiments, complement numerical results, and provide insight into and\ndesign rules for nanolasers.\n",
"title": "Collective Effects in Nanolasers Explained by Generalized Rate Equations"
} | null | null | null | null | true | null | 2256 | null | Default | null | null |
null | {
"abstract": " Many applied settings in empirical economics involve simultaneous estimation\nof a large number of parameters. In particular, applied economists are often\ninterested in estimating the effects of many-valued treatments (like teacher\neffects or location effects), treatment effects for many groups, and prediction\nmodels with many regressors. In these settings, machine learning methods that\ncombine regularized estimation and data-driven choices of regularization\nparameters are useful to avoid over-fitting. In this article, we analyze the\nperformance of a class of machine learning estimators that includes ridge,\nlasso and pretest in contexts that require simultaneous estimation of many\nparameters. Our analysis aims to provide guidance to applied researchers on (i)\nthe choice between regularized estimators in practice and (ii) data-driven\nselection of regularization parameters. To address (i), we characterize the\nrisk (mean squared error) of regularized estimators and derive their relative\nperformance as a function of simple features of the data generating process. To\naddress (ii), we show that data-driven choices of regularization parameters,\nbased on Stein's unbiased risk estimate or on cross-validation, yield\nestimators with risk uniformly close to the risk attained under the optimal\n(unfeasible) choice of regularization parameters. We use data from recent\nexamples in the empirical economics literature to illustrate the practical\napplicability of our results.\n",
"title": "The Risk of Machine Learning"
} | null | null | null | null | true | null | 2257 | null | Default | null | null |
null | {
"abstract": " Periodograms are used as a key significance assessment and visualisation tool\nto display the significant periodicities in unevenly sampled time series. We\nintroduce a framework of periodograms, called \"Agatha\", to disentangle periodic\nsignals from correlated noise and to solve the 2-dimensional model selection\nproblem: signal dimension and noise model dimension. These periodograms are\ncalculated by applying likelihood maximization and marginalization and combined\nin a self-consistent way. We compare Agatha with other periodograms for the\ndetection of Keplerian signals in synthetic radial velocity data produced for\nthe Radial Velocity Challenge as well as in radial velocity datasets of several\nSun-like stars. In our tests we find Agatha is able to recover signals to the\nadopted detection limit of the radial velocity challenge. Applied to real\nradial velocity, we use Agatha to confirm previous analysis of CoRoT-7 and to\nfind two new planet candidates with minimum masses of 15.1 $M_\\oplus$ and 7.08\n$M_\\oplus$ orbiting HD177565 and HD41248, with periods of 44.5 d and 13.4 d,\nrespectively. We find that Agatha outperforms other periodograms in terms of\nremoving correlated noise and assessing the significances of signals with more\nrobust metrics. Moreover, it can be used to select the optimal noise model and\nto test the consistency of signals in time. Agatha is intended to be flexible\nenough to be applied to time series analyses in other astronomical and\nscientific disciplines. Agatha is available at this http URL.\n",
"title": "Agatha: disentangling periodic signals from correlated noise in a periodogram framework"
} | null | null | null | null | true | null | 2258 | null | Default | null | null |
null | {
"abstract": " In this work, we consider the problem of predicting the course of a\nprogressive disease, such as cancer or Alzheimer's. Progressive diseases often\nstart with mild symptoms that might precede a diagnosis, and each patient\nfollows their own trajectory. Patient trajectories exhibit wild variability,\nwhich can be associated with many factors such as genotype, age, or sex. An\nadditional layer of complexity is that, in real life, the amount and type of\ndata available for each patient can differ significantly. For example, for one\npatient we might have no prior history, whereas for another patient we might\nhave detailed clinical assessments obtained at multiple prior time-points. This\npaper presents a probabilistic model that can handle multiple modalities\n(including images and clinical assessments) and variable patient histories with\nirregular timings and missing entries, to predict clinical scores at future\ntime-points. We use a sigmoidal function to model latent disease progression,\nwhich gives rise to clinical observations in our generative model. We\nimplemented an approximate Bayesian inference strategy on the proposed model to\nestimate the parameters on data from a large population of subjects.\nFurthermore, the Bayesian framework enables the model to automatically\nfine-tune its predictions based on historical observations that might be\navailable on the test subject. We applied our method to a longitudinal\nAlzheimer's disease dataset with more than 3000 subjects [23] and present a\ndetailed empirical analysis of prediction performance under different\nscenarios, with comparisons against several benchmarks. We also demonstrate how\nthe proposed model can be interrogated to glean insights about temporal\ndynamics in Alzheimer's disease.\n",
"title": "A Probabilistic Disease Progression Model for Predicting Future Clinical Outcome"
} | null | null | null | null | true | null | 2259 | null | Default | null | null |
null | {
"abstract": " In this paper we present the state of the art about the quarks: group SU(3),\nLie algebra, the electric charge and mass. The quarks masses are generated in\nthe same way as the lepton masses. It is constructed a term in the Lagrangian\nthat couples the Higgs doublet to the fermion fields.\n",
"title": "On the generation of the quarks through spontaneous symmetry breaking"
} | null | null | null | null | true | null | 2260 | null | Default | null | null |
null | {
"abstract": " Link discovery is an active field of research to support data integration in\nthe Web of Data. Due to the huge size and number of available data sources,\nefficient and effective link discovery is a very challenging task. Common\npairwise link discovery approaches do not scale to many sources with very large\nentity sets. We here propose a distributed holistic approach to link many data\nsources based on a clustering of entities that represent the same real-world\nobject. Our clustering approach provides a compact and fused representation of\nentities, and can identify errors in existing links as well as many new links.\nWe support a distributed execution of the clustering approach to achieve faster\nexecution times and scalability for large real-world data sets. We provide a\nnovel gold standard for multi-source clustering, and evaluate our methods with\nrespect to effectiveness and efficiency for large data sets from the geographic\nand music domains.\n",
"title": "Distributed Holistic Clustering on Linked Data"
} | null | null | null | null | true | null | 2261 | null | Default | null | null |
null | {
"abstract": " The minimum $k$-enclosing ball problem seeks the ball with smallest radius\nthat contains at least~$k$ of~$m$ given points in a general $n$-dimensional\nEuclidean space. This problem is NP-hard. We present a branch-and-bound\nalgorithm on the tree of the subsets of~$k$ points to solve this problem. The\nnodes on the tree are ordered in a suitable way, which, complemented with a\nlast-in-first-out search strategy, allows for only a small fraction of nodes to\nbe explored. Additionally, an efficient dual algorithm to solve the subproblems\nat each node is employed.\n",
"title": "A branch-and-bound algorithm for the minimum radius $k$-enclosing ball problem"
} | null | null | null | null | true | null | 2262 | null | Default | null | null |
null | {
"abstract": " Dependency parses are an effective way to inject linguistic knowledge into\nmany downstream tasks, and many practitioners wish to efficiently parse\nsentences at scale. Recent advances in GPU hardware have enabled neural\nnetworks to achieve significant gains over the previous best models, these\nmodels still fail to leverage GPUs' capability for massive parallelism due to\ntheir requirement of sequential processing of the sentence. In response, we\npropose Dilated Iterated Graph Convolutional Neural Networks (DIG-CNNs) for\ngraph-based dependency parsing, a graph convolutional architecture that allows\nfor efficient end-to-end GPU parsing. In experiments on the English Penn\nTreeBank benchmark, we show that DIG-CNNs perform on par with some of the best\nneural network parsers.\n",
"title": "Dependency Parsing with Dilated Iterated Graph CNNs"
} | null | null | null | null | true | null | 2263 | null | Default | null | null |
null | {
"abstract": " The motion of a mechanical system can be defined as a path through its\nconfiguration space. Computing such a path has a computational complexity\nscaling exponentially with the dimensionality of the configuration space. We\npropose to reduce the dimensionality of the configuration space by introducing\nthe irreducible path --- a path having a minimal swept volume. The paper\nconsists of three parts: In part I, we define the space of all irreducible\npaths and show that planning a path in the irreducible path space preserves\ncompleteness of any motion planning algorithm. In part II, we construct an\napproximation to the irreducible path space of a serial kinematic chain under\ncertain assumptions. In part III, we conduct motion planning using the\nirreducible path space for a mechanical snake in a turbine environment, for a\nmechanical octopus with eight arms in a pipe system and for the sideways motion\nof a humanoid robot moving through a room with doors and through a hole in a\nwall. We demonstrate that the concept of an irreducible path can be applied to\nany motion planning algorithm taking curvature constraints into account.\n",
"title": "Motion Planning in Irreducible Path Spaces"
} | null | null | null | null | true | null | 2264 | null | Default | null | null |
null | {
"abstract": " We give a polynomial-time algorithm for learning neural networks with one\nlayer of sigmoids feeding into any Lipschitz, monotone activation function\n(e.g., sigmoid or ReLU). We make no assumptions on the structure of the\nnetwork, and the algorithm succeeds with respect to {\\em any} distribution on\nthe unit ball in $n$ dimensions (hidden weight vectors also have unit norm).\nThis is the first assumption-free, provably efficient algorithm for learning\nneural networks with two nonlinear layers.\nOur algorithm-- {\\em Alphatron}-- is a simple, iterative update rule that\ncombines isotonic regression with kernel methods. It outputs a hypothesis that\nyields efficient oracle access to interpretable features. It also suggests a\nnew approach to Boolean learning problems via real-valued conditional-mean\nfunctions, sidestepping traditional hardness results from computational\nlearning theory.\nAlong these lines, we subsume and improve many longstanding results for PAC\nlearning Boolean functions to the more general, real-valued setting of {\\em\nprobabilistic concepts}, a model that (unlike PAC learning) requires non-i.i.d.\nnoise-tolerance.\n",
"title": "Learning Neural Networks with Two Nonlinear Layers in Polynomial Time"
} | null | null | [
"Computer Science",
"Statistics"
]
| null | true | null | 2265 | null | Validated | null | null |
null | {
"abstract": " Machine learning techniques have been used in the past using Monte Carlo\nsamples to construct predictors of the dynamic stability of power systems. In\nthis paper we move beyond the task of prediction and propose a comprehensive\napproach to use predictors, such as Decision Trees (DT), within a standard\noptimization framework for pre- and post-fault control purposes. In particular,\nwe present a generalizable method for embedding rules derived from DTs in an\noperation decision-making model. We begin by pointing out the specific\nchallenges entailed when moving from a prediction to a control framework. We\nproceed with introducing the solution strategy based on generalized disjunctive\nprogramming (GDP) as well as a two-step search method for identifying optimal\nhyper-parameters for balancing cost and control accuracy. We showcase how the\nproposed approach constructs security proxies that cover multiple contingencies\nwhile facing high-dimensional uncertainty with respect to operating conditions\nwith the use of a case study on the IEEE 39-bus system. The method is shown to\nachieve efficient system control at a marginal increase in system price\ncompared to an oracle model.\n",
"title": "Sample-Derived Disjunctive Rules for Secure Power System Operation"
} | null | null | [
"Statistics"
]
| null | true | null | 2266 | null | Validated | null | null |
null | {
"abstract": " The main properties of the climate of waves in the seasonally ice-covered\nBaltic Sea and its decadal changes since 1990 are estimated from satellite\naltimetry data. The data set of significant wave heights (SWH) from all\nexisting nine satellites, cleaned and cross-validated against in situ\nmeasurements, shows overall a very consistent picture. A comparison with visual\nobservations shows a good correspondence with correlation coefficients of\n0.6-0.8. The annual mean SWH reveals a tentative increase of 0.005 m yr-1, but\nhigher quantiles behave in a cyclic manner with a timescale of 10-15 yr.\nChanges in the basin-wide average SWH have a strong meridional pattern: an\nincrease in the central and western parts of the sea and decrease in the east.\nThis pattern is likely caused by a rotation of wind directions rather than by\nan increase in the wind speed.\n",
"title": "Satellite altimetry reveals spatial patterns of variations in the Baltic Sea wave climate"
} | null | null | [
"Physics"
]
| null | true | null | 2267 | null | Validated | null | null |
null | {
"abstract": " Estimating covariances between financial assets plays an important role in\nrisk management and optimal portfolio allocation. In practice, when the sample\nsize is small compared to the number of variables, i.e. when considering a wide\nuniverse of assets over just a few years, this poses considerable challenges\nand the empirical estimate is known to be very unstable.\nHere, we propose a novel covariance estimator based on the Gaussian Process\nLatent Variable Model (GP-LVM). Our estimator can be considered as a non-linear\nextension of standard factor models with readily interpretable parameters\nreminiscent of market betas. Furthermore, our Bayesian treatment naturally\nshrinks the sample covariance matrix towards a more structured matrix given by\nthe prior and thereby systematically reduces estimation errors.\n",
"title": "Estimation of Covariance Matrices for Portfolio Optimization using Gaussian Processes"
} | null | null | null | null | true | null | 2268 | null | Default | null | null |
null | {
"abstract": " We present a thorough analysis of the interplay of magnetic moment and the\nJahn-Teller effect in the $\\Gamma_8$ cubic multiplet. We find that in the\npresence of dynamical Jahn-Teller effect, the Zeeman interaction remains\nisotropic, whereas the $g$ and $G$ factors can change their signs. The static\nJahn-Teller distortion also can change the sign of these $g$ factors as well as\nthe nature of the magnetic anisotropy. Combining the theory with\nstate-of-the-art {\\it ab initio} calculations, we analyzed the magnetic\nproperties of Np$^{4+}$ and Ir$^{4+}$ impurity ions in cubic environment. The\ncalculated $g$ factors of Np$^{4+}$ impurity agree well with experimental data.\nThe {\\it ab initio} calculation predicts strong Jahn-Teller effect in Ir$^{4+}$\nion in cubic environment and the strong vibronic reduction of $g$ and $G$\nfactors.\n",
"title": "Zeeman interaction and Jahn-Teller effect in $Γ_8$ multiplet"
} | null | null | null | null | true | null | 2269 | null | Default | null | null |
null | {
"abstract": " Android has been the most popular smartphone system, with multiple platform\nversions (e.g., KITKAT and Lollipop) active in the market. To manage the\napplication's compatibility with one or more platform versions, Android allows\napps to declare the supported platform SDK versions in their manifest files. In\nthis paper, we make a first effort to study this modern software mechanism. Our\nobjective is to measure the current practice of the declared SDK versions\n(which we term as DSDK versions afterwards) in real apps, and the consistency\nbetween the DSDK versions and their app API calls. To this end, we perform a\nthree-dimensional analysis. First, we parse Android documents to obtain a\nmapping between each API and their corresponding platform versions. We then\nanalyze the DSDK-API consistency for over 24K apps, among which we pre-exclude\n1.3K apps that provide different app binaries for different Android versions\nthrough Google Play analysis. Besides shedding light on the current DSDK\npractice, our study quantitatively measures the two side effects of\ninappropriate DSDK versions: (i) around 1.8K apps have API calls that do not\nexist in some declared SDK versions, which causes runtime crash bugs on those\nplatform versions; (ii) over 400 apps, due to claiming the outdated targeted\nDSDK versions, are potentially exploitable by remote code execution. These\nresults indicate the importance and difficulty of declaring correct DSDK, and\nour work can help developers fulfill this goal.\n",
"title": "Measuring the Declared SDK Versions and Their Consistency with API Calls in Android Apps"
} | null | null | [
"Computer Science"
]
| null | true | null | 2270 | null | Validated | null | null |
null | {
"abstract": " This paper presents a framework for the implementation of online programming\ncompetitions, including a set of principles for the design of the multiplayer\ngame and a practical framework for the construction of the competition\nenvironment. The paper presents a successful example competition, the 2016-17\nHalite challenge, and briefly mentions a second competition, the Halite II\nchallenge, which launched in October 2017.\n",
"title": "The Design and Implementation of Modern Online Programming Competitions"
} | null | null | null | null | true | null | 2271 | null | Default | null | null |
null | {
"abstract": " Motivated by the intriguing behavior displayed in a dynamic network that\nmodels a population of extreme introverts and extroverts (XIE), we consider the\nspectral properties of ensembles of random split graph adjacency matrices. We\ndiscover that, in general, a gap emerges in the bulk spectrum between -1 and 0\nthat contains a single eigenvalue. An analytic expression for the bulk\ndistribution is derived and verified with numerical analysis. We also examine\ntheir relation to chiral ensembles, which are associated with bipartite graphs.\n",
"title": "Emergence of a spectral gap in a class of random matrices associated with split graphs"
} | null | null | null | null | true | null | 2272 | null | Default | null | null |
null | {
"abstract": " We present a software tool that employs state-of-the-art natural language\nprocessing (NLP) and machine learning techniques to help newspaper editors\ncompose effective headlines for online publication. The system identifies the\nmost salient keywords in a news article and ranks them based on both their\noverall popularity and their direct relevance to the article. The system also\nuses a supervised regression model to identify headlines that are likely to be\nwidely shared on social media. The user interface is designed to simplify and\nspeed the editor's decision process on the composition of the headline. As\nsuch, the tool provides an efficient way to combine the benefits of automated\npredictors of engagement and search-engine optimization (SEO) with human\njudgments of overall headline quality.\n",
"title": "Helping News Editors Write Better Headlines: A Recommender to Improve the Keyword Contents & Shareability of News Headlines"
} | null | null | null | null | true | null | 2273 | null | Default | null | null |
null | {
"abstract": " Knowledge graph embedding aims at translating the knowledge graph into\nnumerical representations by transforming the entities and relations into\ncontinuous low-dimensional vectors. Recently, many methods [1, 5, 3, 2, 6] have\nbeen proposed to deal with this problem, but existing single-thread\nimplementations of them are time-consuming for large-scale knowledge graphs.\nHere, we design a unified parallel framework to parallelize these methods,\nwhich achieves a significant time reduction without influencing the accuracy.\nWe name our framework as ParaGraphE, which provides a library for parallel\nknowledge graph embedding. The source code can be downloaded from\nthis https URL .\n",
"title": "ParaGraphE: A Library for Parallel Knowledge Graph Embedding"
} | null | null | null | null | true | null | 2274 | null | Default | null | null |
null | {
"abstract": " I present a simple phenomenological model for the observed linear scaling of\nthe stellar mass in old globular clusters (GCs) with $z=0$ halo mass in which\nthe stellar mass in GCs scales linearly with progenitor halo mass at $z=6$\nabove a minimum halo mass for GC formation. This model reproduces the observed\n$M_{\\rm GCs}-M_{\\rm halo}$ relation at $z=0$ and results in a prediction for\nthe minimum halo mass at $z=6$ required for hosting one GC: $M_{\\rm\nmin}(z=6)=1.07 \\times 10^9\\,M_{\\odot}$. Translated to $z=0$, the mean threshold\nmass is $M_{\\rm halo}(z=0) \\approx 2\\times 10^{10}\\,M_{\\odot}$. I explore the\nobservability of GCs in the reionization era and their contribution to cosmic\nreionization, both of which depend sensitively on the (unknown) ratio of GC\nbirth mass to present-day stellar mass, $\\xi$. Based on current detections of\n$z \\gtrsim 6$ objects with $M_{1500} < -17$, values of $\\xi > 10$ are strongly\ndisfavored; this, in turn, has potentially important implications for GC\nformation scenarios. Even for low values of $\\xi$, some observed high-$z$\ngalaxies may actually be GCs, complicating estimates of reionization-era galaxy\nultraviolet luminosity functions and constraints on dark matter models. GCs are\nlikely important reionization sources if $5 \\lesssim \\xi \\lesssim 10$. I also\nexplore predictions for the fraction of accreted versus in situ GCs in the\nlocal Universe and for descendants of systems at the halo mass threshold of GC\nformation (dwarf galaxies). An appealing feature of the model presented here is\nthe ability to make predictions for GC properties based solely on dark matter\nhalo merger trees.\n",
"title": "The Globular Cluster - Dark Matter Halo Connection"
} | null | null | null | null | true | null | 2275 | null | Default | null | null |
null | {
"abstract": " Convolutional neural networks (CNNs) have become the dominant neural network\narchitecture for solving many state-of-the-art (SOA) visual processing tasks.\nEven though Graphical Processing Units (GPUs) are most often used in training\nand deploying CNNs, their power efficiency is less than 10 GOp/s/W for\nsingle-frame runtime inference. We propose a flexible and efficient CNN\naccelerator architecture called NullHop that implements SOA CNNs useful for\nlow-power and low-latency application scenarios. NullHop exploits the sparsity\nof neuron activations in CNNs to accelerate the computation and reduce memory\nrequirements. The flexible architecture allows high utilization of available\ncomputing resources across kernel sizes ranging from 1x1 to 7x7. NullHop can\nprocess up to 128 input and 128 output feature maps per layer in a single pass.\nWe implemented the proposed architecture on a Xilinx Zynq FPGA platform and\npresent results showing how our implementation reduces external memory\ntransfers and compute time in five different CNNs ranging from small ones up to\nthe widely known large VGG16 and VGG19 CNNs. Post-synthesis simulations using\nMentor Modelsim in a 28nm process with a clock frequency of 500 MHz show that\nthe VGG19 network achieves over 450 GOp/s. By exploiting sparsity, NullHop\nachieves an efficiency of 368%, maintains over 98% utilization of the MAC\nunits, and achieves a power efficiency of over 3TOp/s/W in a core area of\n6.3mm$^2$. As further proof of NullHop's usability, we interfaced its FPGA\nimplementation with a neuromorphic event camera for real time interactive\ndemonstrations.\n",
"title": "NullHop: A Flexible Convolutional Neural Network Accelerator Based on Sparse Representations of Feature Maps"
} | null | null | null | null | true | null | 2276 | null | Default | null | null |
null | {
"abstract": " The positive impacts of platooning on travel time reliability, congestion,\nemissions, and energy consumption have been shown for homogeneous roadway\nsegments. However, speed limit changes frequently throughout the transportation\nnetwork, due to either safety-related considerations (e.g., workzone\noperations) or congestion management schemes (e.g., speed harmonization\nsystems). These abrupt changes in speed limit can result in shock- wave\nformation and cause travel time unreliability. Therefore, designing a\nplatooning strategy for tracking a reference velocity profile is critical to\nenabling end-to-end platooning. Accordingly, this study introduces a\ngeneralized control model to track a desired velocity profile, while ensuring\nsafety in the platoon of autonomous vehicles. We define appropriate natural\nerror terms and the target curve in the state space of the control system,\nwhich is the set of points where all error terms vanish and corresponds to the\ncase when all vehicles move with the desired velocities and in the minimum safe\ndistance between them. In this way, we change the tracking velocity profile\nproblem into a state- feedback stabilization problem with respect to the target\ncurve. Under certain mild assumptions on the Lipschitz constant of the speed\ndrop profile, we show that the stabilizing feedback can be obtained via\nintroducing a natural dynamics for the maximum of the error terms for each\nvehicle. Moreover, we show that with this stabilizing feedback collisions will\nnot occur if the initial state of the system of vehicles is sufficiently close\nto the target curve. We also show that the error terms remain bounded\nthroughout the time and space. Two scenarios were simulated, with and without\ninitial perturbations, and results confirmed the effectiveness of the proposed\ncontrol model in tracking the speed drop while ensuring safety and string\nstability.\n",
"title": "Platooning in the Presence of a Speed Drop: A Generalized Control Model"
} | null | null | [
"Computer Science"
]
| null | true | null | 2277 | null | Validated | null | null |
null | {
"abstract": " This paper studies the role of dg-Lie algebroids in derived deformation\ntheory. More precisely, we provide an equivalence between the homotopy theories\nof formal moduli problems and dg-Lie algebroids over a commutative dg-algebra\nof characteristic zero. At the level of linear objects, we show that the\ncategory of representations of a dg-Lie algebroid is an extension of the\ncategory of quasi-coherent sheaves on the corresponding formal moduli problem.\nWe describe this extension geometrically in terms of pro-coherent sheaves.\n",
"title": "Koszul duality for Lie algebroids"
} | null | null | null | null | true | null | 2278 | null | Default | null | null |
null | {
"abstract": " This paper presents a novel method to reduce the scale drift for indoor\nmonocular simultaneous localization and mapping (SLAM). We leverage the prior\nknowledge that in the indoor environment, the line segments form tight\nclusters, e.g. many door frames in a straight corridor are of the same shape,\nsize and orientation, so the same edges of these door frames form a tight line\nsegment cluster. We implement our method in the popular ORB-SLAM2, which also\nserves as our baseline. In the front end we detect the line segments in each\nframe and incrementally cluster them in the 3D space. In the back end, we\noptimize the map imposing the constraint that the line segments of the same\ncluster should be the same. Experimental results show that our proposed method\nsuccessfully reduces the scale drift for indoor monocular SLAM.\n",
"title": "Semi-Semantic Line-Cluster Assisted Monocular SLAM for Indoor Environments"
} | null | null | [
"Computer Science"
]
| null | true | null | 2279 | null | Validated | null | null |
null | {
"abstract": " Artificial ice systems have unique physical properties promising for\npotential applications. One of the most challenging issues in this field is to\nfind novel ice systems that allows a precise control over the geometries and\nmany-body interactions. Superconducting vortex matter has been proposed as a\nvery suitable candidate to study artificial ice, mainly due to availability of\ntunable vortex-vortex interactions and the possibility to fabricate a variety\nof nanoscale pinning potential geometries. So far, a detailed imaging of the\nlocal configurations in a vortex-based artificial ice system is still lacking.\nHere we present a direct visualization of the vortex ice state in a\nnanostructured superconductor. By using the scanning Hall probe microscopy, a\nlarge area with the vortex ice ground state configuration has been detected,\nwhich confirms the recent theoretical predictions for this new ice system.\nBesides the defects analogous to artificial spin ice systems, other types of\ndefects have been visualized and identified. We also demonstrate the\npossibility to realize different types of defects by varying the magnetic\nfield.\n",
"title": "Direct visualization of vortex ice in a nanostructured superconductor"
} | null | null | null | null | true | null | 2280 | null | Default | null | null |
null | {
"abstract": " Node-perturbation learning is a type of statistical gradient descent\nalgorithm that can be applied to problems where the objective function is not\nexplicitly formulated, including reinforcement learning. It estimates the\ngradient of an objective function by using the change in the object function in\nresponse to the perturbation. The value of the objective function for an\nunperturbed output is called a baseline. Cho et al. proposed node-perturbation\nlearning with a noisy baseline. In this paper, we report on building the\nstatistical mechanics of Cho's model and on deriving coupled differential\nequations of order parameters that depict learning dynamics. We also show how\nto derive the generalization error by solving the differential equations of\norder parameters. On the basis of the results, we show that Cho's results are\nalso apply in general cases and show some general performances of Cho's model.\n",
"title": "Statistical Mechanics of Node-perturbation Learning with Noisy Baseline"
} | null | null | null | null | true | null | 2281 | null | Default | null | null |
null | {
"abstract": " Recently $S_{b}$-metric spaces have been introduced as the generalizations of\nmetric and $S$-metric spaces. In this paper we investigate some basic\nproperties of this new space. We generalize the classical Banach's contraction\nprinciple using the theory of a complete $S_{b}$-metric space. Also we give an\napplication to linear equation systems using the $S_{b}$-metric which is\ngenerated by a metric.\n",
"title": "New Generalized Fixed Point Results on $S_{b}$-Metric Spaces"
} | null | null | null | null | true | null | 2282 | null | Default | null | null |
null | {
"abstract": " Designing decentralized policies for wireless communication networks is a\ncrucial problem, which has only been partially solved in the literature so far.\nIn this paper, we propose the Decentralized Markov Decision Process (Dec-MDP)\nframework to analyze a wireless sensor network with multiple users which access\na common wireless channel. We consider devices with energy harvesting\ncapabilities, so that they aim at balancing the energy arrivals with the data\ndepartures and with the probability of colliding with other nodes. Randomly\nover time, an access point triggers a SYNC slot, wherein it recomputes the\noptimal transmission parameters of the whole network, and distributes this\ninformation. Every node receives its own policy, which specifies how it should\naccess the channel in the future, and, thereafter, proceeds in a fully\ndecentralized fashion, without interacting with other entities in the network.\nWe propose a multi-layer Markov model, where an external MDP manages the jumps\nbetween SYNC slots, and an internal Dec-MDP computes the optimal policy in the\nnear future. We numerically show that, because of the harvesting, a fully\northogonal scheme (e.g., TDMA-like) is suboptimal in energy harvesting\nscenarios, and the optimal trade-off lies between an orthogonal and a random\naccess system.\n",
"title": "A Decentralized Optimization Framework for Energy Harvesting Devices"
} | null | null | null | null | true | null | 2283 | null | Default | null | null |
null | {
"abstract": " In this paper we propose and analyze a finite element method for both the\nharmonic map heat and Landau--Lifshitz--Gilbert equation, the time variable\nremaining continuous. Our starting point is to set out a unified saddle point\napproach for both problems in order to impose the unit sphere constraint at the\nnodes since the only polynomial function satisfying the unit sphere constraint\neverywhere are constants. A proper inf-sup condition is proved for the Lagrange\nmultiplier leading to the well-posedness of the unified formulation. \\emph{A\npriori} energy estimates are shown for the proposed method.\nWhen time integrations are combined with the saddle point finite element\napproximation some extra elaborations are required in order to ensure both\n\\emph{a priori} energy estimates for the director or magnetization vector\ndepending on the model and an inf-sup condition for the Lagrange multiplier.\nThis is due to the fact that the unit length at the nodes is not satisfied in\ngeneral when a time integration is performed. We will carry out a linear Euler\ntime-stepping method and a non-linear Crank--Nicolson method. The latter is\nsolved by using the former as a non-linear solver.\n",
"title": "Inf-sup stable finite-element methods for the Landau--Lifshitz--Gilbert and harmonic map heat flow equation"
} | null | null | null | null | true | null | 2284 | null | Default | null | null |
null | {
"abstract": " This paper investigates power control and relay selection in Full Duplex\nCognitive Relay Networks (FDCRNs), where the secondary-user (SU) relays can\nsimultaneously receive and forward the signal from the SU source. We study both\nnon-coherent and coherent scenarios. In the non-coherent case, the SU relay\nforwards the signal from the SU source without regulating the phase, while in\nthe coherent scenario, the SU relay regulates the phase when forwarding the\nsignal to minimize the interference at the primary-user (PU) receiver. We\nconsider the problem of maximizing the transmission rate from the SU source to\nthe SU destination subject to the interference constraint at the PU receiver\nand power constraints at both the SU source and SU relay. We develop\nlow-complexity and high-performance joint power control and relay selection\nalgorithms. The superior performance of the proposed algorithms are confirmed\nusing extensive numerical evaluation. In particular, we demonstrate the\nsignificant gain of phase regulation at the SU relay (i.e., the gain of the\ncoherent mechanism over the noncoherent mechanism).\n",
"title": "Power Control and Relay Selection in Full-Duplex Cognitive Relay Networks: Coherent versus Non-coherent Scenarios"
} | null | null | null | null | true | null | 2285 | null | Default | null | null |
null | {
"abstract": " We introduce and study ternary $f$-distributive structures, Ternary\n$f$-quandles and more generally their higher $n$-ary analogues. A\nclassification of ternary $f$-quandles is provided in low dimensions. Moreover,\nwe study extension theory and introduce a cohomology theory for ternary, and\nmore generally $n$-ary, $f$-quandles. Furthermore, we give some computational\nexamples.\n",
"title": "Ternary and $n$-ary $f$-distributive Structures"
} | null | null | null | null | true | null | 2286 | null | Default | null | null |
null | {
"abstract": " The important unsolved problem in theory of integrable systems is to find\nconditions guaranteeing existence of a Lax representation for a given PDE. The\nuse of the exotic cohomology of the symmetry algebras opens a way to formulate\nsuch conditions in internal terms of the PDEs under the study. In this paper we\nconsider certain examples of infinite-dimensional Lie algebras with nontrivial\nsecond exotic cohomology groups and show that the Maurer-Cartan forms of the\nassociated extensions of these Lie algebras generate Lax representations for\nintegrable systems, both known and new ones.\n",
"title": "Deformations of infinite-dimensional Lie algebras, exotic cohomology, and integrable nonlinear partial differential equations"
} | null | null | null | null | true | null | 2287 | null | Default | null | null |
null | {
"abstract": " In this work we study the impact of chromatic focusing of few-cycle laser\npulses on high-order harmonic generation (HHG) through analysis of the emitted\nextreme ultraviolet (XUV) radiation. Chromatic focusing is usually avoided in\nthe few-cycle regime, as the pulse spatio-temporal structure may be highly\ndistorted by the spatiotemporal aberrations. Here, however, we demonstrate it\nas an additional control parameter to modify the generated XUV radiation. We\npresent experiments where few-cycle pulses are focused by a singlet lens in a\nKr gas jet. The chromatic distribution of focal lengths allows us to tune HHG\nspectra by changing the relative singlet-target distance. Interestingly, we\nalso show that the degree of chromatic aberration needed to this control does\nnot degrade substantially the harmonic conversion efficiency, still allowing\nfor the generation of supercontinua with the chirped-pulse scheme, demonstrated\npreviously for achromatic focussing. We back up our experiments with\ntheoretical simulations reproducing the experimental HHG results depending on\ndiverse parameters (input pulse spectral phase, pulse duration, focus position)\nand proving that, under the considered parameters, the attosecond pulse train\nremains very similar to the achromatic case, even showing cases of isolated\nattosecond pulse generation for near single-cycle driving pulses.\n",
"title": "Tunable high-harmonic generation by chromatic focusing of few-cycle laser pulses"
} | null | null | null | null | true | null | 2288 | null | Default | null | null |
null | {
"abstract": " Anisotropy describes the directional dependence of a material's properties\nsuch as transport and optical response. In conventional bulk materials,\nanisotropy is intrinsically related to the crystal structure, and thus not\ntunable by the gating techniques used in modern electronics. Here we show that,\nin bilayer black phosphorus with an interlayer twist angle of 90°, the\nanisotropy of its electronic structure and optical transitions is tunable by\ngating. Using first-principles calculations, we predict that a\nlaboratory-accessible gate voltage can induce a hole effective mass that is 30\ntimes larger along one Cartesian axis than along the other axis, and the two\naxes can be exchanged by flipping the sign of the gate voltage. This\ngate-controllable band structure also leads to a switchable optical linear\ndichroism, where the polarization of the lowest-energy optical transitions\n(absorption or luminescence) is tunable by gating. Thus, anisotropy is a\ntunable degree of freedom in twisted bilayer black phosphorus.\n",
"title": "Gate Switchable Transport and Optical Anisotropy in 90° Twisted Bilayer Black Phosphorus"
} | null | null | null | null | true | null | 2289 | null | Default | null | null |
null | {
"abstract": " Logistic linear mixed model is widely used in experimental designs and\ngenetic analysis with binary traits. Motivated by modern applications, we\nconsider the case with many groups of random effects and each group corresponds\nto a variance component. When the number of variance components is large,\nfitting the logistic linear mixed model is challenging. We develop two\nefficient and stable minorization-maximization (MM) algorithms for the\nestimation of variance components based on the Laplace approximation of the\nlogistic model. One of them leads to a simple iterative soft-thresholding\nalgorithm for variance component selection using maximum penalized approximated\nlikelihood. We demonstrate the variance component estimation and selection\nperformance of our algorithms by simulation studies and a real data analysis.\n",
"title": "MM Algorithms for Variance Component Estimation and Selection in Logistic Linear Mixed Model"
} | null | null | [
"Statistics"
]
| null | true | null | 2290 | null | Validated | null | null |
null | {
"abstract": " We prove that any non-adaptive algorithm that tests whether an unknown\nBoolean function $f: \\{0, 1\\}^n\\to \\{0, 1\\}$ is a $k$-junta or $\\epsilon$-far\nfrom every $k$-junta must make $\\widetilde{\\Omega}(k^{3/2} / \\epsilon)$ many\nqueries for a wide range of parameters $k$ and $\\epsilon$. Our result\ndramatically improves previous lower bounds from [BGSMdW13, STW15], and is\nessentially optimal given Blais's non-adaptive junta tester from [Blais08],\nwhich makes $\\widetilde{O}(k^{3/2})/\\epsilon$ queries. Combined with the\nadaptive tester of [Blais09] which makes $O(k\\log k + k /\\epsilon)$ queries,\nour result shows that adaptivity enables polynomial savings in query complexity\nfor junta testing.\n",
"title": "Settling the query complexity of non-adaptive junta testing"
} | null | null | null | null | true | null | 2291 | null | Default | null | null |
null | {
"abstract": " Language understanding is a key component in a spoken dialogue system. In\nthis paper, we investigate how the language understanding module influences the\ndialogue system performance by conducting a series of systematic experiments on\na task-oriented neural dialogue system in a reinforcement learning based\nsetting. The empirical study shows that among different types of language\nunderstanding errors, slot-level errors can have more impact on the overall\nperformance of a dialogue system compared to intent-level errors. In addition,\nour experiments demonstrate that the reinforcement learning based dialogue\nsystem is able to learn when and what to confirm in order to achieve better\nperformance and greater robustness.\n",
"title": "Investigation of Language Understanding Impact for Reinforcement Learning Based Dialogue Systems"
} | null | null | null | null | true | null | 2292 | null | Default | null | null |
null | {
"abstract": " A novel delay-based spacing policy for the control of vehicle platoons is\nintroduced together with a notion of disturbance string stability. The\ndelay-based spacing policy specifies the desired inter-vehicular distance\nbetween vehicles and guarantees that all vehicles track the same spatially\nvarying reference velocity profile, as is for example required for heavy-duty\nvehicles driving over hilly terrain. Disturbance string stability is a notion\nof string stability of vehicle platoons subject to external disturbances on all\nvehicles that guarantees that perturbations do not grow unbounded as they\npropagate through the platoon. Specifically, a control design approach in the\nspatial domain is presented that achieves tracking of the desired spacing\npolicy and guarantees disturbance string stability with respect to a spatially\nvarying reference velocity. The results are illustrated by means of\nsimulations.\n",
"title": "String stability and a delay-based spacing policy for vehicle platoons subject to disturbances"
} | null | null | null | null | true | null | 2293 | null | Default | null | null |
null | {
"abstract": " The Reidemeister number of an endomorphism of a group is the number of\ntwisted conjugacy classes determined by that endomorphism. The collection of\nall Reidemeister numbers of all automorphisms of a group $G$ is called the\nReidemeister spectrum of $G$. In this paper, we determine the Reidemeister\nspectra of all fundamental groups of solvmanifolds up to Hirsch length 4.\n",
"title": "Reidemeister spectra for solvmanifolds in low dimensions"
} | null | null | null | null | true | null | 2294 | null | Default | null | null |
null | {
"abstract": " In this paper, we propose a framework for cross-layer optimization to ensure\nultra-high reliability and ultra-low latency in radio access networks, where\nboth transmission delay and queueing delay are considered. With short\ntransmission time, the blocklength of channel codes is finite, and the Shannon\nCapacity cannot be used to characterize the maximal achievable rate with given\ntransmission error probability. With randomly arrived packets, some packets may\nviolate the queueing delay. Moreover, since the queueing delay is shorter than\nthe channel coherence time in typical scenarios, the required transmit power to\nguarantee the queueing delay and transmission error probability will become\nunbounded even with spatial diversity. To ensure the required\nquality-of-service (QoS) with finite transmit power, a proactive packet\ndropping mechanism is introduced. Then, the overall packet loss probability\nincludes transmission error probability, queueing delay violation probability,\nand packet dropping probability. We optimize the packet dropping policy, power\nallocation policy, and bandwidth allocation policy to minimize the transmit\npower under the QoS constraint. The optimal solution is obtained, which depends\non both channel and queue state information. Simulation and numerical results\nvalidate our analysis, and show that setting packet loss probabilities equal is\na near optimal solution.\n",
"title": "Cross-layer Optimization for Ultra-reliable and Low-latency Radio Access Networks"
} | null | null | null | null | true | null | 2295 | null | Default | null | null |
null | {
"abstract": " Many important problems can be modeled as a system of interconnected\nentities, where each entity is recording time-dependent observations or\nmeasurements. In order to spot trends, detect anomalies, and interpret the\ntemporal dynamics of such data, it is essential to understand the relationships\nbetween the different entities and how these relationships evolve over time. In\nthis paper, we introduce the time-varying graphical lasso (TVGL), a method of\ninferring time-varying networks from raw time series data. We cast the problem\nin terms of estimating a sparse time-varying inverse covariance matrix, which\nreveals a dynamic network of interdependencies between the entities. Since\ndynamic network inference is a computationally expensive task, we derive a\nscalable message-passing algorithm based on the Alternating Direction Method of\nMultipliers (ADMM) to solve this problem in an efficient way. We also discuss\nseveral extensions, including a streaming algorithm to update the model and\nincorporate new observations in real time. Finally, we evaluate our TVGL\nalgorithm on both real and synthetic datasets, obtaining interpretable results\nand outperforming state-of-the-art baselines in terms of both accuracy and\nscalability.\n",
"title": "Network Inference via the Time-Varying Graphical Lasso"
} | null | null | null | null | true | null | 2296 | null | Default | null | null |
null | {
"abstract": " A functional version of the Kato one-parametric regularisation for the\nconstruction of a dynamical semigroup generator of a relative bound one\nperturbation is introduced. It does not require that the minus generator of the\nunperturbed semigroup is a positivity preserving operator. The regularisation\nis illustrated by an example of a boson-number cut-off regularisation.\n",
"title": "Construction of dynamical semigroups by a functional regularisation à la Kato"
} | null | null | null | null | true | null | 2297 | null | Default | null | null |
null | {
"abstract": " Recent breakthroughs in computer vision and natural language processing have\nspurred interest in challenging multi-modal tasks such as visual\nquestion-answering and visual dialogue. For such tasks, one successful approach\nis to condition image-based convolutional network computation on language via\nFeature-wise Linear Modulation (FiLM) layers, i.e., per-channel scaling and\nshifting. We propose to generate the parameters of FiLM layers going up the\nhierarchy of a convolutional network in a multi-hop fashion rather than all at\nonce, as in prior work. By alternating between attending to the language input\nand generating FiLM layer parameters, this approach is better able to scale to\nsettings with longer input sequences such as dialogue. We demonstrate that\nmulti-hop FiLM generation achieves state-of-the-art for the short input\nsequence task ReferIt --- on-par with single-hop FiLM generation --- while also\nsignificantly outperforming prior state-of-the-art and single-hop FiLM\ngeneration on the GuessWhat?! visual dialogue task.\n",
"title": "Visual Reasoning with Multi-hop Feature Modulation"
} | null | null | [
"Statistics"
]
| null | true | null | 2298 | null | Validated | null | null |
null | {
"abstract": " Objective: To evaluate unsupervised clustering methods for identifying\nindividual-level behavioral-clinical phenotypes that relate personal biomarkers\nand behavioral traits in type 2 diabetes (T2DM) self-monitoring data. Materials\nand Methods: We used hierarchical clustering (HC) to identify groups of meals\nwith similar nutrition and glycemic impact for 6 individuals with T2DM who\ncollected self-monitoring data. We evaluated clusters on: 1) correspondence to\ngold standards generated by certified diabetes educators (CDEs) for 3\nparticipants; 2) face validity, rated by CDEs, and 3) impact on CDEs' ability\nto identify patterns for another 3 participants. Results: Gold standard (GS)\nincluded 9 patterns across 3 participants. Of these, all 9 were re-discovered\nusing HC: 4 GS patterns were consistent with patterns identified by HC (over\n50% of meals in a cluster followed the pattern); another 5 were included as\nsub-groups in broader clusers. 50% (9/18) of clusters were rated over 3 on\n5-point Likert scale for validity, significance, and being actionable. After\nreviewing clusters, CDEs identified patterns that were more consistent with\ndata (70% reduction in contradictions between patterns and participants'\nrecords). Discussion: Hierarchical clustering of blood glucose and\nmacronutrient consumption appears suitable for discovering behavioral-clinical\nphenotypes in T2DM. Most clusters corresponded to gold standard and were rated\npositively by CDEs for face validity. Cluster visualizations helped CDEs\nidentify more robust patterns in nutrition and glycemic impact, creating new\npossibilities for visual analytic solutions. Conclusion: Machine learning\nmethods can use diabetes self-monitoring data to create personalized\nbehavioral-clinical phenotypes, which may prove useful for delivering\npersonalized medicine.\n",
"title": "Behavioral-clinical phenotyping with type 2 diabetes self-monitoring data"
} | null | null | null | null | true | null | 2299 | null | Default | null | null |
null | {
"abstract": " We report structural, optical, temperature and frequency dependent\ndielectric, and energy storage properties of pulsed laser deposited (100)\nhighly textured BaZr(x)Ti(1-x)O3 (x = 0.3, 0.4 and 0.5) relaxor ferroelectric\nthin films on La0.7Sr0.3MnO3/MgO substrates which make this compound as a\npotential lead-free capacitive energy storage material for scalable electronic\ndevices. A high dielectric constant of ~1400 - 3500 and a low dielectric loss\nof <0.025 were achieved at 10 kHz for all three compositions at ambient\nconditions. Ultrahigh stored and recoverable electrostatic energy densities as\nhigh as 214 +/- 1 and 156 +/- 1 J/cm3, respectively, were demonstrated at a\nsustained high electric field of ~3 MV/cm with an efficiency of 72.8 +/- 0.6 %\nin optimum 30% Zr substituted BaTiO3 composition.\n",
"title": "Ultrahigh capacitive energy storage in highly oriented BaZr(x)Ti(1-x)O3 thin films prepared by pulsed laser deposition"
} | null | null | [
"Physics"
]
| null | true | null | 2300 | null | Validated | null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.