diff --git "a/adjusted_data_1.json" "b/adjusted_data_1.json" deleted file mode 100644--- "a/adjusted_data_1.json" +++ /dev/null @@ -1,1132 +0,0 @@ -[ - { - "abstract": " Sparse structures are frequently sought when pursuing tractability in\noptimization problems. They are exploited from both theoretical and\ncomputational perspectives to handle complex problems that become manageable\nwhen sparsity is present. An example of this type of structure is given by\ntreewidth: a graph theoretical parameter that measures how \"tree-like\" a graph\nis. This parameter has been used for decades for analyzing the complexity of\nvarious optimization problems and for obtaining tractable algorithms for\nproblems where this parameter is bounded. The goal of this work is to\ncontribute to the understanding of the limits of the treewidth-based\ntractability in optimization. Our results are as follows. First, we prove that,\nin a certain sense, the already known positive results on extension complexity\nbased on low treewidth are the best possible. Secondly, under mild assumptions,\nwe prove that treewidth is the only graph-theoretical parameter that yields\ntractability a wide class of optimization problems, a fact well known in\nGraphical Models in Machine Learning and in Constraint Satisfaction Problems,\nwhich here we extend to an approximation setting in Optimization.\n", - "method": "Our results are as follows.", - "id": "1807.02551" - }, - { - "abstract": " The place and role of parsing analysis in formation of professional\ninformatics competences of future informatics teachers is determined. Separated\nautomation tools for lexical (lex) and syntax (yacc) analysis invariant to the\nprogramming language used. The expediency of using functional programming\nlanguages Scheme and SML is shown for learning how to develop compilers in the\ncourse of programming theory. The example of the MosML dialect illustrates the\nmain components of the methodic of joint using the tools of automation of\nlexical and parsing analysis in the process of teaching the programming theory\nof future informatics teachers. The main conclusions and recommendations: 1)\nthe considered example of the expanded calculator can be refined by changing\nthe grammar, in particular - for the introduction of conditional and cyclic\nconstructions; 2) the proposed scheme can be used to implement the interpreter\nof any formal language with an arbitrary typing method - the appropriate\nexamples of study will be subsets of procedural languages Basic and C and\nfunctional languages Scheme and SML: provided the addition of the machine code\ngeneration phase, this provides an opportunity to demonstrate the full\ndevelopment cycle for programming language compiler.\n", - "method": "The main conclusions and recommendations: 1)\nthe considered example of the expanded calculator can be refined by changing\nthe grammar, in particular - for the introduction of conditional and cyclic\nconstructions; 2) the proposed scheme can be used to implement the interpreter\nof any formal language with an arbitrary typing method - the appropriate\nexamples of study will be subsets of procedural languages Basic and C and\nfunctional languages Scheme and SML: provided the addition of the machine code\ngeneration phase, this provides an opportunity to demonstrate the full\ndevelopment cycle for programming language compiler.", - "id": "1807.02554" - }, - { - "abstract": " We introduce an approximation technique for nonlinear hyperbolic systems with\nsources that is invariant domain preserving. The method is\ndiscretization-independent provided elementary symmetry and skew-symmetry\nproperties are satisfied by the scheme. The method is formally first-order\naccurate in space. A series of higher-order methods is also introduced. When\nthese methods violate the invariant domain properties, they are corrected by a\nlimiting technique that we call convex limiting. After limiting, the resulting\nmethods satisfy all the invariant domain properties that are imposed by the\nuser (see Theorem~7.24). A key novelty is that the bounds that are enforced on\nthe solution at each time step are necessarily satisfied by the low-order\napproximation.\n", - "method": "A series of higher-order methods is also introduced.", - "id": "1807.02563" - }, - { - "abstract": " A huge optical luminosity of the supercritical accretion disc and powerful\nstellar wind in the high-mass X-ray binary SS433 make it difficult to reliably\nestimate the mass ratio of the binary components from spectroscopic\nobservations. We analyze different indirect methods of the mass ratio estimate.\nWe show that with an account of the possible Roche lobe overflow by the optical\nstar, the analysis of X-ray eclipses in the standard and hard X-ray bands\nsuggests the estimate $q=M_\\mathrm{x}/M_\\mathrm{v}\\gtrsim 0.3$. We argue that\nthe double-peak hydrogen Brackett lines in SS433 should form not in the\naccretion disc but in a circumbinary envelope, suggesting a total mass of\n$M_\\mathrm{v}+M_\\mathrm{x}\\gtrsim 40 M_\\odot$. The observed long-term stability\nof the orbital period in SS433 $|\\dot P_b/P_b|\\le 1.793\\times\n10^{-14}$~s$^{-1}$ over $\\sim 28$ year period is used to place an independent\nconstraint of $q\\gtrsim 0.6$ in SS433, confirming its being a Galactic\nmicroquasar hosting a superaccreting black hole.\n", - "method": "We analyze different indirect methods of the mass ratio estimate.", - "id": "1807.02626" - }, - { - "abstract": " We extend results of Bongiovanni et al. on double bubbles on the line with\nlog-convex density to the case where the derivative of the log of the density\nis bounded. We show that the tie function between the double interval and the\ntriple interval still exists but may blow up to infinity in finite time. For\nthe first time, a density is presented for which the blowup time is positive\nand finite.\n", - "method": "We extend results of Bongiovanni et al.", - "id": "1807.02661" - }, - { - "abstract": " Domination game [SIAM J.\\ Discrete Math.\\ 24 (2010) 979--991] and total\ndomination game [Graphs Combin.\\ 31 (2015) 1453--1462] are by now well\nestablished games played on graphs by two players, named Dominator and Staller.\nIn this paper, Z-domination game, L-domination game, and LL-domination game are\nintroduced as natural companions of the standard domination games.\n Versions of the Continuation Principle are proved for the new games. It is\nproved that in each of these games the outcome of the game, which is a\ncorresponding graph invariant, differs by at most one depending whether\nDominator or Staller starts the game. The hierarchy of the five domination\ngames is established. The invariants are also bounded with respect to the\n(total) domination number and to the order of a graph. Values of the three new\ninvariants are determined for paths up to a small constant independent from the\nlength of a path. Several open problems and a conjecture are listed. The latter\nasserts that the L-domination game number is not greater than $6/7$ of the\norder of a graph.\n", - "method": "Versions of the Continuation Principle are proved for the new games.", - "id": "1807.02695" - }, - { - "abstract": " In permissive environments, E. coli can double its dry mass every 21 minutes.\nDuring this time, ribosomes, RNA polymerases, and the proteome are all doubled.\nYet, the question of how to relate bacterial doubling time to other\nbiologically relevant time scales in the growth process remains illusive, due\nto the complex temporal nesting pattern of these processes. In particular, the\nrelation between the cell's doubling time and the ribosome assembly time is not\nknown. Here we develop a model that connects growth rate to ribosome assembly\ntime and show that the existence of a self-assembly step increases the overall\ngrowth rate, because during ribosome self-assembly existing ribosomes can start\na new round of reproduction, by making a new batch of ribosomal proteins prior\nto the completion of the previous round. This overlapping of ribosome\nreproduction cycles increases growth rate beyond the serial-limit that is\ntypically assumed to hold. Using recent data from ribosome profiling and well\nknown measurements of the average translation rate, rigid bounds on the in-vivo\nribosome self-assembly time are set, which are robust to the assumptions\nregarding the biological noises involved. At 21 minutes doubling time, the\nribosome assembly time is found to be approximately 6 minutes --- three fold\nlarger than the common estimate. We further use our model to explain the\ndetrimental effect of a recently discovered ribosome assembly inhibitor drug,\nand predict the effect of limiting the expression of ribosome assembly\nchaperons on the overall growth rate.\n", - "method": "During this time, ribosomes, RNA polymerases, and the proteome are all doubled.", - "id": "1807.02771" - }, - { - "abstract": " An automatic program that generates constant profit from the financial market\nis lucrative for every market practitioner. Recent advance in deep\nreinforcement learning provides a framework toward end-to-end training of such\ntrading agent. In this paper, we propose an Markov Decision Process (MDP) model\nsuitable for the financial trading task and solve it with the state-of-the-art\ndeep recurrent Q-network (DRQN) algorithm. We propose several modifications to\nthe existing learning algorithm to make it more suitable under the financial\ntrading setting, namely 1. We employ a substantially small replay memory (only\na few hundreds in size) compared to ones used in modern deep reinforcement\nlearning algorithms (often millions in size.) 2. We develop an action\naugmentation technique to mitigate the need for random exploration by providing\nextra feedback signals for all actions to the agent. This enables us to use\ngreedy policy over the course of learning and shows strong empirical\nperformance compared to more commonly used epsilon-greedy exploration. However,\nthis technique is specific to financial trading under a few market assumptions.\n3. We sample a longer sequence for recurrent neural network training. A side\nproduct of this mechanism is that we can now train the agent for every T steps.\nThis greatly reduces training time since the overall computation is down by a\nfactor of T. We combine all of the above into a complete online learning\nalgorithm and validate our approach on the spot foreign exchange market.\n", - "method": "We propose several modifications to\nthe existing learning algorithm to make it more suitable under the financial\ntrading setting, namely 1.. We sample a longer sequence for recurrent neural network training.", - "id": "1807.02787" - }, - { - "abstract": " The prevalence of different kinds of electronic devices and the volume of\ncontent on the Web have increased the amount of plagiarism, which is considered\nan unethical act. If we want to be efficient in the detection and prevention of\nthese acts, we have to improve today's methods of discovering plagiarism. The\npaper presents a research study where a framework for the improved detection of\nplagiarism is proposed. The framework focuses on the integration of social\nnetwork information, information from the Web, and an advanced semantically\nenriched visualization of information about authors and documents that enables\nthe exploration of obtained data by seeking of advanced patterns of plagiarism.\nTo support the proposed framework, a special software tool was also developed.\nThe statistical evaluation confirmed that the employment of social network\nanalysis and advanced visualization techniques led to improvements in the\nconfirmation and investigation stages of the plagiarism detection process,\nthereby enhancing the overall efficiency of the plagiarism detection process.\n", - "method": "To support the proposed framework, a special software tool was also developed.", - "id": "1807.02830" - }, - { - "abstract": " A new forecasting method based on the concept of the profile predictive the\nlikelihood function is proposed for discrete-valued processes. In particular,\ngeneralized autoregressive and moving average (GARMA) models for Poisson\ndistributed data are explored in details. Highest density regions are used to\nconstruct forecasting regions. The proposed forecast estimates and regions are\ncoherent. Large sample results are derived for the forecasting distribution.\nNumerical studies using simulations and a real data set are used to establish\nthe performance of the proposed forecasting method. Robustness of the proposed\nmethod to possible misspecification in the model is also studied.\n", - "method": "Large sample results are derived for the forecasting distribution.", - "id": "1807.02850" - }, - { - "abstract": " We consider the following distributed service model: jobs with unit mean,\ngeneral distribution, and independent processing times arrive as a renewal\nprocess of rate $\\lambda n$, with $0<\\lambda<1$, and are immediately dispatched\nto one of several queues associated with $n$ identical servers with unit\nprocessing rate. We assume that the dispatching decisions are made by a central\ndispatcher endowed with a finite memory, and with the ability to exchange\nmessages with the servers.\n We study the fundamental resource requirements (memory bits and message\nexchange rate), in order to drive the expected queueing delay in steady-state\nof a typical job to zero, as $n$ increases. We develop a novel approach to show\nthat, within a certain broad class of \"symmetric\" policies, every dispatching\npolicy with a message rate of the order of $n$, and with a memory of the order\nof $\\log n$ bits, results in an expected queueing delay which is bounded away\nfrom zero, uniformly as $n\\to\\infty$.\n", - "method": "We consider the following distributed service model: jobs with unit mean,\ngeneral distribution, and independent processing times arrive as a renewal\nprocess of rate $\\lambda n$, with $0<\\lambda<1$, and are immediately dispatched\nto one of several queues associated with $n$ identical servers with unit\nprocessing rate.", - "id": "1807.02882" - }, - { - "abstract": " We carried out multi-color optical monitoring of a sample of ten blazars from\n2005 to 2011. The sample contains 3 LBLs, 2 IBLs, 4 HBLs, and 1 FSRQ. Our\nmonitoring focused on the long-term variability and the sample included nine BL\nLac objects and one flat-spectrum radio quasar. A total number of 14799 data\npoints were collected. This is one of the largest optical database for a sample\nof ten blazars. All objects showed significant variability except OT 546.\nBecause of the low sampling on each single night, only BL Lacertae was observed\nto have intra-day variability on 2006 November 6. Most BL Lac objects showed a\nbluer-when-brighter chromatism, while the flat-spectrum radio quasar, 3C 454.3,\ndisplayed a redder-when-brighter trend. The BWB color behaviors of most BL Lacs\ncan be at least partly interpreted by the fact of increasing variation\namplitude with increasing frequency observed in these objects. The average\nspectral index of LBLs is around 1.5, as expected from the model dominated by\nSynchrotron Self-Compton (SSC) loss. The optical emission of HBL is probably\ncontaminated by the thermal emission from the host galaxies. Correlation\nanalysis did not reveal any time delay between variations at different\nwavelengths.\n", - "method": "All objects showed significant variability except OT 546.", - "id": "1807.02926" - }, - { - "abstract": " I suggest a new extension of the standard model of particle physics, which\nintroduces a dark sector with the $SU(2)_{D}\\otimes U(1)_{D'}$ symmetry besides\nthe SM sector. The new particles of the model all inhabit in the dark sector.\nThe dark gauge symmetry breaking will bring about fruitful physics beyond the\nSM. The tiny neutrino mass is generated through the Dirac-type seesaw\nmechanism. The inflaton decay can not only provide the universe inflation and\nreheating, but also lead to the baryon asymmetry and the asymmetric cold dark\nmatter. In short, the model provides an unification of the neutrino mass, the\nbaryon asymmetry, the asymmetric CDM and the inflation, and it can account for\ntheir common origin. Finally, it is very possible to test the model predictions\nand probe the dark sector physics in near future experiments.\n", - "method": "The new particles of the model all inhabit in the dark sector.", - "id": "1807.03036" - }, - { - "abstract": " We report the observations of a moderate but relatively intense geo-effective\nsolar eruption on 2015 November 4 from the peripheral diffusive polarities of\nactive region 12443. We use space-borne Solar Dynamics Observatory and ACE\nobservations. EUV images identified helical pattern along a filament channel\nand we regard this channel as flux-rope structure. Flow velocity derived from\ntracked magnetograms infers converging motion along the polarity inversion line\nbeneath the filament channel. An associated magnetic cancellation process was\ndetected in the converging region. Further, the pre-eruptive EUV brightening\nwas observed in the converging region, the most intense part of which appeared\nin the magnetic cancellation region. These observations imply that the\nconverging and cancelling flux probably contributed to the formation of the\nhelical magnetic fields associated with the flux rope. A filament-height\nestimation method suggests that the middle part of the filament probably lies\nat a low altitude and was consistent with the initial place of the eruption. A\nthick current channel associated with the flux rope is also determined. For a\nexpanding thick current channel, the critical height of the decay index for\ntorus instability lies in the range of 37 - 47 Mm. Southward magnetic fields in\nthe sheath and the ejecta induced a geomagnetic storm with a Dst global minimum\nof ~-90 nT.\n", - "method": "EUV images identified helical pattern along a filament channel\nand we regard this channel as flux-rope structure.", - "id": "1807.03047" - }, - { - "abstract": " In this paper we present a system based on SVM ensembles trained on\ncharacters and words to discriminate between five similar languages of the\nIndo-Aryan family: Hindi, Braj Bhasha, Awadhi, Bhojpuri, and Magahi. We\ninvestigate the performance of individual features and combine the output of\nsingle classifiers to maximize performance. The system competed in the\nIndo-Aryan Language Identification (ILI) shared task organized within the\nVarDial Evaluation Campaign 2018. Our best entry in the competition, named\nILIdentification, scored 88:95% F1 score and it was ranked 3rd out of 8 teams.\n", - "method": "In this paper we present a system based on SVM ensembles trained on\ncharacters and words to discriminate between five similar languages of the\nIndo-Aryan family: Hindi, Braj Bhasha, Awadhi, Bhojpuri, and Magahi.", - "id": "1807.03108" - }, - { - "abstract": " Recent studies have shown evidence of a significant decline of the Posidonia\noceanica (P.O.) meadows on a global scale. The monitoring and mapping of these\nmeadows are fundamental tools for measuring their status. We present an\napproach based on a deep neural network to automatically perform a\nhigh-precision semantic segmentation of P.O. meadows in sea-floor images,\noffering several improvements over the state of the art techniques. Our network\ndemonstrates outstanding performance over diverse test sets, reaching a\nprecision of 96.57% and an accuracy of 96.81%, surpassing the reliability of\nlabelling the images manually. Also, the network is implemented in an\nAutonomous Underwater Vehicle (AUV), performing an online P.O. segmentation,\nwhich will be used to generate real-time semantic coverage maps.\n", - "method": "Recent studies have shown evidence of a significant decline of the Posidonia\noceanica (P.O.) meadows on a global scale.", - "id": "1807.03117" - }, - { - "abstract": " Horticultural enterprises are becoming more sophisticated as the range of the\ncrops they target expands. Requirements for enhanced efficiency and\nproductivity have driven the demand for automating on-field operations.\nHowever, various problems remain yet to be solved for their reliable, safe\ndeployment in real-world scenarios. This paper examines major research trends\nand current challenges in horticultural robotics. Specifically, our work\nfocuses on sensing and perception in the three main horticultural procedures:\npollination, yield estimation, and harvesting. For each task, we expose major\nissues arising from the unstructured, cluttered, and rugged nature of field\nenvironments, including variable lighting conditions and difficulties in\nfruit-specific detection, and highlight promising contemporary studies.\n", - "method": "Specifically, our work\nfocuses on sensing and perception in the three main horticultural procedures:\npollination, yield estimation, and harvesting.", - "id": "1807.03124" - }, - { - "abstract": " In this paper, the first of two, we introduce an alternative definition of\nthe Chang--Weinberger--Yu relative higher index, which is thought of as a\nrelative analogue of the Mishchenko--Fomenko index pairing. A main result of\nthis paper is that our map coincides with the existing relative higher index\nmaps. We make use of this fact for understanding the relative higher index.\nFirst, we relate the relative higher index with the higher index of amalgamated\nfree product groups. Second, we define the dual relative higher index map and\nshow its rational surjectivity under certain assumptions.\n", - "method": "We make use of this fact for understanding the relative higher index.", - "id": "1807.03181" - }, - { - "abstract": " This paper is concerned with the $L^p$ integrability of $N$-harmonic\nfunctions with respect to the standard weights $(1-|x|^2)^{\\alpha}$ on the unit\nball $\\mathbb{B}$ of $\\mathbb{R}^n$, $n\\geq 2$. More precisely, our goal is to\ndetermine the real (negative) parameters $\\alpha$, for which\n$(1-|x|^2)^{\\alpha/p} u(x) \\in L^p(\\mathbb{B})$ implies that $u\\equiv 0$,\nwhenever $u$ is a solution of the $N$-Laplace equation on $\\mathbb{B}$. This\nquestion is motivated by the uniqueness considerations of the Dirichlet problem\nfor the $N$-Laplacian $\\Delta^N$.\n Our study is inspired by a recent work of Borichev and Hedenmalm [Adv. Math.,\n264(2014), pp. 464-505], where a complete answer to the above question in the\ncase $n=2$ is given for the full scale $00$ such that the coupled system possesses\npositive ground state solution for all $\\mu\\geq\\mu_{0}$. In these cases, we use\na minimization method based on Nehari manifold. Finally, we consider the case\n$p=q=2^{*}$, and we prove that the coupled system has no positive solutions.\nFor that matter, we use a Pohozaev identity type.\n", - "method": "Finally, we consider the case\n$p=q=2^{*}$, and we prove that the coupled system has no positive solutions.. For that matter, we use a Pohozaev identity type.", - "id": "1807.03436" - }, - { - "abstract": " The empirical coupled-channel (ECC) model and the universal fusion function\n(UFF) prescription are used to analyse the data of capture cross sections for\nreactions ${}^{39}$K$+{}^{181}$Ta and ${}^{46}$K$+{}^{181}$Ta reported recently\nby A. Wakhle {\\it et al.} [Phys. Rev. C 97, 021602(R) (2018)]. The results of\nthe ECC model are in good agreement with the data of ${}^{39}$K$+{}^{181}$Ta\nwhile, for ${}^{46}$K$+{}^{181}$Ta, the predictions of the ECC model\noverestimate the above-barrier capture cross sections. Comparing the reduced\ndata of these two reactions, it is found that the above-barrier cross sections\nof ${}^{39}$K$+{}^{181}$Ta are consistent with the UFF and are larger than\nthose of ${}^{46}$K$+{}^{181}$Ta. This implies that the capture cross sections\nof ${}^{46}$K$+{}^{181}$Ta are suppressed at energies above the Coulomb\nbarrier. Furthermore, at sub-barrier energies, the reduced calculated capture\ncross sections of ${}^{39}$K$+{}^{181}$Ta are a little larger than those of\n${}^{46}$K$+{}^{181}$Ta, which is owing to the coupling to the positive\n$Q$-value two-neutron transfer channel.\n", - "method": "Wakhle {\\it et al.} [Phys.. Rev.. C 97, 021602(R) (2018)].", - "id": "1807.03441" - }, - { - "abstract": " In this paper, we aim to examine the relationship between the unstable\ncircular photon orbit and the thermodynamic phase transition for a rotating\nKerr-AdS black hole. On one side, we give a brief review of the phase\ntransition for the Kerr-AdS black hole. The coexistence curve and the\nmetastable curve corresponding to the phase transition are clearly shown. On\nthe other side, we calculate the radius and the angular momentum of the\nunstable circular orbits for a photon by analyzing the effective potential.\nThen combining these two sides, we find the following results. i) The radius\nand the angular momentum of the unstable circular photon orbits demonstrate the\nnon-monotonic behaviors when the thermodynamic phase transition takes place. So\nfrom the behavior of the circular orbit, one can determine whether there exists\na thermodynamic phase transition. ii) The difference of the radius or the\nangular momentum for the coexistence small and large black holes can be treated\nas an order parameter to describe the phase transition. And near the critical\npoint, it has a critical exponent of $\\frac{1}{2}$. iii) The temperature and\npressure corresponding to the extremal points of the radius or the angular\nmomentum of the unstable circular photon orbit completely agree with that of\nthe metastable curves from the thermodynamic side. Thus, these results confirm\nthe relationship between the geodesics and thermodynamic phase transition for\nthe Kerr-AdS black hole. Therefore, on one hand, we are allowed to probe the\nthermodynamic phase transition from the gravity side. On the other hand, the\nsignature of the strong gravitational effect can also be revealed from the\nblack hole thermodynamics.\n", - "method": "Then combining these two sides, we find the following results.", - "id": "1807.03455" - }, - { - "abstract": " Scientific applications consist of large and computationally-intensive loops.\nDynamic loop scheduling (DLS) techniques are used to load balance the execution\nof such applications. Load imbalance can be caused by variations in loop\niteration execution times due to problem, algorithmic, or systemic\ncharacteristics (also, perturbations). The following question motivates this\nwork: \"Given an application, a high-performance computing (HPC) system, and\nboth their characteristics and interplay, which DLS technique will achieve\nimproved performance under unpredictable perturbations?\" Existing work only\nconsiders perturbations caused by variations in the HPC system delivered\ncomputational speeds. However, perturbations in available network bandwidth or\nlatency are inevitable on production HPC systems. Simulator in the loop (SiL)\nis introduced, herein, as a new control-theoretic inspired approach to\ndynamically select DLS techniques that improve the performance of applications\non heterogeneous HPC systems under perturbations. The present work examines the\nperformance of six applications on a heterogeneous system under all above\nsystem perturbations. The SiL proof of concept is evaluated using simulation.\nThe performance results confirm the initial hypothesis that no single DLS\ntechnique can deliver best performance in all scenarios, while the SiL-based\nDLS selection delivered improved application performance in most experiments.\n", - "method": "The SiL proof of concept is evaluated using simulation.", - "id": "1807.03577" - }, - { - "abstract": " We have developed a method to estimate surface gravity (log g) from light\ncurves by measuring the granulation background, similar to the \"flicker\" method\nby Bastien et al. (2016) but working in the Fourier power spectrum. We\ncalibrated the method using Kepler stars for which asteroseismology has been\npossible with short-cadence data, demonstrating a precision in log g of about\n0.05 dex. We also derived a correction for white noise as a function of Kepler\nmagnitude by measuring white noise directly from observations. We then applied\nthe method to the same sample of long-cadence stars as Bastien et al. We found\nthat about half the stars are too faint for the granulation background to be\nreliably detected above the white noise. We provide a catalogue of log g values\nfor about 15,000 stars having uncertainties better than 0.5 dex. We used Gaia\nDR2 parallaxes to validate that granulation is a powerful method to measure log\ng from light curves. Our method can also be applied to the large number of\nlight curves collected by K2 and TESS.\n", - "method": "(2016) but working in the Fourier power spectrum.", - "id": "1807.03624" - }, - { - "abstract": " In the restricted Santa Claus problem we are given resources $\\mathcal R$ and\nplayers $\\mathcal P$. Every resource $j\\in\\mathcal R$ has a value $v_j$ and\nevery player $i$ desires a set $\\mathcal R(i)$ of resources. We are interested\nin distributing the resources to players that desire them. The quality of a\nsolution is measured by the least happy player, i.e., the lowest sum of\nresource values. This value should be maximized. The local search algorithm by\nAsadpour et al. and its connection to the configuration LP has proved itself to\nbe a very influential technique for this and related problems. In the original\nproof, a local search was used to obtain a bound of $4$ for the ratio of the\nfractional to the integral optimum of the configuration LP (integrality gap).\nThis bound is non-constructive since the local search has not been shown to\nterminate in polynomial time. On the negative side, the worst instance known\nhas an integrality gap of $2$. Although much progress was made in this area,\nneither bound has been improved since. We present a better analysis that shows\nthe integrality gap is not worse than $3 + 5/6 \\approx 3.8333$.\n", - "method": "This value should be maximized.", - "id": "1807.03626" - }, - { - "abstract": " In this paper, we introduce a new class of stochastic multilayer networks. A\nstochastic multilayer network is the aggregation of $M$ networks (one per\nlayer) where each is a subgraph of a foundational network $G$. Each layer\nnetwork is the result of probabilistically removing links and nodes from $G$.\nThe resulting network includes any link that appears in at least $K$ layers.\nThis model is an instance of a non-standard site-bond percolation model. Two\nsets of results are obtained: first, we derive the probability distribution\nthat the $M$-layer network is in a given configuration for some particular\ngraph structures (explicit results are provided for a line and an algorithm is\nprovided for a tree), where a configuration is the collective state of all\nlinks (each either active or inactive). Next, we show that for appropriate\nscalings of the node and link selection processes in a layer, links are\nasymptotically independent as the number of layers goes to infinity, and follow\nPoisson distributions. Numerical results are provided to highlight the impact\nof having several layers on some metrics of interest (including expected size\nof the cluster a node belongs to in the case of the line). This model finds\napplications in wireless communication networks with multichannel radios,\nmultiple social networks with overlapping memberships, transportation networks,\nand, more generally, in any scenario where a common set of nodes can be linked\nvia co-existing means of connectivity.\n", - "method": "This model is an instance of a non-standard site-bond percolation model.", - "id": "1807.03650" - }, - { - "abstract": " Information extraction traditionally focuses on extracting relations between\nidentifiable entities, such as . Yet, texts\noften also contain Counting information, stating that a subject is in a\nspecific relation with a number of objects, without mentioning the objects\nthemselves, for example, \"California is divided into 58 counties\". Such\ncounting quantifiers can help in a variety of tasks such as query answering or\nknowledge base curation, but are neglected by prior work. This paper develops\nthe first full-fledged system for extracting counting information from text,\ncalled CINEX. We employ distant supervision using fact counts from a knowledge\nbase as training seeds, and develop novel techniques for dealing with several\nchallenges: (i) non-maximal training seeds due to the incompleteness of\nknowledge bases, (ii) sparse and skewed observations in text sources, and (iii)\nhigh diversity of linguistic patterns. Experiments with five human-evaluated\nrelations show that CINEX can achieve 60% average precision for extracting\ncounting information. In a large-scale experiment, we demonstrate the potential\nfor knowledge base enrichment by applying CINEX to 2,474 frequent relations in\nWikidata. CINEX can assert the existence of 2.5M facts for 110 distinct\nrelations, which is 28% more than the existing Wikidata facts for these\nrelations.\n", - "method": "CINEX can assert the existence of 2.5M facts for 110 distinct\nrelations, which is 28% more than the existing Wikidata facts for these\nrelations.", - "id": "1807.03656" - }, - { - "abstract": " Aim. The unusual transient ATLAS17aeu was serendipitously detected within the\nsky localisation of the gravitational wave trigger GW170104. The importance of\na possible association with gravitational waves coming from a binary black hole\nmerger led to an extensive follow-up campaign, with the aim of assessing a\npossible connection with GW170104. Methods. With several telescopes, we carried\nout both photometric and spectroscopic observations of ATLAS17aeu, for several\nepochs, between $\\sim 3$ and $\\sim 230$ days after the first detection.\nResults. We studied in detail the temporal and spectroscopic properties of\nATLAS17aeu and its host galaxy. We detected spectral features similar to those\nof a broad lined supernova superposed to an otherwise typical long-GRB\nafterglow. Based on analysis of the optical light curve, spectrum and host\ngalaxy SED, we conclude that the redshift of the source is probably $z \\simeq\n0.5 \\pm 0.2$. Conclusions. While the redshift range we have determined is\nmarginally compatible with that of the gravitational wave event, the presence\nof a supernova component and the consistency of this transient with the E$_{\\rm\np}$-E$_{\\rm iso}$ correlation support the conclusion that ATLAS17aeu was\nassociated with the long gamma-ray burst GRB170105A. This rules out the\nassociation of the GRB170105A/ATLAS17aeu transient with the gravitational wave\nevent GW170104, which was due to a binary black hole merger.\n", - "method": "Methods.", - "id": "1807.03681" - }, - { - "abstract": " The number of diagrams of stationary points free vector fields in the 2-disk\n$\\mathbb{B}^{2}$ is counted in the article. It is shown that the number of such\ndiagrams with $2k$ exceptional points on the boundary $\\mathbb{S}^{1}$ equals\n$3^{k-2}(C_{k}+2C_{k-1})$, where $C_{k}$ is the corresponding Catalan number.\nAn algorithm for finding all such diagrams is discussed.\n", - "method": "An algorithm for finding all such diagrams is discussed.", - "id": "1807.03714" - }, - { - "abstract": " We compute the homology of the spaces in the Omega spectrum for $BoP$. There\nis no torsion in $H_*(\\underline{BoP}_{\\; i})$ for $i \\ge 2$, and things are\nonly slightly more complicated for $i < 2$. We find the complete homotopy type\nof $\\underline{BoP}_{\\; i}$ for $i \\le 6$ and conjecture the homotopy type for\n$i > 6$. This completes the computation of all $H_*(\\underline{MSU}_{\\;*})$.\n", - "method": "We compute the homology of the spaces in the Omega spectrum for $BoP$.", - "id": "1807.03738" - }, - { - "abstract": " Yield and its prediction is one of the most important tasks in grapevine\nbreeding purposes and vineyard management. Commonly, this trait is estimated\nmanually right before harvest by extrapolation, which mostly is\nlabor-intensive, destructive and inaccurate. In the present study an automated\nimage-based workflow was developed quantifying inflorescences and single\nflowers in unprepared field images of grapevines, i.e. no artificial background\nor light was applied. It is a novel approach for non-invasive, inexpensive and\nobjective phenotyping with high-throughput.\n First, image regions depicting inflorescences were identified and localized.\nThis was done by segmenting the images into the classes \"inflorescence\" and\n\"non-inflorescence\" using a Fully Convolutional Network (FCN). Efficient image\nsegmentation hereby is the most challenging step regarding the small geometry\nand dense distribution of flowers (several hundred flowers per inflorescence),\nsimilar color of all plant organs in the fore- and background as well as the\ncircumstance that only approximately 5% of an image show inflorescences. The\ntrained FCN achieved a mean Intersection Over Union (IOU) of 87.6% on the test\ndata set. Finally, individual flowers were extracted from the\n\"inflorescence\"-areas using Circular Hough Transform. The flower extraction\nachieved a recall of 80.3% and a precision of 70.7% using the segmentation\nderived by the trained FCN model.\n Summarized, the presented approach is a promising strategy in order to\npredict yield potential automatically in the earliest stage of grapevine\ndevelopment which is applicable for objective monitoring and evaluations of\nbreeding material, genetic repositories or commercial vineyards.\n", - "method": "First, image regions depicting inflorescences were identified and localized.", - "id": "1807.03770" - }, - { - "abstract": " As with EAST (Experimental Advanced Superconducting Tokamak) experimental\ndata analyzed by more and more collaborators, the experimental videos which\ndirectly reflect the real status of vacuum attract more and more researchers'\nattention. The real time VOD (Video On Demand) system based on MDSplus allows\nusers reading the video frames in real time as same as the signal data which is\nalso stored in the MDSplus database. User can display the plasma discharge\nvideos and analyze videos frame by frame through jScope or our VOD web station.\nThe system mainly includes the frames storing and frames displaying. The frames\nstoring application accepts shot information by using socket TCP communication\nfirstly, then reads video frames through disk mapping, finally stores them into\nMDSplus. The displaying process is implemented through B/S (Browser/Server)\nframework, it uses PHP and JavaScript to realize VOD function and read frames\ninformation from MDSplus. The system offers a unit way to access and backup\nexperimental data and video during the EAST experiment, which is of great\nbenefit to EAST experimenter than the formal VOD system in VOD function and\nreal time performance.\n", - "method": "The system mainly includes the frames storing and frames displaying.", - "id": "1807.03773" - }, - { - "abstract": " We propose a new sequential classification model for astronomical objects\nbased on a recurrent convolutional neural network (RCNN) which uses sequences\nof images as inputs. This approach avoids the computation of light curves or\ndifference images. This is the first time that sequences of images are used\ndirectly for the classification of variable objects in astronomy. The second\ncontribution of this work is the image simulation process. We generate\nsynthetic image sequences that take into account the instrumental and observing\nconditions, obtaining a realistic, set of movies for each astronomical object.\nThe simulated dataset is used to train our RCNN classifier. This approach\nallows us to generate datasets to train and test our RCNN model for different\nastronomical surveys and telescopes. We aim at building a simulated dataset\nwhose distribution is close enough to the real dataset, so that a fine tuning\ncould match the distributions between real and simulated dataset. To test the\nRCNN classifier trained with the synthetic dataset, we used real-world data\nfrom the High cadence Transient Survey (HiTS) obtaining an average recall of\n85%, improved to 94% after performing fine tuning with 10 real samples per\nclass. We compare the results of our model with those of a light curve random\nforest classifier. The proposed RCNN with fine tuning has a similar performance\non the HiTS dataset compared to the light curve classifier, trained on an\naugmented training set with 10 real samples per class. The RCNN approach\npresents several advantages in an alert stream classification scenario, such as\na reduction of the data pre-processing, faster online evaluation and easier\nperformance improvement using a few real data samples. These results encourage\nus to use this method for alert brokers systems that will process alert streams\ngenerated by new telescopes such as the Large Synoptic Survey Telescope.\n", - "method": "The simulated dataset is used to train our RCNN classifier.", - "id": "1807.03869" - }, - { - "abstract": " In this study, we address the challenge of measuring the ability of a\nrecommender system to make surprising recommendations. Although current\nevaluation methods make it possible to determine if two algorithms can make\nrecommendations with a significant difference in their average surprise\nmeasure, it could be of interest to our community to know how competent an\nalgorithm is at embedding surprise in its recommendations, without having to\nresort to making a direct comparison with another algorithm. We argue that a)\nsurprise is a finite resource in a recommender system, b) there is a limit to\nhow much surprise any algorithm can embed in a recommendation, and c) this\nlimit can provide us with a scale against which the performance of any\nalgorithm can be measured. By exploring these ideas, it is possible to define\nthe concepts of maximum and minimum potential surprise and design a surprise\nmetric called \"normalised surprise\" that employs these limits to potential\nsurprise. Two experiments were conducted to test the proposed metric. The aim\nof the first was to validate the quality of the estimates of minimum and\nmaximum potential surprise produced by a greedy algorithm. The purpose of the\nsecond experiment was to analyse the behaviour of the proposed metric using the\nMovieLens dataset. The results confirmed the behaviour that was expected, and\nshowed that the proposed surprise metric is both effective and consistent for\ndiffering choices of recommendation algorithms, data representations and\ndistance functions.\n", - "method": "Two experiments were conducted to test the proposed metric.", - "id": "1807.03905" - }, - { - "abstract": " Let Sn denote the network of all RNA secondary structures of length n, in\nwhich undirected edges exist between structures s, t such that t is obtained\nfrom s by the addition, removal or shift of a single base pair. Using\ncontext-free grammars, generating functions and complex analysis, we show that\nthe asymptotic average degree is O(n) and that the asymptotic clustering\ncoeffcient is O(1/n), from which it follows that the family Sn, n = 1,2,3,...\nof secondary structure networks is not small-world.\n", - "method": "of secondary structure networks is not small-world.", - "id": "1807.03911" - }, - { - "abstract": " We show that symbolic finite-to-one extensions of the type constructed by O.\nSarig for surface diffeomorphisms induce H\\\"older-continuous conjugacies on\nlarge sets. We deduce this from their Bowen property. This notion, introduced\nin a joint work with M. Boyle, generalizes a fact first observed by R.\\ Bowen\nfor Markov partitions. We rely on the notion of degree from finite equivalence\ntheory and magic word isomorphisms. As an application, we give lower bounds on\nthe number of periodic points first for surface diffeomorphisms (improving a\nresult of Sarig) and for Sinai billiards maps (building on a result of Baladi\nand Demers). Finally we characterize surface diffeomorphisms admitting a\nH\\\"older-continuous coding of all their aperiodic hyperbolic measures and give\na slightly weaker construction preserving local compactness.\n", - "method": "We deduce this from their Bowen property.", - "id": "1807.04017" - }, - { - "abstract": " There are two major routes to address the ubiquitous family of inverse\nproblems appearing in signal and image processing, such as denoising or\ndeblurring. A first route relies on Bayesian modeling, where prior\nprobabilities are used to embody models of both the distribution of the unknown\nvariables and their statistical dependence with the observed data. The\nestimation process typically relies on the minimization of an expected loss\n(e.g. minimum mean squared error, or MMSE). The second route has received much\nattention in the context of sparse regularization and compressive sensing: it\nconsists in designing (often convex) optimization problems involving the sum of\na data fidelity term and a penalty term promoting certain types of unknowns\n(e.g., sparsity, promoted through an 1 norm). Well known relations between\nthese two approaches have lead to some widely spread misconceptions. In\nparticular, while the so-called Maximum A Posterori (MAP) estimate with a\nGaussian noise model does lead to an optimization problem with a quadratic\ndata-fidelity term, we disprove through explicit examples the common belief\nthat the converse would be true. It has already been shown [7, 9] that for\ndenoising in the presence of additive Gaussian noise, for any prior probability\non the unknowns, MMSE estimation can be expressed as a penalized least squares\nproblem, with the apparent characteristics of a MAP estimation problem with\nGaussian noise and a (generally) different prior on the unknowns. In other\nwords, the variational approach is rich enough to build all possible MMSE\nestimators associated to additive Gaussian noise via a well chosen penalty. We\ngeneralize these results beyond Gaussian denoising and characterize noise\nmodels for which the same phenomenon occurs. In particular, we prove that with\n(a variant of) Poisson noise and any prior probability on the unknowns, MMSE\nestimation can again be expressed as the solution of a penalized least squares\noptimization problem. For additive scalar denois-ing the phenomenon holds if\nand only if the noise distribution is log-concave. In particular, Laplacian\ndenoising can (perhaps surprisingly) be expressed as the solution of a\npenalized least squares problem. In the multivariate case, the same phenomenon\noccurs when the noise model belongs to a particular subset of the exponential\nfamily. For multivariate additive denoising, the phenomenon holds if and only\nif the noise is white and Gaussian.\n", - "method": "The\nestimation process typically relies on the minimization of an expected loss\n(e.g. minimum mean squared error, or MMSE).", - "id": "1807.04021" - }, - { - "abstract": " We extend our previous study of Markov chains on finite commutative rings\n(arXiv:1605.05089) to arbitrary finite rings with identity. At each step, we\neither add or multiply by a randomly chosen element of the ring, where the\naddition (resp. multiplication) distribution is uniform (resp. conjugacy\ninvariant). We prove explicit formulas for some of the eigenvalues of the\ntransition matrix and give lower bounds on their multiplicities. We also give\nrecursive formulas for the stationary distribution and prove that the mixing\ntime is bounded by an absolute constant. For the matrix rings $M_2(\\mathbb\nF_q),$ we compute the entire spectrum explicitly using the representation\ntheory of $\\text{GL}_2(\\mathbb F_q),$ as well as the stationary probabilities.\n", - "method": "multiplication) distribution is uniform (resp.", - "id": "1807.04082" - }, - { - "abstract": " The key to any nameless representation of syntax is how it indicates the\nvariables we choose to use and thus, implicitly, those we discard. Standard de\nBruijn representations delay discarding maximally till the leaves of terms\nwhere one is chosen from the variables in scope at the expense of the rest.\nConsequently, introducing new but unused variables requires term traversal.\nThis paper introduces a nameless 'co-de-Bruijn' representation which makes the\nopposite canonical choice, delaying discarding minimally, as near as possible\nto the root. It is literate Agda: dependent types make it a practical joy to\nexpress and be driven by strong intrinsic invariants which ensure that scope is\naggressively whittled down to just the support of each subterm, in which every\nremaining variable occurs somewhere. The construction is generic, delivering a\nuniverse of syntaxes with higher-order metavariables, for which the appropriate\nnotion of substitution is hereditary. The implementation of simultaneous\nsubstitution exploits tight scope control to avoid busywork and shift terms\nwithout traversal. Surprisingly, it is also intrinsically terminating, by\nstructural recursion alone.\n", - "method": "Consequently, introducing new but unused variables requires term traversal.", - "id": "1807.04085" - }, - { - "abstract": " It is well known that many types of artificial neural networks, including\nrecurrent networks, can achieve a high classification accuracy even with\nlow-precision weights and activations. The reduction in precision generally\nyields much more efficient hardware implementations in regards to hardware\ncost, memory requirements, energy, and achievable throughput. In this paper, we\npresent the first systematic exploration of this design space as a function of\nprecision for Bidirectional Long Short-Term Memory (BiLSTM) neural network.\nSpecifically, we include an in-depth investigation of precision vs. accuracy\nusing a fully hardware-aware training flow, where during training quantization\nof all aspects of the network including weights, input, output and in-memory\ncell activations are taken into consideration. In addition, hardware resource\ncost, power consumption and throughput scalability are explored as a function\nof precision for FPGA-based implementations of BiLSTM, and multiple approaches\nof parallelizing the hardware. We provide the first open source HLS library\nextension of FINN for parameterizable hardware architectures of LSTM layers on\nFPGAs which offers full precision flexibility and allows for parameterizable\nperformance scaling offering different levels of parallelism within the\narchitecture. Based on this library, we present an FPGA-based accelerator for\nBiLSTM neural network designed for optical character recognition, along with\nnumerous other experimental proof points for a Zynq UltraScale+ XCZU7EV MPSoC\nwithin the given design space.\n", - "method": "Specifically, we include an in-depth investigation of precision vs.", - "id": "1807.04093" - }, - { - "abstract": " We generalise and improve a result of Stoll, Walsh and Yuan by showing that\nthere are at most two solutions in coprime positive integers of the equation in\nthe title when $b=p^{m}$ where $m$ is a non-negative integer, $p$ is prime,\n$(a,p)=1$, $a^{2}+p^{2m}$ not a perfect square and $x^{2}- \\left( a^{2}+p^{2m}\n\\right) y^{2}=-1$ has an integer solution. This result is best possible. We\nalso obtain best possible results for all positive integer solutions when $m=1$\nand $2$.\n When $b$ is an arbitrary square with $(a,b)=1$ and $a^{2}+b^{2}$ not a\nperfect square, we are able to prove there are at most three solutions in\ncoprime positive integers provided $x^{2}- \\left( a^{2}+b^{2} \\right) y^{2}=-1$\nhas an integer solution and $x^{2}- \\left( a^{2}+b^{2} \\right) y^{2}=-b^{2}$\nhas only one family of solutions. Our proof is based on a novel use of the\nhypergeometric method that may also be useful for other problems.\n", - "method": "This result is best possible.", - "id": "1807.04116" - }, - { - "abstract": " Understanding systems level behaviour of many interacting agents is\nchallenging in various ways, here we'll focus on the how the interaction\nbetween components can lead to hierarchical structures with different types of\ndynamics, or causations, at different levels. We use the Tangled Nature model\nto discuss the co-evolutionary aspects connecting the microscopic level of the\nindividual to the macroscopic systems level. At the microscopic level the\nindividual agent may undergo evolutionary changes due to mutations of\nstrategies. The micro-dynamics always run at a constant rate. Nevertheless, the\nsystem's level dynamics exhibit a completely different type of intermittent\nabrupt dynamics where major upheavals keep throwing the system between\nmeta-stable configurations. These dramatic transitions are described by a\nlog-Poisson time statistics. The long time effect is a collectively adapted of\nthe ecological network. We discuss the ecological and macroevolutionary\nconsequences of the adaptive dynamics and briefly describe work using the\nTangled Nature framework to analyse problems in economics, sociology,\ninnovation and sustainability\n", - "method": "The micro-dynamics always run at a constant rate.", - "id": "1807.04228" - }, - { - "abstract": " Motivated by the Koml\\'os conjecture in combinatorial discrepancy, we study\nthe discrepancy of random matrices with $m$ rows and $n$ independent columns\ndrawn from a bounded lattice random variable. It is known that for $n$ tending\nto infinity and $m$ fixed, with high probability the $\\ell_\\infty$-discrepancy\nis at most twice the $\\ell_\\infty$-covering radius of the integer span of the\nsupport of the random variable. However, the easy argument for the above fact\ngives no concrete bounds on the failure probability in terms of $n$. We prove\nthat the failure probability is inverse polynomial in $m, n$ and some\nwell-motivated parameters of the random variable. We also obtain the analogous\nbounds for the discrepancy in arbitrary norms.\n We apply these results to two random models of interest. For random\n$t$-sparse matrices, i.e. uniformly random matrices with $t$ ones and $m-t$\nzeroes in each column, we show that the $\\ell_\\infty$-discrepancy is at most\n$2$ with probability $1 - O(\\sqrt{ \\log n/n})$ for $n = \\Omega(m^3 \\log^2 m)$.\nThis improves on a bound proved by Ezra and Lovett (Ezra and Lovett,\nApprox+Random, 2016) showing that the same is true for $n$ at least $m^t$. For\nmatrices with random unit vector columns, we show that the\n$\\ell_\\infty$-discrepancy is $O(\\exp(\\sqrt{ n/m^3}))$ with probability $1 -\nO(\\sqrt{ \\log n/n})$ for $n = \\Omega(m^3 \\log^2 m).$ Our approach, in the\nspirit of Kuperberg, Lovett and Peled (G. Kuperberg, S. Lovett and R. Peled,\nSTOC 2012), uses Fourier analysis to prove that for $m \\times n$ matrices $M$\nwith i.i.d. columns, and $m$ sufficiently large, the distribution of $My$ for\nrandom $y \\in \\{-1,1\\}^n$ obeys a local limit theorem.\n", - "method": "Kuperberg, S.. Lovett and R.", - "id": "1807.04318" - }, - { - "abstract": " We consider two quantization approaches to the Bateman oscillator model. One\nis Feshbach-Tikochinsky's quantization approach reformulated concisely without\ninvoking the ${\\mathit{SU}(1,1)}$ Lie algebra, and the other is the\nimaginary-scaling quantization approach developed originally for the\nPais-Uhlenbeck oscillator model. The latter approach overcomes the problem of\nunbounded-below energy spectrum that is encountered in the former approach. In\nboth the approaches, the positive-definiteness of the squared-norms of the\nHamiltonian eigenvectors is ensured. Unlike Feshbach-Tikochinsky's quantization\napproach, the imaginary-scaling quantization approach allows to have stable\nstates in addition to decaying and growing states.\n", - "method": "We consider two quantization approaches to the Bateman oscillator model.", - "id": "1807.04403" - }, - { - "abstract": " We analyze the security of two multipartite quantum key distribution (QKD)\nprotocols, specifically we introduce an $N$-partite version of the BB84\nprotocol and we discuss the $N$-partite six-state protocol proposed in\narXiv:1612.05585v2. The security analysis proceeds from the generalization of\nknown results in bipartite QKD to the multipartite scenario, and takes into\naccount finite resources. In this context we derive a computable expression for\nthe achievable key rate of both protocols by employing the best-known\nstrategies: the uncertainty relation and the postselection technique. We\ncompare the performances of the two protocols both for finite resources and\ninfinitely many signals.\n", - "method": "We analyze the security of two multipartite quantum key distribution (QKD)\nprotocols, specifically we introduce an $N$-partite version of the BB84\nprotocol and we discuss the $N$-partite six-state protocol proposed in\narXiv:1612.05585v2.", - "id": "1807.04472" - }, - { - "abstract": " We consider the quantum linear solver for $Ax=b$ with the circulant\npreconditioner $C$. The main technique is the singular value estimation (SVE)\nintroduced in [I. Kerenidis and A. Prakash, Quantum recommendation system, in\nITCS 2017]. However, some modifications of SVE should be made to solve the\npreconditioned linear system $C^{-1} Ax = C^{-1} b$. Moreover, different from\nthe preconditioned linear system considered in [B. D. Clader, B. C. Jacobs, C.\nR. Sprouse, Preconditioned quantum linear system algorithm, Phys. Rev. Lett.,\n2013], the circulant preconditioner is easy to construct and can be directly\napplied to general dense non-Hermitian cases. The time complexity depends on\nthe condition numbers of $C$ and $C^{-1} A$, as well as the Frobenius norm\n$\\|A\\|_F$.\n", - "method": "Kerenidis and A.", - "id": "1807.04563" - }, - { - "abstract": " We present a wormhole solution in four dimensions. It is a solution of an\nEinstein Maxwell theory plus charged massless fermions. The fermions give rise\nto a negative Casimir-like energy, which makes the wormhole possible. It is a\nlong wormhole that does not lead to causality violations in the ambient space.\nIt can be viewed as a pair of entangled near extremal black holes with an\ninteraction term generated by the exchange of fermion fields. The solution can\nbe embedded in the Standard Model by making its overall size small compared to\nthe electroweak scale.\n", - "method": "We present a wormhole solution in four dimensions.", - "id": "1807.04726" - }, - { - "abstract": " We study a form of refined class number formula (resp. type number formula)\nfor maximal orders in totally definite quaternion algebras over real quadratic\nfields, by taking into consideration the automorphism groups of right ideal\nclasses (resp. unit groups of maximal orders). For each finite noncyclic group\n$G$, we give an explicit formula for the number of conjugacy classes of maximal\norders whose unit groups modulo center are isomorphic to $G$, and write down a\nrepresentative for each conjugacy class. This leads to a complete recipe (even\nexplicit formulas in special cases) for the refined class number formula for\nall finite groups. As an application, we prove the existence of superspecial\nabelian surfaces whose endomorphism algebras coincide with\n$\\mathbb{Q}(\\sqrt{p})$ in all positive characteristic $p\\not\\equiv 1\\pmod{24}$.\n", - "method": "We study a form of refined class number formula (resp.. unit groups of maximal orders).", - "id": "1807.04736" - }, - { - "abstract": " Finite volume simulations of turbulent boundary layer flow over an\naxisymmetric hill are performed for $Re_H$ = 6500 using Morphing Continuum\nTheory (MCT), and compared with DNS data from Castagna et. al. and experimental\ndata obtained by Simpson. The inlet profile was specified by inputting values\nfrom the profile specified by Castagna et. al. Root- mean-square velocity\nfluctuations are inputted using the new variable of gyration from MCT.\nAdditional terms introduced by MCT lead to a new formulation of the\nQ-criterion, which allows for the visualization of three-dimensional turbulent\nstructures including a hairpin vortices. Streamline plots of the separation\nbubble show a more confined bubble than Castagna et. al., but agree well on the\nseparation and reconnection points. The surface pressure coefficient matches\nSimpson much more closely than Castagna et. al. MCT data was obtained on a\n$6.72 \\times 10^6$ cell mesh containing a smaller number of cells by an order\nof magnitude than that of the DNS mesh of $5.4 \\times 10^7$ by Castagna et. al.\nThe effects of the particle on the larger structures are inferred from the\nvariation in the new variable of gyration.\n", - "method": "al.", - "id": "1807.04793" - }, - { - "abstract": " We give an $\\tilde{O}(n^{7/5} \\log (nC))$-time algorithm to compute a\nminimum-cost maximum cardinality matching (optimal matching) in $K_h$-minor\nfree graphs with $h=O(1)$ and integer edge weights having magnitude at most\n$C$. This improves upon the $\\tilde{O}(n^{10/7}\\log{C})$ algorithm of Cohen et\nal. [SODA 2017] and the $O(n^{3/2}\\log (nC))$ algorithm of Gabow and Tarjan\n[SIAM J. Comput. 1989].\n For a graph with $m$ edges and $n$ vertices, the well-known Hungarian\nAlgorithm computes a shortest augmenting path in each phase in $O(m)$ time,\nyielding an optimal matching in $O(mn)$ time. The Hopcroft-Karp [SIAM J.\nComput. 1973], and Gabow-Tarjan [SIAM J. Comput. 1989] algorithms compute, in\neach phase, a maximal set of vertex-disjoint shortest augmenting paths (for\nappropriately defined costs) in $O(m)$ time. This reduces the number of phases\nfrom $n$ to $O(\\sqrt{n})$ and the total execution time to $O(m\\sqrt{n})$.\n In order to obtain our speed-up, we relax the conditions on the augmenting\npaths and iteratively compute, in each phase, a set of carefully selected\naugmenting paths that are not restricted to be shortest or vertex-disjoint. As\na result, our algorithm computes substantially more augmenting paths in each\nphase, reducing the number of phases from $O(\\sqrt{n})$ to $O(n^{2/5})$. By\nusing small vertex separators, the execution of each phase takes $\\tilde{O}(m)$\ntime on average. For planar graphs, we combine our algorithm with efficient\nshortest path data structures to obtain a minimum-cost perfect matching in\n$\\tilde{O}(n^{6/5} \\log{(nC)})$ time. This improves upon the recent\n$\\tilde{O}(n^{4/3}\\log{(nC)})$ time algorithm by Asathulla et al. [SODA 2018].\n", - "method": "The Hopcroft-Karp [SIAM J.. Comput.. 1973], and Gabow-Tarjan [SIAM J.. Comput.", - "id": "1807.04802" - }, - { - "abstract": " The abstraction of the study of stochastic processes to Banach lattices and\nvector lattices has received much attention by Grobler, Kuo, Labuschagne,\nStoica, Troitsky and Watson over the past fifteen years. By contrast mixing\nprocesses have received very little attention. In particular mixingales were\ngeneralized to the Riesz space setting in {\\sc W.-C. Kuo, J.J. Vardy, B.A.\nWatson,} Mixingales on Riesz spaces, {\\em J. Math. Anal. Appl.}, \\textbf{402}\n(2013), 731-738. The concepts of strong and uniform mixing as well as related\nmixing inequalities were extended to this setting in {\\sc W.-C. Kuo, M.J.\nRogans, B.A. Watson,} Mixing inequalities in Riesz spaces, {\\em J. Math. Anal.\nAppl.}, \\textbf{456} (2017), 992-1004. In the present work we formulate the\nconcept of near-epoch dependence for Riesz space processes and show that if a\nprocess is near-epoch dependent and either strong or uniform mixing then the\nprocess is a mixingale, giving access to a law of large numbers. The above is\napplied to autoregessive processes of order 1 in Riesz spaces.\n", - "method": "Kuo, J.J. Vardy, B.A.\nWatson,} Mixingales on Riesz spaces, {\\em J.. Kuo, J.J. Vardy, B.A.\nWatson,} Mixingales on Riesz spaces, {\\em J.. Math.. Anal.. Kuo, M.J.\nRogans, B.A. Watson,} Mixing inequalities in Riesz spaces, {\\em J.. Kuo, M.J.\nRogans, B.A. Watson,} Mixing inequalities in Riesz spaces, {\\em J.. Math.. Anal.", - "id": "1807.04869" - }, - { - "abstract": " We have conducted a deep survey (with a central rms of $55\\mu\\textrm{Jy}$)\nwith the LOw Frequency ARray (LOFAR) at 120-168 MHz of the Bo\\\"otes field, with\nan angular resolution of $3.98^{''}\\times6.45^{''}$, and obtained a sample of\n10091 radio sources ($5\\sigma$ limit) over an area of $20\\:\\textrm{deg}^{2}$.\nThe astrometry and flux scale accuracy of our source catalog is investigated.\nThe resolution bias, incompleteness and other systematic effects that could\naffect our source counts are discussed and accounted for. The derived 150 MHz\nsource counts present a flattening below sub-mJy flux densities, that is in\nagreement with previous results from high- and low- frequency surveys. This\nflattening has been argued to be due to an increasing contribution of\nstar-forming galaxies and faint active galactic nuclei. Additionally, we use\nour observations to evaluate the contribution of cosmic variance to the scatter\nin source counts measurements. The latter is achieved by dividing our Bo\\\"otes\nmosaic into 10 non-overlapping circular sectors, each one with an approximate\narea of $2\\:\\textrm{deg}^{2}.$ The counts in each sector are computed in the\nsame way as done for the entire mosaic. By comparing the induced scatter with\nthat of counts obtained from depth observations scaled to 150MHz, we find that\nthe $1\\sigma$ scatter due to cosmic variance is larger than the Poissonian\nerrors of the source counts, and it may explain the dispersion from previously\nreported depth source counts at flux densities $S<1\\,\\textrm{mJy}$. This work\ndemonstrates the feasibility of achieving deep radio imaging at low-frequencies\nwith LOFAR.\n", - "method": "The astrometry and flux scale accuracy of our source catalog is investigated.", - "id": "1807.04878" - }, - { - "abstract": " We introduce a new entity typing task: given a sentence with an entity\nmention, the goal is to predict a set of free-form phrases (e.g. skyscraper,\nsongwriter, or criminal) that describe appropriate types for the target entity.\nThis formulation allows us to use a new type of distant supervision at large\nscale: head words, which indicate the type of the noun phrases they appear in.\nWe show that these ultra-fine types can be crowd-sourced, and introduce new\nevaluation sets that are much more diverse and fine-grained than existing\nbenchmarks. We present a model that can predict open types, and is trained\nusing a multitask objective that pools our new head-word supervision with prior\nsupervision from entity linking. Experimental results demonstrate that our\nmodel is effective in predicting entity types at varying granularity; it\nachieves state of the art performance on an existing fine-grained entity typing\nbenchmark, and sets baselines for our newly-introduced datasets. Our data and\nmodel can be downloaded from: http://nlp.cs.washington.edu/entity_type\n", - "method": "This formulation allows us to use a new type of distant supervision at large\nscale: head words, which indicate the type of the noun phrases they appear in.", - "id": "1807.04905" - }, - { - "abstract": " The Riccati equation method is used to establish some oscillatory criteria\nfor the second order linear functional - differential equations of multiple\nterms with locally integrable coefficients. An interval oscillation criterion\nfor the second order linear functional - differential equations is proved. We\nhave obtained a generalization of an oscillation theorem of L. Berezanski and\nE. Braverman, a generalization of the well known Fite's oscillation criterion\nand a new global solvability criterion for the second order linear functional -\ndifferential equations with advanced and retarded arguments.\n", - "method": "The Riccati equation method is used to establish some oscillatory criteria\nfor the second order linear functional - differential equations of multiple\nterms with locally integrable coefficients.", - "id": "1807.04943" - }, - { - "abstract": " Contrary to the claims made in Ref. [1], the recent Bell tests by Giustina et\nal. and Shalm et al. do not suffer from a \"photon identification loophole\". The\nmodel discussed in Ref. [1] is exploiting the well-known detection loophole,\nthat is precisely what was closed in the recent experiments.\n", - "method": "and Shalm et al.", - "id": "1807.04999" - }, - { - "abstract": " We discuss a Curie-Weiss model with two groups in the critical regime. This\nis the region where the central limit theorem does not hold any more but the\nmean magnetization still goes to zero as the number of spins grows. We show\nthat the total magnetization normalized by $N^{3/4}$ converges to a non-trivial\ndistribution which is not Gaussian, just as in the single-group Curie-Weiss\nmodel.\n", - "method": "We discuss a Curie-Weiss model with two groups in the critical regime.", - "id": "1807.05020" - }, - { - "abstract": " We characterize the monomial ideals $I\\subset K[x_1,\\ldots,x_n]$ with the\nproperty that the polarization $I^p$ and $I^{\\sigma^n}:=$ the ideal obtained\nfrom $I$ by the $n$-th iterated squarefree operator $\\sigma$ are isomorphic via\na permutation of variables. We give several methods to construct such ideals.\nWe also compare the depth and sdepth of $I$ and $I^{\\sigma^n}$.\n", - "method": "We give several methods to construct such ideals.", - "id": "1807.05028" - }, - { - "abstract": " An attitude of satellite is not always static, sometimes it moves randomly\nand the antenna pointing of satellite is harder to achieve line of sight\ncommunication to other satellite when it is outage by tumbling effect. In order\nto determine an appropriate direction of satellite antenna in inter-satellite\nlink, this paper analyze estimation performance of the direction of arrival\n(DoA) using MUSIC algorithm from connected satellite signal source. It differs\nfrom optical measurement, magnetic field measurement, inertial measurement, and\nglobal positioning system (GPS) attitude determination. The proposed method is\ncharacterized by taking signal source from connected satellites, after that the\nmain satellite processed the information to obtain connected satellites antenna\ndirection. The simulation runs only on the direction of azimuth. The simulation\nresult shows that MUSIC algorithm processing time is faster than satellite\nmovement time in orbit on altitude of 830 km with the period of 101 minutes.\nWith the use of a 50 elements array antenna in spacing of 0.5 wavelength, the\ntotal of 20 angle of arrival (AoA) can be detected in 0.98 seconds of\nprocessing time when using MUSIC algorithm.\n", - "method": "The simulation runs only on the direction of azimuth.", - "id": "1807.05108" - }, - { - "abstract": " Electricity cost is a dominant and rapidly growing expense in data centers.\nUnfortunately, much of the consumed energy is wasted because servers are idle\nfor extended periods of time. We study a capacity management problem that\ndynamically right-sizes a data center, matching the number of active servers\nwith the varying demand for computing capacity. We resort to a data-center\noptimization problem introduced by Lin, Wierman, Andrew and Thereska that, over\na time horizon, minimizes a combined objective function consisting of operating\ncost, modeled by a sequence of convex functions, and server switching cost. All\nprior work addresses a continuous setting in which the number of active\nservers, at any time, may take a fractional value.\n In this paper, we investigate for the first time the discrete data-center\noptimization problem where the number of active servers, at any time, must be\ninteger valued. Thereby we seek truly feasible solutions. First, we show that\nthe offline problem can be solved in polynomial time. Second, we study the\nonline problem and extend the algorithm {\\em Lazy Capacity Provisioning\\/}\n(LCP) by Lin et al. to the discrete setting. We prove that LCP is\n3-competitive. Moreover, we show that no deterministic online algorithm can\nachieve a competitive ratio smaller than~3. In addition, we develop a\nrandomized online algorithm that is 2-competitive against an oblivious\nadversary. Moreover, we prove that 2 is a lower bound for the competitive ratio\nof randomized online algorithms, so our algorithm is optimal.\n Finally, we address the continuous setting and give a lower bound of~2 on the\nbest competitiveness of online algorithms. This matches an upper bound by\nBansal et al. We prove that all lower bounds still hold in a problem variant\nwith more restricted operating cost functions, introduced by Lin et al.\n", - "method": "to the discrete setting.", - "id": "1807.05112" - }, - { - "abstract": " Finite difference schemes are the method of choice for solving nonlinear,\ndegenerate elliptic PDEs, because the Barles-Sougandis convergence framework\n[Barles and Sougandidis, Asymptotic Analysis, 4(3):271-283, 1991] provides\nsufficient conditions for convergence to the unique viscosity solution\n[Crandall, Ishii and Lions, Bull. Amer. Math Soc., 27(1):1-67, 1992]. For\nanisotropic operators, such as the Monge-Ampere equation, wide stencil schemes\nare needed [Oberman, SIAM J. Numer. Anal., 44(2):879-895]. The accuracy of\nthese schemes depends on both the distances to neighbors, $R$, and the angular\nresolution, $d\\theta$. On uniform grids, the accuracy is $\\mathcal O(R^2 +\nd\\theta)$. On point clouds, the most accurate schemes are of $\\mathcal O(R +\nd\\theta)$, by Froese [Numerische Mathematik, 138(1):75-99, 2018]. In this work,\nwe construct geometrically motivated schemes of higher accuracy in both cases:\norder $\\mathcal O(R + d\\theta^2)$ on point clouds, and $\\mathcal O(R^2 +\nd\\theta^2)$ on uniform grids.\n", - "method": "Amer.. Math Soc., 27(1):1-67, 1992].", - "id": "1807.05150" - }, - { - "abstract": " The Gauss law plays a basic role in gauge theories, enforcing gauge\ninvariance and creating edge states and superselection sectors. This article\nsurveys these aspects of the Gauss law in QED, QCD and nonlinear $G/H$ models.\nIt is argued that nonabelian superselection rules are spontaneously broken.\nThat is the case with $SU(3)$ of colour which is spontaneously broken to\n$U(1)\\times U(1)$. Nonlinear $G/H$ models are reformulated as gauge theories\nand the existence of edge states and superselection sectors in these models is\nalso established.\n", - "method": "It is argued that nonabelian superselection rules are spontaneously broken.", - "id": "1807.05161" - }, - { - "abstract": " We study locally compact groups having all dense subgroups (locally) minimal.\nWe call such groups densely (locally) minimal. In 1972 Prodanov proved that the\ninfinite compact abelian groups having all subgroups minimal are precisely the\ngroups $\\mathbb Z_p$ of $p$-adic integers. In [31], we extended Prodanov's\ntheorem to the non-abelian case at several levels. In this paper, we focus on\nthe densely (locally) minimal abelian groups.\n We prove that in case that a topological abelian group $G$ is either compact\nor connected locally compact, then $G$ is densely locally minimal if and only\nif $G$ either is a Lie group or has an open subgroup isomorphic to $\\mathbb\nZ_p$ for some prime $p$. This should be compared with the main result of [9].\nOur Theorem C provides another extension of Prodanov's theorem: an infinite\nlocally compact group is densely minimal if and only if it is isomorphic to\n$\\mathbb Z_p$. In contrast, we show that there exists a densely minimal,\ncompact, two-step nilpotent group that neither is a Lie group nor it has an\nopen subgroup isomorphic to $\\mathbb Z_p$.\n", - "method": "We study locally compact groups having all dense subgroups (locally) minimal.", - "id": "1807.05166" - }, - { - "abstract": " This thesis is devoted to develop tunable devices on the base of\none-dimensional photonic crystals (1DPhC) which can sustain Bloch surface waves\n(BSWs). First, we explore the possibilities to control the BSW propagation\ndirection with polarization of incident light. In this case we manufacture\nadditional passive structures such as gratings on the top of the 1DPhC, which\nare working both as a BSW launcher and polarization-controlled wave-splitters.\nWe test this type of launcher in air and in water as an external medium. Then,\nwe demonstrate the tunability of the BSW by adding an active layers into the\nmultilayer stack. Here a crystalline X-cut thin film lithium niobate (TFLN) is\nused to introduce anisotropic properties to the whole 1DPhC. Different ways to\nmanufacture 1D PhCs with LiNbO3 on the top would be described. Finally, we\nexplore the concept of the electro-optically tuned BSW.\n", - "method": "We test this type of launcher in air and in water as an external medium.", - "id": "1807.05268" - }, - { - "abstract": " Neural network models and deep models are one of the leading and state of the\nart models in machine learning. Most successful deep neural models are the ones\nwith many layers which highly increases their number of parameters. Training\nsuch models requires a large number of training samples which is not always\navailable. One of the fundamental issues in neural networks is overfitting\nwhich is the issue tackled in this thesis. Such problem often occurs when the\ntraining of large models is performed using few training samples. Many\napproaches have been proposed to prevent the network from overfitting and\nimprove its generalization performance such as data augmentation, early\nstopping, parameters sharing, unsupervised learning, dropout, batch\nnormalization, etc.\n In this thesis, we tackle the neural network overfitting issue from a\nrepresentation learning perspective by considering the situation where few\ntraining samples are available which is the case of many real world\napplications. We propose three contributions. The first one presented in\nchapter 2 is dedicated to dealing with structured output problems to perform\nmultivariate regression when the output variable y contains structural\ndependencies between its components. The second contribution described in\nchapter 3 deals with the classification task where we propose to exploit prior\nknowledge about the internal representation of the hidden layers in neural\nnetworks. Our last contribution presented in chapter 4 showed the interest of\ntransfer learning in applications where only few samples are available. In this\ncontribution, we provide an automatic system based on such learning scheme with\nan application to medical domain. In this application, the task consists in\nlocalizing the third lumbar vertebra in a 3D CT scan. This work has been done\nin collaboration with the clinic Rouen Henri Becquerel Center who provided us\nwith data.\n", - "method": "We propose three contributions.", - "id": "1807.05292" - }, - { - "abstract": " We investigate the dependence of steady-state properties of Schelling's\nsegregation model on the agents' activation order. Our basic formalism is the\nPollicott-Weiss version of Schelling's segregation model. Our main result\nmodifies this baseline scenario by incorporating contagion in the decision to\nmove: (pairs of) agents are connected by a second, agent influence network.\nPair activation is specified by a random walk on this network.\n The considered schedulers choose the next pair nonadaptively. We can\ncomplement this result by an example of adaptive scheduler (even one that is\nquite fair) that is able to preclude maximal segregation. Thus scheduler\nnonadaptiveness seems to be required for the validity of the original result\nunder arbitrary asynchronous scheduling. The analysis (and our result) are part\nof an adversarial scheduling approach we are advocating to evolutionary games\nand social simulations.\n", - "method": "Pair activation is specified by a random walk on this network.", - "id": "1807.05419" - }, - { - "abstract": " We present here an extension of our search for EISERS (Extremely Inverted\nSpectrum Extragalactic Radio Sources) to the northern hemisphere. With an\ninverted radio spectrum of slope $\\alpha$ $>$ + 2.5, these rare sources would\neither require a non-standard particle acceleration mechanism (in the framework\nof synchrotron self-absorption hypothesis), or a severe free-free absorption\nwhich attenuates practically all of their synchrotron radiation at metre\nwavelengths. A list of 15 EISERS candidates is presented here. It was assembled\nby applying a sequence of selection filters, starting with the two available\nlarge-sky radio surveys, namely the WENSS (325 MHz) and the ADR-TGSS (150 MHz).\nThese surveys offer the twin advantages of being fairly deep (typical rms $<$\n10 mJy/beam) and having a sub-arcminute resolution. Their zone of overlap\nspreads over 1.3$\\pi$ steradian in the northern hemisphere. Radio spectra are\npresented for the entire sample of 15 EISERS candidates, of which 8 spectra are\nof GPS type. Eleven members of the sample are associated with previously known\nquasars. Information on the parsec-scale radio structure, available for several\nof these sources, is also summarized.\n", - "method": "A list of 15 EISERS candidates is presented here.", - "id": "1807.05441" - }, - { - "abstract": " The problem of identifiability of finite mixtures of finite product measures\nis studied. A mixture model with $K$ mixture components and $L$ observed\nvariables is considered, where each variable takes its value in a finite set\nwith cardinality $M$.The variables are independent in each mixture component.\nThe identifiability of a mixture model means the possibility of attaining the\nmixture components parameters by observing its mixture distribution. In this\npaper, we investigate fundamental relations between the identifiability of\nmixture models and the separability of their observed variables by introducing\ntwo types of separability: strongly and weakly separable variables. Roughly\nspeaking, a variable is said to be separable, if and only if it has some\ndifferences among its probability distributions in different mixture\ncomponents. We prove that mixture models are identifiable if the number of\nstrongly separable variables is greater than or equal to $2K-1$, independent\nform $M$. This fundamental threshold is shown to be tight, where a family of\nnon-identifiable mixture models with less than $2K-1$ strongly separable\nvariables is provided. We also show that mixture models are identifiable if\nthey have at least $2K$ weakly separable variables. To prove these theorems, we\nintroduce a particular polynomial, called characteristic polynomial, which\ntranslates the identifiability conditions to identity of polynomials and allows\nus to construct an inductive proof.\n", - "method": "A mixture model with $K$ mixture components and $L$ observed\nvariables is considered, where each variable takes its value in a finite set\nwith cardinality $M$.The variables are independent in each mixture component.", - "id": "1807.05444" - }, - { - "abstract": " In this paper C-Slow Retiming (CSR) on RTL is discussed. CSR multiplies the\nfunctionality of cores by adding the same number of registers into each path.\nThe technique is ideal for FPGAs with their already existing registers.\nPreviously publications are limited to adding registers on netlist level, which\ngenerates a lot of system verification problems and which is assumed to be the\nmajor drawback to use this technology in the modern multicore times. The paper\nshows how CSR can efficiently be done with timing driven automatic RTL\nmodification. The methodology provided with this paper can be used as guidance\nfor using CSR in high level synthesis (HLS). The paper shows the results of a\nCSR-ed complex RISC core on RTL implemented on FPGAs.\n", - "method": "The technique is ideal for FPGAs with their already existing registers.", - "id": "1807.05446" - }, - { - "abstract": " The Submodular Welfare Maximization problem (SWM) captures an important\nsubclass of combinatorial auctions and has been studied extensively from both\ncomputational and economic perspectives. In particular, it has been studied in\na natural online setting in which items arrive one-by-one and should be\nallocated irrevocably upon arrival. In this setting, it is well known that the\ngreedy algorithm achieves a competitive ratio of 1/2, and recently Kapralov et\nal. (2013) showed that this ratio is optimal for the problem. Surprisingly,\ndespite this impossibility result, Korula et al. (2015) were able to show that\nthe same algorithm is 0.5052-competitive when the items arrive in a uniformly\nrandom order, but unfortunately, their proof is very long and involved. In this\nwork, we present an (arguably) much simpler analysis that provides a slightly\nbetter guarantee of 0.5096-competitiveness for the greedy algorithm in the\nrandom-arrival model. Moreover, this analysis applies also to a generalization\nof online SWM in which the sets defining a (simple) partition matroid arrive\nonline in a uniformly random order, and we would like to maximize a monotone\nsubmodular function subject to this matroid. Furthermore, for this more general\nproblem, we prove an upper bound of 0.576 on the competitive ratio of the\ngreedy algorithm, ruling out the possibility that the competitiveness of this\nnatural algorithm matches the optimal offline approximation ratio of 1-1/e.\n", - "method": "(2013) showed that this ratio is optimal for the problem.. Surprisingly,\ndespite this impossibility result, Korula et al.", - "id": "1807.05529" - }, - { - "abstract": " Consensus and Broadcast are two fundamental problems in distributed\ncomputing, whose solutions have several applications. Intuitively, Consensus\nshould be no harder than Broadcast, and this can be rigorously established in\nseveral models. Can Consensus be easier than Broadcast?\n In models that allow noiseless communication, we prove a reduction of (a\nsuitable variant of) Broadcast to binary Consensus, that preserves the\ncommunication model and all complexity parameters such as randomness, number of\nrounds, communication per round, etc., while there is a loss in the success\nprobability of the protocol. Using this reduction, we get, among other\napplications, the first logarithmic lower bound on the number of rounds needed\nto achieve Consensus in the uniform GOSSIP model on the complete graph. The\nlower bound is tight and, in this model, Consensus and Broadcast are\nequivalent.\n We then turn to distributed models with noisy communication channels that\nhave been studied in the context of some bio-inspired systems. In such models,\nonly one noisy bit is exchanged when a communication channel is established\nbetween two nodes, and so one cannot easily simulate a noiseless protocol by\nusing error-correcting codes. An $\\Omega(\\epsilon^{-2} n)$ lower bound on the\nnumber of rounds needed for Broadcast is proved by Boczkowski et al. [PLOS\nComp. Bio. 2018] in one such model (noisy uniform PULL, where $\\epsilon$ is a\nparameter that measures the amount of noise). In such model, we prove a new\n$\\Theta(\\epsilon^{-2} n \\log n)$ bound for Broadcast and a\n$\\Theta(\\epsilon^{-2} \\log n)$ bound for binary Consensus, thus establishing an\nexponential gap between the number of rounds necessary for Consensus versus\nBroadcast.\n", - "method": "Bio.", - "id": "1807.05626" - }, - { - "abstract": " In this paper, we consider the following weakly coupled nonlinear\nSchr\\\"{o}dinger system \\begin{equation*}\n \\left\\{\n \\begin{array}{ll}\n -\\epsilon^{2}\\Delta u_1 + V_1(x)u_1 = |u_1|^{2p - 2}u_1 + \\beta|u_1|^{p -\n2}|u_2|^pu_1, & x\\in \\mathbb{R}^N,\\\\\n -\\epsilon^{2}\\Delta u_2 + V_2(x)u_2 = |u_2|^{2p - 2}u_2 + \\beta|u_2|^{p -\n2}|u_1|^pu_2, & x\\in \\mathbb{R}^N,\n \\end{array}\n \\right. \\end{equation*} where $\\epsilon>0$, $\\beta\\in\\mathbb{R}$ is a\ncoupling constant, $2p\\in (2,2^*)$ with $2^* = \\frac{2N}{N - 2}$ if $N\\geq 3$\nand $+\\infty$ if $N = 1,2$, $V_1$ and $V_2$ belong to\n$C(\\mathbb{R}^N,[0,\\infty))$. This type of systems arise from models of\nnonlinear optics.\n Supposing that $V_1$ and $V_2$ have local minimum in an open set $\\Lambda$,\nwe construct a new penalized function to show the problem has a family of\nnonstandard solutions $\\{w_{\\epsilon} =\n(u^1_{\\epsilon},u^2_{\\epsilon}):0<\\epsilon<\\epsilon_{0}\\}$ concentrating\nsynchronously provided $\\beta>0$ is suitably large and small. All decay rates\nof $V$ are admissible in this paper and our work is the first semiclassical\nstudy for the problem when $\\beta>0$ is small. Moreover, we know the location\nof concentration points in the case $\\beta>0$ small and $p\\ge 2$ by local\nPohozaev identities, see their applications in [23, 28] for example.\n", - "method": "In this paper, we consider the following weakly coupled nonlinear\nSchr\\\"{o}dinger system \\begin{equation*}\n \\left\\{\n \\begin{array}{ll}\n -\\epsilon^{2}\\Delta u_1 + V_1(x)u_1 = |u_1|^{2p - 2}u_1 + \\beta|u_1|^{p -\n2}|u_2|^pu_1, & x\\in \\mathbb{R}^N,\\\\\n -\\epsilon^{2}\\Delta u_2 + V_2(x)u_2 = |u_2|^{2p - 2}u_2 + \\beta|u_2|^{p -\n2}|u_1|^pu_2, & x\\in \\mathbb{R}^N,\n \\end{array}\n \\right.", - "id": "1807.05644" - }, - { - "abstract": " We perform a statistical standard siren analysis of GW170817. Our analysis\ndoes not utilize knowledge of NGC 4993 as the unique host galaxy of the optical\ncounterpart to GW170817. Instead, we consider each galaxy within the GW170817\nlocalization region as a potential host; combining the redshift from each\ngalaxy with the distance estimate from GW170817 provides an estimate of the\nHubble constant, $H_0$. We then combine the $H_0$ values from all the galaxies\nto provide a final measurement of $H_0$. We explore the dependence of our\nresults on the thresholds by which galaxies are included in our sample, as well\nas the impact of weighting the galaxies by stellar mass and star-formation\nrate. Considering all galaxies brighter than $0.01 L^\\star_B$ as equally likely\nto host a BNS merger, we find $H_0= 76^{+48}_{-23}$ km s$^{-1}$ Mpc$^{-1}$\n(maximum a posteriori and 68.3% highest density posterior interval; assuming a\nflat $H_0$ prior in the range $\\left[ 10, 220 \\right]$ km s$^{-1}$ Mpc$^{-1}$).\nRestricting only to galaxies brighter than $0.626 L^\\star_B$ tightens the\nmeasurement to $H_0= 77^{+37}_{-18}$ km s$^{-1}$ Mpc$^{-1}$. We show that\nweighting the host galaxies by stellar mass or star-formation rate provides\nentirely consistent results with potentially tighter constraints. While these\nstatistical estimates are inferior to the value from the counterpart standard\nsiren measurement utilizing NGC 4993 as the unique host, $H_0=76^{+19}_{-13}$\nkm s$^{-1}$ Mpc$^{-1}$ (determined from the same publicly available data), our\nanalysis is a proof-of-principle demonstration of the statistical approach\nfirst proposed by Bernard Schutz over 30 years ago.\n", - "method": "We perform a statistical standard siren analysis of GW170817.", - "id": "1807.05667" - }, - { - "abstract": " For a given ensemble of input and target states, the classical fidelity\nthreshold (CFT) is the maximum valve of the averaged fidelity, and it can be\nachieved with a measure-and-prepare operation. This quantity can be employed to\nverify whether the channel is in the quantum domain or not. In a recent work by\nChiribella and Xie [Phys. Rev. Lett. {\\bf 110}, 213601 (2013)], it was showed\nthat all the information about the input and target states can be equivalently\ndescribed by an entangled state and an effective entanglement-braking (EB)\nchannel, and the CFTs can be defined with the Choi matrix of the effective EB\nchannel. Following this idea, the protocol proposed by Fuchs and Sasaki\n[Quantum. Inf. Comput, {\\bf 3}, 377 (2003)] are reformulated in terms of the\neffective EB channel, and ss applications, the deterministic and probabilistic\nCFTs for qubit states and the coherent states are derived in this paper.\n", - "method": "Inf.", - "id": "1807.05775" - }, - { - "abstract": " In the scenario with four generation quarks and leptons and using a 3 + 1\nneutrino model having one sterile and the three standard active neutrinos with\na $4 \\times 4$ unitary transformation matrix, $U_{PMNS_{4}}$, we perform a\nmodel-based analysis using the latest global data and determine bounds on the\nsterile neutrino parameters i.e. the neutrino mixing angles. Motivated by our\nprevious results, where, in a quark-lepton complementarity (QLC) model we\npredicted the values of $\\theta_{13}^{PMNS}=(9_{-2}^{+1})^{\\circ}$ and\n$\\theta_{23}^{PMNS}=(40.60_{-0.3}^{+0.1})^{\\circ}$. In the QLC model the\nnon-trivial correlation between $CKM_4$ and $PMNS_4$ mixing matrix is given by\nthe correlation matrix $V_{c_{4}}$. Monte Carlo simulations are performed to\nestimate the texture of $V_{c4}$ followed by the calculation of $PMNS_4$ using\nthe equation, $U_{PMNS_{4}}= (U_{CKM_{4}} . \\psi_{4})^{-1}.V_{c_{4}}$, where\n$\\psi_{4}$ is a diagonal phase matrix. The sterile neutrino mixing angles,\n$\\theta_{14}^{PMNS}$, $\\theta_{24}^{PMNS}$ and $\\theta_{34}^{PMNS}$ are assumed\nto be freely varying between $(0-\\pi/4)$ and obtained results which are\nconsistent with the data available from various experiments, like No$\\nu$A,\nMINOS, SuperK, Ice Cube-DeepCore. In further investigation, we analytically\nobtain approximately similar ranges for various neutrino mixing parameters\n$\\mid{ U_{\\mu 4}}\\mid ^2$ and $\\mid{ U_{\\tau 4}}\\mid ^2$.\n", - "method": "In the scenario with four generation quarks and leptons and using a 3 + 1\nneutrino model having one sterile and the three standard active neutrinos with\na $4 \\times 4$ unitary transformation matrix, $U_{PMNS_{4}}$, we perform a\nmodel-based analysis using the latest global data and determine bounds on the\nsterile neutrino parameters i.e. the neutrino mixing angles.", - "id": "1807.05785" - }, - { - "abstract": " The EcoLexicon English Corpus (EEC) is a 23.1-million-word corpus of\ncontemporary environmental texts. It was compiled by the LexiCon research group\nfor the development of EcoLexicon (Faber, Leon-Arauz & Reimerink 2016; San\nMartin et al. 2017), a terminological knowledge base on the environment. It is\navailable as an open corpus in the well-known corpus query system Sketch Engine\n(Kilgarriff et al. 2014), which means that any user, even without a\nsubscription, can freely access and query the corpus. In this paper, the EEC is\nintroduced by de- scribing how it was built and compiled and how it can be\nqueried and exploited, based both on the functionalities provided by Sketch\nEngine and on the parameters in which the texts in the EEC are classified.\n", - "method": "2017), a terminological knowledge base on the environment.", - "id": "1807.05797" - }, - { - "abstract": " CD16 is a MATLAB-based dosimetry system designed for rapid estimation of\norgan doses for x-ray guided cardiac catheterization procedures. CD16 is\nprincipally designed for epidemiological studies in which dose estimates are\nrequired for thousands of cohort members, often based on limited exposure\ninformation. The minimum data required for dose estimates are (1) total\nexamination kerma area product (PKA), (2) patient size (age or mass), (3) x-ray\nspectrum represented by half value layer (HVL, in mm Al) and (4) examination\ntype. These data can be inputted using an Excel spreadsheet. Organ dose\nestimates are outputted into the same spreadsheet. Doses are estimated using\nlookup tables of k-factors relating PKA to organ doses for a wide range of beam\nangles, six patient sizes and two field sizes. These K-factors were calculated\nusing the Monte Carlo code PCXMC V2.0. CD16 initially calculates organ doses\nfor a beam energy represented by a HVL of 4.8 mm Al, then adjusts to other beam\nenergies if specified. Beam angles can be specified if known, otherwise CD16\nassigns beam angles based on examination type. CD16 is the final version of\nthis dosimetry system to utilise PCXMC. Future versions will use the general\npurpose Monte Carlo code MCNP and hybrid phantoms.\n", - "method": "These data can be inputted using an Excel spreadsheet.", - "id": "1807.05823" - }, - { - "abstract": " We derive a connection between the intrinsic tribological properties and the\nelectronic properties of a solid interface. In particular, we show that the\nadhesion and frictional forces are dictated by the electronic charge\nredistribution occurring due to the relative displacements of the two surfaces\nin contact. We define a figure of merit to quantify such charge redistribution\nand show that simple functional relations hold for a wide series of\ninteractions including metallic, covalent and physical bonds. This suggests\nunconventional ways of measuring friction by recording the evolution of the\ninterfacial electronic charge during sliding. Finally, we explain that the key\nmechanism to reduce adhesive friction is to inhibit the charge flow at the\ninterface and provide examples of this mechanism in common lubricant additives.\n", - "method": "We define a figure of merit to quantify such charge redistribution\nand show that simple functional relations hold for a wide series of\ninteractions including metallic, covalent and physical bonds.", - "id": "1807.05840" - }, - { - "abstract": " First-passage percolation is a random growth model which has a metric\nstructure. An infinite geodesic is an infinite sequence whose all sub-sequences\nare shortest paths. One of the important quantity is the number of infinite\ngeodesics originating from the origin. When $d=2$ and an edge distribution is\ncontinuous, it is proved to be almost surely constant [D. Ahlberg, C. Hoffman.\nRandom coalescing geodesics in first-passage percolation]. In this paper, we\nwill prove the same result for higher dimensions and general distributions.\n", - "method": "Ahlberg, C.. Hoffman.. Random coalescing geodesics in first-passage percolation].", - "id": "1807.05900" - }, - { - "abstract": " Chemical design of block copolymers makes it possible to create polymer\nvesicles with tunable microscopic structure. Here we focus on a model of a\nvesicle made of smectic liquid-crystalline block copolymers at zero\ntemperature. The vesicle assumes a faceted tetrahedral shape and the smectic\nlayers arrange in a stack of parallel straight lines with topological defects\nlocalized at the vertices. We counted the number of allowed states at $T=0$.\nFor any fixed shape, we found a two-dimensional countable degeneracy in the\nsmectic pattern depending on the tilt angle between the smectic layers and the\nedge of the tetrahedral shell. For most values of the tilt angle, the smectic\nlayers contain spiral topological defects. The system can spontaneously break\nchiral symmetry when the layers organize into spiral patterns, composed of a\nbound pair of $+1/2$ disclinations. Finally, we suggest possible applications\nof tetrahedral smectic vesicles in the context of functionalizing defects and\nthe possible consequences of the spiral structures for the rigidity of the\nvesicle.\n", - "method": "We counted the number of allowed states at $T=0$.", - "id": "1807.05988" - }, - { - "abstract": " We establish a set of equations for moments of the distribution function. In\nthe relaxation time approximations, these moments obey a coupled set of\nequations that can be truncated order-by-order. Solving the equations of\nmoments, we are able to identify an attractor solution that controls a\ntransition from a free streaming fixed point to a hydrodynamic fixed point. In\nparticular, this attractor solution provides a renormalization of the effective\nvalue of the shear viscosity to entropy density ratio, $\\eta/s$, taking into\naccount off-equilibrium effects.\n", - "method": "We establish a set of equations for moments of the distribution function.", - "id": "1807.06104" - }, - { - "abstract": " This paper presents a novel data-driven approach for predicting the number of\nvegetation-related outages that occur in power distribution systems on a\nmonthly basis. In order to develop an approach that is able to successfully\nfulfill this objective, there are two main challenges that ought to be\naddressed. The first challenge is to define the extent of the target area. An\nunsupervised machine learning approach is proposed to overcome this difficulty.\nThe second challenge is to correctly identify the main causes of\nvegetation-related outages and to thoroughly investigate their nature. In this\npaper, these outages are categorized into two main groups: growth-related and\nweather-related outages, and two types of models, namely time series and\nnon-linear machine learning regression models are proposed to conduct the\nprediction tasks, respectively. Moreover, various features that can explain the\nvariability in vegetation-related outages are engineered and employed. Actual\noutage data, obtained from a major utility in the U.S., in addition to\ndifferent types of weather and geographical data are utilized to build the\nproposed approach. Finally, by utilizing various time series models and machine\nlearning methods, a comprehensive case study is carried out to demonstrate how\nthe proposed approach can be used to successfully predict the number of\nvegetation-related outages and to help decision-makers to detect vulnerable\nzones in their systems.\n", - "method": "The first challenge is to define the extent of the target area.", - "id": "1807.06180" - }, - { - "abstract": " The flashing Brownian ratchet is a stochastic process that alternates between\ntwo regimes, a one-dimensional Brownian motion and a Brownian ratchet, the\nlatter being a one-dimensional diffusion process that drifts towards a minimum\nof a periodic asymmetric sawtooth potential. The result is directed motion. In\nthe presence of a static homogeneous force that acts in the direction opposite\nthat of the directed motion, there is a reduction (or even a reversal) of the\ndirected motion effect. Such a process may be called a tilted flashing Brownian\nratchet. We show how one can study this process numerically, using a random\nwalk approximation or, equivalently, using numerical solution of the\nFokker-Planck equation. Stochastic simulation is another viable method.\n", - "method": "Stochastic simulation is another viable method.", - "id": "1807.06226" - }, - { - "abstract": " We introduce a method for sparsifying distributed algorithms and exhibit how\nit leads to improvements that go past known barriers in two algorithmic\nsettings of large-scale graph processing: Massively Parallel Computation (MPC),\nand Local Computation Algorithms (LCA).\n - MPC with Strongly Sublinear Memory: Recently, there has been growing\ninterest in obtaining MPC algorithms that are faster than their classic $O(\\log\nn)$-round parallel counterparts for problems such as MIS, Maximal Matching,\n2-Approximation of Minimum Vertex Cover, and $(1+\\epsilon)$-Approximation of\nMaximum Matching. Currently, all such MPC algorithms require\n$\\tilde{\\Omega}(n)$ memory per machine. Czumaj et al. [STOC'18] were the first\nto handle $\\tilde{\\Omega}(n)$ memory, running in $O((\\log\\log n)^2)$ rounds. We\nobtain $\\tilde{O}(\\sqrt{\\log \\Delta})$-round MPC algorithms for all these four\nproblems that work even when each machine has memory $n^{\\alpha}$ for any\nconstant $\\alpha\\in (0, 1)$. Here, $\\Delta$ denotes the maximum degree. These\nare the first sublogarithmic-time algorithms for these problems that break the\nlinear memory barrier.\n - LCAs with Query Complexity Below the Parnas-Ron Paradigm: Currently, the\nbest known LCA for MIS has query complexity $\\Delta^{O(\\log \\Delta)} poly(\\log\nn)$, by Ghaffari [SODA'16]. As pointed out by Rubinfeld, obtaining a query\ncomplexity of $poly(\\Delta\\log n)$ remains a central open question. Ghaffari's\nbound almost reaches a $\\Delta^{\\Omega\\left(\\frac{\\log \\Delta}{\\log\\log\n\\Delta}\\right)}$ barrier common to all known MIS LCAs, which simulate\ndistributed algorithms by learning the local topology, \\`{a} la Parnas-Ron\n[TCS'07]. This barrier follows from the $\\Omega(\\frac{\\log \\Delta}{\\log\\log\n\\Delta})$ distributed lower bound of Kuhn, et al. [JACM'16]. We break this\nbarrier and obtain an MIS LCA with query complexity $\\Delta^{O(\\log\\log\n\\Delta)} poly(\\log n)$.\n", - "method": "Czumaj et al.. [JACM'16].", - "id": "1807.06251" - }, - { - "abstract": " We prove lower bounds on complexity measures, such as the approximate degree\nof a Boolean function and the approximate rank of a Boolean matrix, using\nquantum arguments. We prove these lower bounds using a quantum query algorithm\nfor the combinatorial group testing problem.\n We show that for any function f, the approximate degree of computing the OR\nof n copies of f is Omega(sqrt{n}) times the approximate degree of f, which is\noptimal. No such general result was known prior to our work, and even the lower\nbound for the OR of ANDs function was only resolved in 2013.\n We then prove an analogous result in communication complexity, showing that\nthe logarithm of the approximate rank (or more precisely, the approximate\ngamma_2 norm) of F: X x Y -> {0,1} grows by a factor of Omega~(sqrt{n}) when we\ntake the OR of n copies of F, which is also essentially optimal. As a\ncorollary, we give a new proof of Razborov's celebrated Omega(sqrt{n}) lower\nbound on the quantum communication complexity of the disjointness problem.\n Finally, we generalize both these results from composition with the OR\nfunction to composition with arbitrary symmetric functions, yielding nearly\noptimal lower bounds in this setting as well.\n", - "method": "No such general result was known prior to our work, and even the lower\nbound for the OR of ANDs function was only resolved in 2013.", - "id": "1807.06256" - }, - { - "abstract": " We show that the complex dynamics of the general Bianchi IX universe in the\nvicinity of the spacelike singularity can be approximated by a simplified\nsystem of equations. Our analysis is mainly based on numerical simulations. The\nproperties of the solution space can be studied by using this simplified\ndynamics. Our results will be useful for the quantization of the general\nBianchi IX model.\n", - "method": "Our analysis is mainly based on numerical simulations.", - "id": "1807.06261" - }, - { - "abstract": " We simulate an actin filament as an automaton network. Every atom takes two\nor three states and updates its state, in discrete time, depending on a ratio\nof its neighbours in some selected state. All atoms/automata simultaneously\nupdate their states by the same rule. Two state transition rules are\nconsidered. In semi-totalistic Game of Life like actin filament automaton atoms\ntake binary states `0' and `1' and update their states depending on a ratio of\nneighbours in the state `1'. In excitable actin filament automaton atoms take\nthree states: resting, excited and refractory. A resting atom excites if a\nratio of its excited neighbours belong to some specified interval; transitions\nfrom excited state to refractory state and from refractory state to resting\nstate are unconditional. In computational experiments we implement mappings of\n8-bit input string to 8-bit output string via dynamics of\nperturbation/excitation on actin filament automata. We assign eight domains in\nan actin filament as I/O ports. To write {\\sc True} to a port we perturb/excite\na certain percentage of the nodes in the domain corresponding to the port. We\nread outputs at the ports after some time interval. A port is considered to be\nin a state {\\sc True} if a number of excited nodes in the port's domain exceed\na certain threshold. A range of eight-argument Boolean functions is uncovered\nin a series of computational trials when all possible configurations of\neight-elements binary strings were mapped onto excitation outputs of the I/O\ndomains.\n", - "method": "We simulate an actin filament as an automaton network.", - "id": "1807.06352" - }, - { - "abstract": " Stability analysis of paramagnetic prolate and oblate spheroidal particle in\nprecessing magnetic field is studied. Bifurcation diagram is calculated\nanalytically in dependence of magnetic field frequency and precession angle.\nThe direction of particle in synchronous regime is calculated. Rotational\ndynamics and mean rotational frequency in asynchronous regime is found.\nObtained theoretical model gives possibility to calculate analytically dynamics\nof the particle in limiting case (when motion is periodic). Theoretically\nobtained models were compared with experimental results of rod like particle\ndynamics in precessing magnetic field. Experimental results are in good\nagreement with proposed theory.\n", - "method": "The direction of particle in synchronous regime is calculated.", - "id": "1807.06379" - }, - { - "abstract": " Many theories of deep learning have shown that a deep network can require\ndramatically fewer resources to represent a given function compared to a\nshallow network. But a question remains: can these efficient representations be\nlearned using current deep learning techniques? In this work, we test whether\nstandard deep learning methods can in fact find the efficient representations\nposited by several theories of deep representation. Specifically, we train deep\nneural networks to learn two simple functions with known efficient solutions:\nthe parity function and the fast Fourier transform. We find that using\ngradient-based optimization, a deep network does not learn the parity function,\nunless initialized very close to a hand-coded exact solution. We also find that\na deep linear neural network does not learn the fast Fourier transform, even in\nthe best-case scenario of infinite training data, unless the weights are\ninitialized very close to the exact hand-coded solution. Our results suggest\nthat not every element of the class of compositional functions can be learned\nefficiently by a deep network, and further restrictions are necessary to\nunderstand what functions are both efficiently representable and learnable.\n", - "method": "Specifically, we train deep\nneural networks to learn two simple functions with known efficient solutions:\nthe parity function and the fast Fourier transform.", - "id": "1807.06399" - }, - { - "abstract": " The electronic and optical properties of 2D hexagonal boron nitride are\nstudied using first principle calculations. GW and BSE methods are employed in\norder to predict with better accuracy the excited and excitonic properties of\nthis material. We determine the values of the band gap, optical gap, excitonic\nbinding energies and analyse the excitonic wave functions. We also calculate\nthe exciton energies following an equation of motion formalism and the Elliot\nformula, and find a very good agreement with the GW+BSE method. The optical\nproperties are studied for both the TM and TE modes, showing that 2D hBN is a\ngood candidate to polaritonics in the UV range. In particular it is shown that\na single layer of h-BN can act as an almost perfect mirror for ultraviolet\nelectromagnetic radiation.\n", - "method": "We determine the values of the band gap, optical gap, excitonic\nbinding energies and analyse the excitonic wave functions.", - "id": "1807.06413" - }, - { - "abstract": " We introduce and study filtrations of a matroid on a linearly ordered ground\nset, which are particular sequences of nested sets. A given basis can be\ndecomposed into a uniquely defined sequence of bases of minors, such that these\nbases have an internal/external activity equal to 1/0 or 0/1 (in the sense of\nTutte polynomial activities). This decomposition, which we call the active\nfiltration/partition of the basis, refines the known partition of the ground\nset into internal and external elements with respect to a given basis. It can\nbe built by a certain closure operator, which we call the active closure. It\nrelies only on the fundamental bipartite graph of the basis and can be\nexpressed also as a decomposition of general bipartite graphs on a linearly\nordered set of vertices.\n From this, first, structurally, we obtain that the set of all bases can be\ncanonically partitioned and decomposed in terms of such bases of minors induced\nby filtrations. Second, enumeratively, we derive an expression of the Tutte\npolynomial of a matroid in terms of beta invariants of minors. This expression\nrefines at the same time the classical expressions in terms of basis activities\nand orientation activities (if the matroid is oriented), and the well-known\nconvolution formula for the Tutte polynomial. Third, in a companion paper of\nthe same series (No. 2.b), we use this decomposition of matroid bases, along\nwith a similar decomposition of oriented matroids, and along with a bijection\nin the 1/0 activity case from a previous paper (No. 1), to define the canonical\nactive bijection between orientations/signatures/reorientations and spanning\ntrees/simplices/bases of a graph/real hyperplane arrangement/oriented matroid,\nas well as various related bijections.\n", - "method": "A given basis can be\ndecomposed into a uniquely defined sequence of bases of minors, such that these\nbases have an internal/external activity equal to 1/0 or 0/1 (in the sense of\nTutte polynomial activities).", - "id": "1807.06516" - }, - { - "abstract": " The proliferation of information disseminated by public/social media has made\ndecision-making highly challenging due to the wide availability of noisy,\nuncertain, or unverified information. Although the issue of uncertainty in\ninformation has been studied for several decades, little work has investigated\nhow noisy (or uncertain) or valuable (or credible) information can be\nformulated into people's opinions, modeling uncertainty both in the quantity\nand quality of evidence leading to a specific opinion. In this work, we model\nand analyze an opinion and information model by using Subjective Logic where\nthe initial set of evidence is mixed with different types of evidence (i.e.,\npro vs. con or noisy vs. valuable) which is incorporated into the opinions of\noriginal propagators, who propagate information over a network. With the help\nof an extensive simulation study, we examine how the different ratios of\ninformation types or agents' prior belief or topic competence affect the\noverall information diffusion. Based on our findings, agents' high uncertainty\nis not necessarily always bad in making a right decision as long as they are\ncompetent enough not to be at least biased towards false information (e.g.,\nneutral between two extremes).\n", - "method": "con or noisy vs.", - "id": "1807.06519" - }, - { - "abstract": " It has been previously shown by the authors that a directed graph on a\nlinearly ordered set of edges (ordered graph) with adjacent unique source and\nsink (bipolar digraph) has a unique fully optimal spanning tree, that satisfies\na simple criterion on fundamental cycle/cocycle directions. This result yields,\nfor any ordered graph, a canonical bijection between bipolar orientations and\nspanning trees with internal activity 1 and external activity 0 in the sense of\nthe Tutte polynomial. This bijection can be extended to all orientations and\nall spanning trees, yielding the active bijection, presented for graphs in a\ncompanion paper. In this paper, we specifically address the problem of the\ncomputation of the fully optimal spanning tree of an ordered bipolar digraph.\nIn contrast with the inverse mapping, built by a straightforward single pass\nover the edge set, the direct computation is not easy and had previously been\nleft aside. We give two independent constructions. The first one is a\ndeletion/contraction recursion, involving an exponential number of minors. It\nis structurally significant but it is efficient only for building the whole\nbijection (i.e. all images) at once. The second one is more complicated and is\nthe main contribution of the paper. It involves just one minor for each edge of\nthe resulting spanning tree, and it is a translation and an adaptation in the\ncase of graphs, in terms of weighted cocycles, of a general geometrical linear\nprogramming type algorithm, which allows for a polynomial time complexity.\n", - "method": "We give two independent constructions.", - "id": "1807.06552" - }, - { - "abstract": " We introduce a family of loop soup models on the hypercubic lattice. The\nmodels involve links on the edges, and random pairings of the link endpoints on\nthe sites. We conjecture that loop correlations of distant points are given by\nPoisson-Dirichlet correlations in dimensions three and higher. We prove that,\nin a specific random wire model that is related to the classical XY spin\nsystem, the probability that distant sites form an even partition is given by\nthe Poisson-Dirichlet counterpart.\n", - "method": "We introduce a family of loop soup models on the hypercubic lattice.", - "id": "1807.06564" - }, - { - "abstract": " The Open Science Grid(OSG) is a world-wide computing system which facilitates\ndistributed computing for scientific research. It can distribute a\ncomputationally intensive job to geo-distributed clusters and process job's\ntasks in parallel. For compute clusters on the OSG, physical resources may be\nshared between OSG and cluster's local user-submitted jobs, with local jobs\npreempting OSG-based ones. As a result, job preemptions occur frequently in\nOSG, sometimes significantly delaying job completion time.\n We have collected job data from OSG over a period of more than 80 days. We\npresent an analysis of the data, characterizing the preemption patterns and\ndifferent types of jobs. Based on observations, we have grouped OSG jobs into 5\ncategories and analyze the runtime statistics for each category. we further\nchoose different statistical distributions to estimate probability density\nfunction of job runtime for different classes.\n", - "method": "We have collected job data from OSG over a period of more than 80 days.", - "id": "1807.06639" - }, - { - "abstract": " Abridged. Methods. We made use of the Karl G. Jansky Very Large Array (VLA)\nin the B configuration at K band, and in the A configuration at both Ku and C\nbands, in order to image the radio continuum emission towards 25 H$_2$O maser\nsites with an angular resolution and thermal rms of the order of $0.1''$ and 10\n$\\mu$Jy beam$^{-1}$, respectively. These targets add to our pilot study of 11\nmaser sites presented in Moscadelli et al. (2016). The sample of H$_2$O maser\nsites was selected among those regions having an accurate distance measurement,\nobtained through maser trigonometric parallaxes, and H$_2$O maser luminosities\nin excess of 10$^{-6}$ L$_{\\odot}$. Results. We present high-resolution radio\ncontinuum images of 33 sources belonging to 25 star-forming regions. In each\nregion, we detect radio continuum emission within a few 1000 au of the H$_2$O\nmasers' position; 50% of the radio continuum sources are associated with\nbolometric luminosities exceeding 5 $\\times$ 10$^{3}$ L$_{\\odot}$, including\nW33A and G240.32$+$0.07. We provide a detailed spectral index analysis for each\nradio continuum source, based on the integrated fluxes at each frequency, and\nproduce spectral index maps with the multi-frequency-synthesis deconvolution\nalgorithm of CASA. The radio continuum emission traces thermal bremsstrahlung\nin (proto)stellar winds and jets, with flux densities at 22 GHz below 3 mJy,\nand spectral index values between $-0.1$ and 1.3. We prove a strong correlation\n($r>$0.8) between the radio continuum luminosity (L$_{\\rm rad}$) and the H$_2$O\nmaser luminosity (L$_{\\rm H_2O}$) of ($\\rm L_{8GHz}/mJy\\times\nkpc^2)=10^{3.8}\\times(L_{H_2O}/L_{\\odot})^{0.74}$. Since H$_2$O masers are\nexcited through shocks driven by (proto)stellar winds and jets, these results\nprovide support to the idea that the radio continuum emission around young\nstars is dominated by shock-ionization.\n", - "method": "We made use of the Karl G.", - "id": "1807.06680" - }, - { - "abstract": " We propose to take a novel approach to robot system design where each\nbuilding block of a larger system is represented as a differentiable program,\ni.e. a deep neural network. This representation allows for integrating\nalgorithmic planning and deep learning in a principled manner, and thus combine\nthe benefits of model-free and model-based methods. We apply the proposed\napproach to a challenging partially observable robot navigation task. The robot\nmust navigate to a goal in a previously unseen 3-D environment without knowing\nits initial location, and instead relying on a 2-D floor map and visual\nobservations from an onboard camera. We introduce the Navigation Networks\n(NavNets) that encode state estimation, planning and acting in a single,\nend-to-end trainable recurrent neural network. In preliminary simulation\nexperiments we successfully trained navigation networks to solve the\nchallenging partially observable navigation task.\n", - "method": "We propose to take a novel approach to robot system design where each\nbuilding block of a larger system is represented as a differentiable program,\ni.e. a deep neural network.", - "id": "1807.06696" - }, - { - "abstract": " We demonstrate low-power amplification process in cavity optomechanics (COM).\nThis operation is based on the nonlinear position-modulated self-Kerr\ninteraction. Owing to this nonlinear term, the effective coupling highly scales\nwith the photon number, resulting in a giant enhancement of the cooperativity.\nEven for small nonlinearity, the system reaches the amplification threshold for\nweak driving strength, leading to low-power phonon lasing. This amplifier can\nbe phase-preserving and provides a practical advantage related to the power\nconsumption issues. This work opens up new avenues to realize low-power and\nefficient amplifiers in optomechanics and related fields.\n", - "method": "We demonstrate low-power amplification process in cavity optomechanics (COM).", - "id": "1807.06852" - }, - { - "abstract": " Stochastic dominance is a crucial tool for the analysis of choice under risk.\nIt is typically analyzed as a property of two gambles that are taken in\nisolation. We study how additional independent sources of risk (e.g.\nuninsurable labor risk, house price risk, etc.) can affect the ordering of\ngambles. We show that, perhaps surprisingly, background risk can be strong\nenough to render lotteries that are ranked by their expectation ranked in terms\nof first-order stochastic dominance. We extend our results to second order\nstochastic dominance, and show how they lead to a novel, and elementary,\naxiomatization of mean-variance preferences.\n", - "method": "We study how additional independent sources of risk (e.g.\nuninsurable labor risk, house price risk, etc.) can affect the ordering of\ngambles.", - "id": "1807.06927" - }, - { - "abstract": " In this paper, we study the conjecture of Gardner and Zvavitch from\n\\cite{GZ}, which suggests that the standard Gaussian measure $\\gamma$ enjoys\n$\\frac{1}{n}$-concavity with respect to the Minkowski addition of\n\\textbf{symmetric} convex sets. We prove this fact up to a factor of 2: that\nis, we show that for symmetric convex $K$ and $L,$ $$ \\gamma(\\lambda\nK+(1-\\lambda)L)^{\\frac{1}{2n}}\\geq \\lambda\n\\gamma(K)^{\\frac{1}{2n}}+(1-\\lambda)\\gamma(L)^{\\frac{1}{2n}}. $$ Further, we\nshow that under suitable dimension-free uniform bounds on the Hessian of the\npotential, the log-concavity of even measures can be strengthened to\n$p$-concavity, with $p>0,$ with respect to the addition of symmetric convex\nsets.\n", - "method": "We prove this fact up to a factor of 2: that\nis, we show that for symmetric convex $K$ and $L,$ $$ \\gamma(\\lambda\nK+(1-\\lambda)L)^{\\frac{1}{2n}}\\geq \\lambda\n\\gamma(K)^{\\frac{1}{2n}}+(1-\\lambda)\\gamma(L)^{\\frac{1}{2n}}.", - "id": "1807.06952" - }, - { - "abstract": " Time-aware encoding of frame sequences in a video is a fundamental problem in\nvideo understanding. While many attempted to model time in videos, an explicit\nstudy on quantifying video time is missing. To fill this lacuna, we aim to\nevaluate video time explicitly. We describe three properties of video time,\nnamely a) temporal asymmetry, b)temporal continuity and c) temporal causality.\nBased on each we formulate a task able to quantify the associated property.\nThis allows assessing the effectiveness of modern video encoders, like C3D and\nLSTM, in their ability to model time. Our analysis provides insights about\nexisting encoders while also leading us to propose a new video time encoder,\nwhich is better suited for the video time recognition tasks than C3D and LSTM.\nWe believe the proposed meta-analysis can provide a reasonable baseline to\nassess video time encoders on equal grounds on a set of temporal-aware tasks.\n", - "method": "Based on each we formulate a task able to quantify the associated property.", - "id": "1807.06980" - }, - { - "abstract": " Period estimation is an important task in the classification of many variable\nastrophysical objects. Here we present GRAPE: Genetic Routine for Astronomical\nPeriod Estimation, a genetic algorithm optimised for the processing of survey\ndata with spurious and aliased artefacts. It uses a Bayesian Generalised\nLomb-Scargle (BGLS) fitness function designed for use with the Skycam survey\nconducted at the Liverpool Telescope. We construct a set of simulated light\ncurves using both regular and Skycam survey cadence with four types of signal:\nsinusoidal, sawtooth, symmetric eclipsing binary and eccentric eclipsing\nbinary. We apply GRAPE and a BGLS periodogram to this data and show that the\nperformance of GRAPE is superior to the periodogram on sinusoidal and sawtooth\nlight curves with relative hit rate improvement of 18.2% and 6.4% respectively.\nThe symmetric and eccentric eclipsing binary light curves have similar\nperformance on both methods. We show the Skycam cadence is sufficient to\ncorrectly estimate the period for all of the sinusoidal shape light curves\nalthough this degrades with increased non-sinusoidal shape with sawtooth,\nsymmetric binary and eccentric binary light curves down by 20%, 30% and 35%\nrespectively. The runtime of GRAPE demonstrates that light curves with more\nthan 500-1000 data points achieve similar performance in less computing time.\nThe GRAPE performance can be matched by a frequency spectrum with an\noversampled fine-tuning grid at the cost of almost doubling the runtime.\nFinally, we propose improvements which will extend this method to the detection\nof quasi-periodic signals and the use of multiband light curves.\n", - "method": "Period estimation is an important task in the classification of many variable\nastrophysical objects.", - "id": "1807.07010" - }, - { - "abstract": " We study the learnability of sums of independent integer random variables\ngiven a bound on the size of the union of their supports. For $\\mathcal{A}\n\\subset \\mathbf{Z}_{+}$, a sum of independent random variables with collective\nsupport $\\mathcal{A}$} (called an $\\mathcal{A}$-sum in this paper) is a\ndistribution $\\mathbf{S} = \\mathbf{X}_1 + \\cdots + \\mathbf{X}_N$ where the\n$\\mathbf{X}_i$'s are mutually independent (but not necessarily identically\ndistributed) integer random variables with $\\cup_i \\mathsf{supp}(\\mathbf{X}_i)\n\\subseteq \\mathcal{A}.$ We give two main algorithmic results for learning such\ndistributions:\n 1. For the case $| \\mathcal{A} | = 3$, we give an algorithm for learning\n$\\mathcal{A}$-sums to accuracy $\\epsilon$ that uses $\\mathsf{poly}(1/\\epsilon)$\nsamples and runs in time $\\mathsf{poly}(1/\\epsilon)$, independent of $N$ and of\nthe elements of $\\mathcal{A}$.\n 2. For an arbitrary constant $k \\geq 4$, if $\\mathcal{A} = \\{ a_1,...,a_k\\}$\nwith $0 \\leq a_1 < ... < a_k$, we give an algorithm that uses\n$\\mathsf{poly}(1/\\epsilon) \\cdot \\log \\log a_k$ samples (independent of $N$)\nand runs in time $\\mathsf{poly}(1/\\epsilon, \\log a_k).$\n We prove an essentially matching lower bound: if $|\\mathcal{A}| = 4$, then\nany algorithm must use $\\Omega(\\log \\log a_4) $ samples even for learning to\nconstant accuracy. We also give similar-in-spirit (but quantitatively very\ndifferent) algorithmic results, and essentially matching lower bounds, for the\ncase in which $\\mathcal{A}$ is not known to the learner.\n", - "method": "For $\\mathcal{A}\n\\subset \\mathbf{Z}_{+}$, a sum of independent random variables with collective\nsupport $\\mathcal{A}$} (called an $\\mathcal{A}$-sum in this paper) is a\ndistribution $\\mathbf{S} = \\mathbf{X}_1 + \\cdots + \\mathbf{X}_N$ where the\n$\\mathbf{X}_i$'s are mutually independent (but not necessarily identically\ndistributed) integer random variables with $\\cup_i \\mathsf{supp}(\\mathbf{X}_i)\n\\subseteq \\mathcal{A}.$ We give two main algorithmic results for learning such\ndistributions:\n 1.", - "id": "1807.07013" - }, - { - "abstract": " At the core of every frustrated system, one can identify the existence of\nfrustrated rings that are usually interpreted in terms of single--particle\nphysics. We check this point of view through a careful analysis of the\nentanglement entropy of both models that admit an exact single--particle\ndecomposition of their Hilbert space due to integrability and those for which\nthe latter is supposed to hold only as a low energy approximation. In\nparticular, we study generic spin chains made by an odd number of sites with\nshort-range antiferromagnetic interactions and periodic boundary conditions,\nthus characterized by a weak, i.e. nonextensive, frustration. While for\ndistances of the order of the correlation length the phenomenology of these\nchains is similar to that of the non-frustrated cases, we find that correlation\nfunctions involving a number of sites scaling like the system size follow\ndifferent rules. We quantify the long-range correlations through the von\nNeumann entanglement entropy, finding that indeed it violates the area law,\nwhile not diverging with the system size. This behavior is well fitted by a\nuniversal law that we derive from the conjectured single--particle picture.\n", - "method": "In\nparticular, we study generic spin chains made by an odd number of sites with\nshort-range antiferromagnetic interactions and periodic boundary conditions,\nthus characterized by a weak, i.e. nonextensive, frustration.", - "id": "1807.07055" - }, - { - "abstract": " The contributions of scalars and fermions to the null polygonal bosonic\nWilson loops/gluon MHV scattering amplitudes in $\\mathcal{N} = 4$ SYM are\nconsidered. We first examine the re-summation of scalars at strong coupling.\nThen, we disentangle the form of the fermion contribution and show its strong\ncoupling expansion. In particular, we derive the leading order with the\nappearance of a fermion-anti-fermion bound state first and then effective\nmultiple bound states thereof. This reproduces the string minimal area result\nand also applies to the Nekrasov instanton partition function $\\mathcal{Z}$ of\nthe $\\mathcal{N}=2$ theories. Especially, in the latter case the method appears\nto be suitable for a systematic expansion.\n", - "method": "We first examine the re-summation of scalars at strong coupling.", - "id": "1807.07058" - }, - { - "abstract": " We study the distribution (w.r.t. the vacuum state) of family of partial sums\nSm of position operators on weakly monotone Fock space. We show that any single\noperator has the Wigner law, and an arbitrary family of them (with the index\nset linearly ordered) is a collection of monotone independent random variables.\nIt turns out that our problem equivalently consists in finding the m-fold\nmonotone convolution of the semicircle law. For m = 2 we compute the explicit\ndistribution. For any m > 2 we give the moments of the measure, and show it is\nabsolutely continuous and compactly supported on a symmetric interval whose\nendpoints can be found by a recurrence relation.\n", - "method": "We study the distribution (w.r.t. the vacuum state) of family of partial sums\nSm of position operators on weakly monotone Fock space.", - "id": "1807.07101" - }, - { - "abstract": " Embedding models typically associate each word with a single real-valued\nvector, representing its different properties. Evaluation methods, therefore,\nneed to analyze the accuracy and completeness of these properties in\nembeddings. This requires fine-grained analysis of embedding subspaces.\nMulti-label classification is an appropriate way to do so. We propose a new\nevaluation method for word embeddings based on multi-label classification given\na word embedding. The task we use is fine-grained name typing: given a large\ncorpus, find all types that a name can refer to based on the name embedding.\nGiven the scale of entities in knowledge bases, we can build datasets for this\ntask that are complementary to the current embedding evaluation datasets in:\nthey are very large, contain fine-grained classes, and allow the direct\nevaluation of embeddings without confounding factors like sentence context\n", - "method": "This requires fine-grained analysis of embedding subspaces.", - "id": "1807.07186" - }, - { - "abstract": " Given a target image as query, person re-identification systems retrieve a\nranked list of candidate matches on a per-camera basis. In deployed systems, a\nhuman operator scans these lists and labels sighted targets by touch or\nmouse-based selection. However, classical re-id approaches generate per-camera\nlists independently. Therefore, target identifications by operator in a subset\nof cameras cannot be utilized to improve ranking of the target in remaining set\nof network cameras. To address this shortcoming, we propose a novel sequential\nmulti-camera re-id approach. The proposed approach can accommodate human\noperator inputs and provides early gains via a monotonic improvement in target\nranking. At the heart of our approach is a fusion function which operates on\ndeep feature representations of query and candidate matches. We formulate an\noptimization procedure custom-designed to incrementally improve query\nrepresentation. Since existing evaluation methods cannot be directly adopted to\nour setting, we also propose two novel evaluation protocols. The results on two\nlarge-scale re-id datasets (Market-1501, DukeMTMC-reID) demonstrate that our\nmulti-camera method significantly outperforms baselines and other popular\nfeature fusion schemes. Additionally, we conduct a comparative subject-based\nstudy of human operator performance. The superior operator performance enabled\nby our approach makes a compelling case for its integration into deployable\nvideo-surveillance systems.\n", - "method": "The results on two\nlarge-scale re-id datasets (Market-1501, DukeMTMC-reID) demonstrate that our\nmulti-camera method significantly outperforms baselines and other popular\nfeature fusion schemes.", - "id": "1807.07295" - }, - { - "abstract": " This paper contributes to the study of the fibers of the commutator map on\nspecial linear groups in characteristic zero. Specifically, we show that the\nfibers over non-central elements all have the same dimension. Also we explain\nthat the fibers over central elements can be of larger dimension and compute\nhow large. We use the character tables of finite general linear groups\nconstructed by J.A. Green to count solutions to the commutator equation\n$[x,y]=g$ over finite fields and use algebraic geometry to go from\ncharacteristic $p$ to characteristic $0$. To deal with fibers over central\nelements, we compute the orbits of the conjugation action of $\\mathrm{GL}_n$ on\nthese fibers.\n", - "method": "Also we explain\nthat the fibers over central elements can be of larger dimension and compute\nhow large.", - "id": "1807.07300" - }, - { - "abstract": " A real semisimple Lie group G_0 embedded in its complexification G has only\nfinitely many orbits in any G-fag manifold Z = G/Q. The complex geometry of its\nopen orbits D (flag domains) is studied from the point of view of compact\ncomplex submanifolds C (cycles) which arise as orbits of certain distinguished\nsubgroups. Normal bundles E of the cycles are analyzed in some detail. It is\nshown that E is trivial if and only if D is holomorphically convex, in fact a\nproduct of C and a Hermitian symmetric space, and otherwise D is pseudoconcave.\n", - "method": "Normal bundles E of the cycles are analyzed in some detail.", - "id": "1807.07311" - }, - { - "abstract": " Context: High-mass young stellar objects (HMYSOs) can undergo accretion\nepisodes that strongly affect the star evolution, the dynamics of the disk, and\nits chemical evolution. Recently reported extraordinary bursts in the methanol\nmaser emission may be the observational signature of accretion events in deeply\nembedded HMYSOs. Aims: We analyze the light curve of 6.7 GHz methanol masers in\nS255IR-NIRS3 during the 2015-2016 burst. Methods: 8.5-year monitoring data with\nan average sampling interval of 5 days were obtained with the Torun 32 m radio\ntelescope. Archival data were added, extending the time series to ~27 years.\nResults: The maser emission showed moderate (25-30%) variability on timescales\nof months to years over ~23 years since its discovery. The main burst was\npreceded by a one-year increase of the total flux density by a factor of 2.5,\nthen it grew by a factor of 10 over ~0.4 years and declined by a factor of 8\nduring the consecutive 2.4 years. The peak maser luminosity was a factor of\n24.5 higher than the pre-burst quiescent value. The light curves of individual\nfeatures showed considerable diversity but indicated a general trend of\nsuppression of the maser emission at blueshifted (<4.7 km s$^{-1}$) velocities\nwhen the redshifted emission rapidly grew and new emission features appeared at\nvelocities >5.8 km s$^{-1}$. This new emission provided a contribution of about\n80% to the maser luminosity around the peak of the burst. Conclusions: The\nonset of the maser burst exactly coincides with that of the infrared burst\nestimated from the motion of the light echo. This strongly supports the\nradiative pumping scheme of the maser transition. The growth of the maser\nluminosity is the result of an increasing volume of gas where the maser\ninversion is achieved.\n", - "method": "Archival data were added, extending the time series to ~27 years.", - "id": "1807.07334" - }, - { - "abstract": " For a finite dimensional unital complex simple Jordan superalgebra $J$, the\nTits-Kantor-Koecher construction yields a 3-graded Lie superalgebra $\\mathfrak\ng_\\flat\\cong \\mathfrak g_\\flat(-1)\\oplus\\mathfrak g_\\flat(0)\\oplus\\mathfrak\ng_\\flat(1)$, such that $\\mathfrak g_\\flat(-1)\\cong J$. Set $V:=\\mathfrak\ng_\\flat(-1)^*$ and $\\mathfrak g:=\\mathfrak g_\\flat(0)$.\n In most cases, the space $\\mathcal P(V)$ of superpolynomials on $V$ is a\ncompletely reducible and multiplicity-free representation of $\\mathfrak g$,\nwith a decomposition $\\mathcal P(V):=\\bigoplus_{\\lambda\\in\\Omega}V_\\lambda$,\nwhere $\\left(V_\\lambda\\right)_{\\lambda\\in\\Omega}$ is a family of irreducible\n$\\mathfrak g$-modules parametrized by a set of partitions $\\Omega$. In these\ncases, one can define a natural basis\n$\\left(D_\\lambda\\right)_{\\lambda\\in\\Omega}$ of \"Capelli operators\" for the\nalgebra $\\mathcal{PD}(V)^{\\mathfrak g}$. In this paper we complete the solution\nto the Capelli eigenvalue problem, which is to determine the scalar\n$c_\\mu(\\lambda)$ by which $D_\\mu$ acts on $V_\\lambda$.\n We associate a restricted root system $\\mathit{\\Sigma}$ to the symmetric pair\n$(\\mathfrak g,\\mathfrak k)$ that corresponds to $J$, which is either a deformed\nroot system of type $\\mathsf{A}(m,n)$ or a root system of type $\\mathsf{Q}(n)$.\nWe prove a necessary and sufficient condition on the structure of\n$\\mathit{\\Sigma}$ for $\\mathcal{P}(V)$ to be completely reducible and\nmultiplicity-free. When $\\mathit{\\Sigma}$ satisfies the latter condition we\nobtain an explicit formula for the eigenvalue $c_\\mu(\\lambda)$, in terms of\nSergeev-Veselov's shifted super Jack polynomials when $\\mathit{\\Sigma}$ is of\ntype $\\mathsf{A}(m,n)$, and Okounkov-Ivanov's factorial Schur $Q$-polynomials\nwhen $\\mathit{\\Sigma}$ is of type $\\mathsf{Q}(n)$.\n", - "method": "For a finite dimensional unital complex simple Jordan superalgebra $J$, the\nTits-Kantor-Koecher construction yields a 3-graded Lie superalgebra $\\mathfrak\ng_\\flat\\cong \\mathfrak g_\\flat(-1)\\oplus\\mathfrak g_\\flat(0)\\oplus\\mathfrak\ng_\\flat(1)$, such that $\\mathfrak g_\\flat(-1)\\cong J$.", - "id": "1807.07340" - }, - { - "abstract": " Deformable Image Registration (DIR) of MR and CT images is one of the most\nchallenging registration task, due to the inherent structural differences of\nthe modalities and the missing dense ground truth. Recently cycle Generative\nAdversarial Networks (cycle-GANs) have been used to learn the intensity\nrelationship between these 2 modalities for unpaired brain data. Yet its\nusefulness for DIR was not assessed.\n In this study we evaluate the DIR performance for thoracic and abdominal\norgans after synthesis by cycle-GAN. We show that geometric changes, which\ndifferentiate the two populations (e.g. inhale vs. exhale), are readily\nsynthesized as well. This causes substantial problems for any application which\nrelies on spatial correspondences being preserved between the real and the\nsynthesized image (e.g. plan, segmentation, landmark propagation). To alleviate\nthis problem, we investigated reducing the spatial information provided to the\ndiscriminator by decreasing the size of its receptive fields.\n Image synthesis was learned from 17 unpaired subjects per modality.\nRegistration performance was evaluated with respect to manual segmentations of\n11 structures for 3 subjects from the VISERAL challenge. State-of-the-art DIR\nmethods based on Normalized Mutual Information (NMI), Modality Independent\nNeighborhood Descriptor (MIND) and their novel combination achieved a mean\nsegmentation overlap ratio of 76.7, 67.7, 76.9%, respectively. This dropped to\n69.1% or less when registering images synthesized by cycle-GAN based on local\ncorrelation, due to the poor performance on the thoracic region, where large\nlung volume changes were synthesized. Performance for the abdominal region was\nsimilar to that of CT-MRI NMI registration (77.4 vs. 78.8%) when using 3D\nsynthesizing MRIs (12 slices) and medium sized receptive fields for the\ndiscriminator.\n", - "method": "We show that geometric changes, which\ndifferentiate the two populations (e.g. inhale vs.. Image synthesis was learned from 17 unpaired subjects per modality.", - "id": "1807.07349" - }, - { - "abstract": " Our team participate in the challenge of Task 1: Lesion Boundary Segmentation\n, and use a combined network, one of which is designed by ourselves named\nupdcnn net and another is an improved VGG 16-layer net. Updcnn net uses reduced\nsize images for training, and VGG 16-layer net utilizes large size images.\nImage enhancement is used to get a richer data set. We use boxes in the VGG\n16-layer net network for local attention regularization to fine-tune the loss\nfunction, which can increase the number of training data, and also make the\nmodel more robust. In the test, the model is used for joint testing and\nachieves good results.\n", - "method": "Updcnn net uses reduced\nsize images for training, and VGG 16-layer net utilizes large size images.. Image enhancement is used to get a richer data set.", - "id": "1807.07391" - }, - { - "abstract": " The emerging blockchain protocols provide a decentralized architecture that\nis suitable of supporting Internet of Things (IoT) interactions. However,\nkeeping a local copy of the blockchain ledger is infeasible for low-power and\nmemory-constrained devices. For this reason, they are equipped with lightweight\nsoftware implementations that only download the useful data structures, e.g.\nstate of accounts, from the blockchain network, when they are updated. In this\npaper, we consider and analyze a novel scheme, implemented by the nodes of the\nblockchain network, which aggregates the blockchain data in periodic updates\nand further reduces the communication cost of the connected IoT devices. We\nshow that the aggregation period should be selected based on the channel\nquality, the offered rate, and the statistics of updates of the useful data\nstructures. The results, obtained for the Ethereum protocol, illustrate the\nbenefits of the aggregation scheme in terms of a reduced duty cycle of the\ndevice, particularly for low signal-to-noise ratios, and the overall reduction\nof the amount of information transmitted in downlink (e.g., from the wireless\nbase station to the IoT device). A potential application of the proposed scheme\nis to let the IoT device request more information than actually needed, hence\nincreasing its privacy, while keeping the communication cost constant. In\nconclusion, our work is the first to provide rigorous guidelines for the design\nof lightweight blockchain protocols with wireless connectivity.\n", - "method": "For this reason, they are equipped with lightweight\nsoftware implementations that only download the useful data structures, e.g.\nstate of accounts, from the blockchain network, when they are updated.", - "id": "1807.07422" - }, - { - "abstract": " Background: Chest X-rays are the most commonly performed, cost-effective\ndiagnostic imaging tests ordered by physicians. A clinically validated AI\nsystem that can reliably separate normals from abnormals can be invaluble\nparticularly in low-resource settings. The aim of this study was to develop and\nvalidate a deep learning system to detect various abnormalities seen on a chest\nX-ray. Methods: A deep learning system was trained on 2.3 million chest X-rays\nand their corresponding radiology reports to identify various abnormalities\nseen on a Chest X-ray. The system was tested against - 1. A three-radiologist\nmajority on an independent, retrospectively collected set of 2000\nX-rays(CQ2000) 2. Radiologist reports on a separate validation set of 100,000\nscans(CQ100k). The primary accuracy measure was area under the ROC curve (AUC),\nestimated separately for each abnormality and for normal versus abnormal scans.\nResults: On the CQ2000 dataset, the deep learning system demonstrated an AUC of\n0.92(CI 0.91-0.94) for detection of abnormal scans, and AUC(CI) of\n0.96(0.94-0.98), 0.96(0.94-0.98), 0.95(0.87-1), 0.95(0.92-0.98),\n0.93(0.90-0.96), 0.89(0.83-0.94), 0.91(0.87-0.96), 0.94(0.93-0.96),\n0.98(0.97-1) for the detection of blunted costophrenic angle, cardiomegaly,\ncavity, consolidation, fibrosis, hilar enlargement, nodule, opacity and pleural\neffusion. The AUCs were similar on the larger CQ100k dataset except for\ndetecting normals where the AUC was 0.86(0.85-0.86). Interpretation: Our study\ndemonstrates that a deep learning algorithm trained on a large, well-labelled\ndataset can accurately detect multiple abnormalities on chest X-rays. As these\nsystems improve in accuracy, applying deep learning to widen the reach of chest\nX-ray interpretation and improve reporting efficiency will add tremendous value\nin radiology workflows and public health screenings globally.\n", - "method": "The system was tested against - 1.", - "id": "1807.07455" - }, - { - "abstract": " In this paper, we provide the first provable linear-time (in the number of\nnon-zero entries of the input) algorithm for approximately solving the\ngeneralized trust region subproblem (GTRS) of minimizing a quadratic function\nover a quadratic constraint under some regular conditions. Our algorithm is\nmotivated by and extends a recent linear-time algorithm for the trust region\nsubproblem by Hazan and Koren [Math. Program., 2016, 158(1-2): 363-381].\nHowever, due to the non-convexity and non-compactness of the feasible region,\nsuch an extension is nontrivial. Our main contribution is to demonstrate that\nunder some regular condition, the optimal solution is in a compact and convex\nset and lower and upper bounds of the optimal value can be computed in linear\ntime. Using these properties, we develop a linear-time algorithm for the GTRS.\n", - "method": "Program., 2016, 158(1-2): 363-381].", - "id": "1807.07563" - }, - { - "abstract": " Nearly all fast radio RRAT-type transients that are pulsars with rare pulses\nhave been previously detected using decimetre wavelengths. We present here 34\ntransients detected at metre wavelengths in our daily monitoring at\ndeclinations -9o < decl. < +42o. 25 transients are new RRATs. We confirmed the\ndetection of 7 RRATs based on our early observations. One of the 34 detected\ntransients was determined to be a new pulsar J1326+3346. At the same time, out\nof 35 RRATs detected at the decimetre wavelengths and included in the studied\narea, only one was detected by us J1848+1518. The periods of 6 RRATs were found\nfrom the time of arrival of single pulses. Three quarters of all RRATs were\nobserved more than once and the total number of RRATs in the area studied has\ndoubled.\n", - "method": "< +42o.", - "id": "1807.07565" - }, - { - "abstract": " We propose a new image denoising algorithm, dubbed as Fully Convolutional\nAdaptive Image DEnoiser (FC-AIDE), that can learn from an offline supervised\ntraining set with a fully convolutional neural network as well as adaptively\nfine-tune the supervised model for each given noisy image. We significantly\nextend the framework of the recently proposed Neural AIDE, which formulates the\ndenoiser to be context-based pixelwise mappings and utilizes the unbiased\nestimator of MSE for such denoisers. The two main contributions we make are; 1)\nimplementing a novel fully convolutional architecture that boosts the base\nsupervised model, and 2) introducing regularization methods for the adaptive\nfine-tuning such that a stronger and more robust adaptivity can be attained. As\na result, FC-AIDE is shown to possess many desirable features; it outperforms\nthe recent CNN-based state-of-the-art denoisers on all of the benchmark\ndatasets we tested, and gets particularly strong for various challenging\nscenarios, e.g., with mismatched image/noise characteristics or with scarce\nsupervised training data. The source code of our algorithm is available at\nhttps://github.com/csm9493/FC-AIDE-Keras.\n", - "method": "The two main contributions we make are; 1)\nimplementing a novel fully convolutional architecture that boosts the base\nsupervised model, and 2) introducing regularization methods for the adaptive\nfine-tuning such that a stronger and more robust adaptivity can be attained.", - "id": "1807.07569" - }, - { - "abstract": " The details of the physical mechanism that drives core-collapse supernovae\n(CCSNe) remain uncertain. While there is an emerging consensus on the\nqualitative outcome of detailed CCSN mechanism simulations in 2D, only recently\nhave high-fidelity 3D simulations become possible. Here we present the results\nof an extensive set of 3D CCSN simulations using high-fidelity multidimensional\nneutrino transport, high-resolution hydrodynamics, and approximate general\nrelativistic gravity. We employ a state-of-the-art 20 solar mass progenitor\ngenerated using the Modules for Experiments in Stellar Astrophysics (MESA;\nFarmer et al. (2016) Paxton et al. (2011, 2013, 2015, 2018) and the SFHo\nequation of state of Steiner et al. (2013). While none of our 3D CCSN\nsimulations explode within ~500ms after core bounce, we find that the presence\nof large scale aspherical motion in the Si and O shells surrounding the\ncollapsing iron core aid shock expansion and bring the models closer to the\nthreshold of explosion. We also find some dependence on resolution and geometry\n(octant vs. full 4$\\pi$). As has been noted in other recent works, we find that\nthe post-shock turbulence plays an important role in determining the overall\ndynamical evolution of our simulations. We find a strong standing accretion\nshock instability (SASI) that develops at late times during the shock recession\nepoch. The SASI aids in transient shock expansion phases, but is not enough to\nresult in shock revival. We also report that for a subset of our simulations,\nwe find conclusive evidence for the LESA first reported in Tamborra et al.\n(2014), but until now, not confirmed by other simulation codes. Both the\nprogenitor asphericities and the SASI-induced transient shock expansion phases\ngenerate transient gravitational waves and neutrino signal modulations via\nperturbations of the protoneutron star by turbulent motions. (abridged)\n", - "method": "(2013).", - "id": "1807.07579" - }, - { - "abstract": " In this note we study the logarithmic derivation module of a non-free\narrangement. We prove a generalized addition theorem for all arrangements. This\naddition theorem allows us to find various relationships between non-free\narrangements, free arrangements and restriction counts. For graphic\narrangements we can use these results to find a lower bound for the maximal\ndegree generator in terms of triangles in the associated graph. We also apply\nthese results to the case of hypersolvable arrangements where we define\nhyperexponents and use them to find a lower bound for their maximal degree\ngenerator.\n", - "method": "We prove a generalized addition theorem for all arrangements.", - "id": "1807.07613" - }, - { - "abstract": " We describe approximation algorithms in Linial's classic LOCAL model of\ndistributed computing to find maximum-weight matchings in a hypergraph of rank\n$r$. Our main result is a deterministic algorithm to generate a matching which\nis an $O(r)$-approximation to the maximum weight matching, running in $\\tilde\nO(r \\log \\Delta + \\log^2 \\Delta + \\log^* n)$ rounds. (Here, the $\\tilde O()$\nnotations hides $\\text{polyloglog } \\Delta$ and $\\text{polylog } r$ factors).\nThis is based on a number of new derandomization techniques extending methods\nof Ghaffari, Harris & Kuhn (2017).\n As a main application, we obtain nearly-optimal algorithms for the\nlong-studied problem of maximum-weight graph matching. Specifically, we get a\n$(1+\\epsilon)$ approximation algorithm using $\\tilde O(\\log \\Delta / \\epsilon^3\n+ \\text{polylog}(1/\\epsilon, \\log \\log n))$ randomized time and $\\tilde\nO(\\log^2 \\Delta / \\epsilon^4 + \\log^*n / \\epsilon)$ deterministic time.\n The second application is a faster algorithm for hypergraph maximal matching,\na versatile subroutine introduced in Ghaffari et al. (2017) for a variety of\nlocal graph algorithms. This gives an algorithm for $(2 \\Delta - 1)$-edge-list\ncoloring in $\\tilde O(\\log^2 \\Delta \\log n)$ rounds deterministically or\n$\\tilde O( (\\log \\log n)^3 )$ rounds randomly. Another consequence (with\nadditional optimizations) is an algorithm which generates an edge-orientation\nwith out-degree at most $\\lceil (1+\\epsilon) \\lambda \\rceil$ for a graph of\narboricity $\\lambda$; for fixed $\\epsilon$ this runs in $\\tilde O(\\log^6 n)$\nrounds deterministically or $\\tilde O(\\log^3 n )$ rounds randomly.\n", - "method": "This is based on a number of new derandomization techniques extending methods\nof Ghaffari, Harris & Kuhn (2017).", - "id": "1807.07645" - }, - { - "abstract": " Employers actively look for talents having not only specific hard skills but\nalso various soft skills. To analyze the soft skill demands on the job market,\nit is important to be able to detect soft skill phrases from job advertisements\nautomatically. However, a naive matching of soft skill phrases can lead to\nfalse positive matches when a soft skill phrase, such as friendly, is used to\ndescribe a company, a team, or another entity, rather than a desired candidate.\n In this paper, we propose a phrase-matching-based approach which\ndifferentiates between soft skill phrases referring to a candidate vs.\nsomething else. The disambiguation is formulated as a binary text\nclassification problem where the prediction is made for the potential soft\nskill based on the context where it occurs. To inform the model about the soft\nskill for which the prediction is made, we develop several approaches,\nincluding soft skill masking and soft skill tagging.\n We compare several neural network based approaches, including CNN, LSTM and\nHierarchical Attention Model. The proposed tagging-based input representation\nusing LSTM achieved the highest recall of 83.92% on the job dataset when fixing\na precision to 95%.\n", - "method": "something else.", - "id": "1807.07741" - }, - { - "abstract": " 3D reconstruction from single view images is an ill-posed problem. Inferring\nthe hidden regions from self-occluded images is both challenging and ambiguous.\nWe propose a two-pronged approach to address these issues. To better\nincorporate the data prior and generate meaningful reconstructions, we propose\n3D-LMNet, a latent embedding matching approach for 3D reconstruction. We first\ntrain a 3D point cloud auto-encoder and then learn a mapping from the 2D image\nto the corresponding learnt embedding. To tackle the issue of uncertainty in\nthe reconstruction, we predict multiple reconstructions that are consistent\nwith the input view. This is achieved by learning a probablistic latent space\nwith a novel view-specific diversity loss. Thorough quantitative and\nqualitative analysis is performed to highlight the significance of the proposed\napproach. We outperform state-of-the-art approaches on the task of single-view\n3D reconstruction on both real and synthetic datasets while generating multiple\nplausible reconstructions, demonstrating the generalizability and utility of\nour approach.\n", - "method": "We propose a two-pronged approach to address these issues.", - "id": "1807.07796" - }, - { - "abstract": " We study the effect of the chirped laser pulse on the transmission and\nassociated ion acceleration by the sub-wavelength target. In the chirped laser\npulses, the pulse frequency has a temporal variation about its fundamental\nfrequency,\n which manifests to the temporal dependence of the critical density ($n_c$).\nIn this work we used a chirp model which is beyond the linear approximation.\nFor negatively (positively) chirped pulses, the high (low) frequency component\nof the pulse interacts with the target initially followed by the low (high)\nfrequency component. The threshold plasma density for the transmission of the\npulse is found to be higher for the negatively chirped laser pulses as compared\nto the unchirped or positively chirped pulses. The enhanced transmission of the\nnegatively chirped pulses for higher densities ($6n_c$) results in very\nefficient heating of the target electrons, creating a very stable and\npersistent longitudinal electrostatic field behind the target. The void of the\nelectrons results in expansion of the target ions in either direction,\nresulting in the broad energy spectrum. We have introduced a very thin, low\ndensity ($< n_c$) secondary layer behind the primary layer. The ions from the\nsecondary layer are then found to be accelerated as a mono-energetic bunch\nunder the influence of the electrostatic field created by the primary layer\nupon interaction by the negatively chirped pulse. Under the optimum conditions,\nthe maximum energy of the protons are found to be $\\sim 100$ MeV for 10 fs\n(intensity fwhm); Circularly Polarized; Gaussian; negatively chirped laser\npulse with peak intensity $\\sim 8.5\\times 10^{20}$ W/cm$^2$.\n", - "method": "In this work we used a chirp model which is beyond the linear approximation.", - "id": "1807.07804" - }, - { - "abstract": " The research goal is to investigate the evolution of the concept of utility\ncomputing in the works of foreign researchers in the years 1959-1966. First the\nA. O. Mann's results and expanded overview of the D. F. Parkhill's results on\nthe concept of computer (information) utility were introduced in the domestic\nscientific circulation. Functionally identity of the computer utility and cloud\ncomputing concepts was proved, as well as refined the primary sources of cloud\nservice models. There was proposed the interpretation of the \"cloud\ntechnologies of learning\" concept. Continuity of the development of cloud\ntechnologies over the past 55 years and their relationship with the development\nof ICT in general was concluded. The research results make it possible to\ndetermine the prospects of the development of cloud computing in general and\ncloud technologies of learning in particular.\n", - "method": "O.. Mann's results and expanded overview of the D.. First the\nA.", - "id": "1807.07849" - }, - { - "abstract": " In this paper, we show that, under some technical assumptions, the\nKolmogorov-Sinai entropy and the permutation entropy are equal for\none-dimensional maps if there exists a countable partition of the domain of\ndefinition into intervals such that the considered map is monotone on each of\nthose intervals. This is a generalization of a result by Bandt, Pompe and G.\nKeller, who showed that the above holds true under the additional assumptions\nthat the number of intervals on which the map is monotone is finite and that\nthe map is continuous on each of those intervals.\n", - "method": "This is a generalization of a result by Bandt, Pompe and G.", - "id": "1807.07881" - }, - { - "abstract": " Shared memory emulation can be used as a fault-tolerant and highly available\ndistributed storage solution or as a low-level synchronization primitive.\nAttiya, Bar-Noy, and Dolev were the first to propose a single-writer,\nmulti-reader linearizable register emulation where the register is replicated\nto all servers. Recently, Cadambe et al. proposed the Coded Atomic Storage\n(CAS) algorithm, which uses erasure coding for achieving data redundancy with\nmuch lower communication cost than previous algorithmic solutions.\n Although CAS can tolerate server crashes, it was not designed to recover from\nunexpected, transient faults, without the need of external (human)\nintervention. In this respect, Dolev, Petig, and Schiller have recently\ndeveloped a self-stabilizing version of CAS, which we call CASSS. As one would\nexpect, self-stabilization does not come as a free lunch; it introduces,\nmainly, communication overhead for detecting inconsistencies and stale\ninformation. So, one would wonder whether the overhead introduced by\nself-stabilization would nullify the gain of erasure coding.\n To answer this question, we have implemented and experimentally evaluated the\nCASSS algorithm on PlanetLab; a planetary scale distributed infrastructure. The\nevaluation shows that our implementation of CASSS scales very well in terms of\nthe number of servers, the number of concurrent clients, as well as the size of\nthe replicated object. More importantly, it shows (a) to have only a constant\noverhead compared to the traditional CAS algorithm (which we also implement)\nand (b) the recovery period (after the last occurrence of a transient fault) is\nas fast as a few client (read/write) operations. Our results suggest that CASSS\ndoes not significantly impact efficiency while dealing with automatic recovery\nfrom transient faults and bounded size of needed resources.\n", - "method": "Recently, Cadambe et al.", - "id": "1807.07901" - }, - { - "abstract": " Superconductivity in FeSe has recently attracted a great deal of attention\nbecause it emerges out of an electronic nematic state of elusive character.\nHere we study both the electronic normal state and the superconducting gap\nstructure using heat-capacity measurements on high-quality single crystals. The\nspecific-heat curve, from 0.4 K to Tc = 9.1 K, is found to be consistent with a\nrecent gap determination using Bogoliubov quasiparticle interference [P. O.\nSprau et al., Science 357, 75 (2017)], however only if nodes are introduced on\neither the electron or the hole Fermi-surface sheets. Our analysis, which is\nconsistent with quantum-oscillation measurements, relies on the presence of\nonly two bands, and thus the fate of the theoretically predicted second\nelectron pocket remains mysterious.\n", - "method": "O.", - "id": "1807.07907" - }, - { - "abstract": " Inverse problems in imaging such as denoising, deblurring, superresolution\n(SR) have been addressed for many decades. In recent years, convolutional\nneural networks (CNNs) have been widely used for many inverse problem areas.\nAlthough their indisputable success, CNNs are not mathematically validated as\nto how and what they learn. In this paper, we prove that during training, CNN\nelements solve for inverse problems which are optimum solutions stored as CNN\nneuron filters. We discuss the necessity of mutual coherence between CNN layer\nelements in order for a network to converge to the optimum solution. We prove\nthat required mutual coherence can be provided by the usage of residual\nlearning and skip connections. We have set rules over training sets and depth\nof networks for better convergence, i.e. performance.\n", - "method": "We have set rules over training sets and depth\nof networks for better convergence, i.e. performance.", - "id": "1807.07998" - }, - { - "abstract": " GW170817, the milestone gravitational-wave event originated from a binary\nneutron star merger, has allowed scientific community to place a constraint on\nthe equation of state of neutron stars by extracting the leading-order,\ntidal-deformability term from the gravitational waveform. Here we incorporate\ntidal corrections to the gravitational-wave phase at next-to-leading and\nnext-to-next-to-leading order, including the magnetic tidal Love numbers, tail\neffects, and the spin-tidal couplings recently computed in Tiziano Abdelsalhin\n[Phys. Rev. D 98, 104046 (2018)]. These effects have not yet been included in\nthe waveform approximants for the analysis of GW170817. We provide a\nqualitative and quantitative analysis of the impact of these new terms by\nstudying the parameter bias induced on events compatible with GW170817 assuming\nsecond-generation (advanced LIGO) and third-generation (Einstein Telescope)\nground-based gravitational-wave interferometers. We find that including the\ntidal-tail term deteriorates the convergence properties of the post-Newtonian\nexpansion in the relevant frequency range. We also find that the effect of\nmagnetic tidal Love numbers could be measurable for an optimal GW170817 event\nwith signal-to-noise ratio $\\rho \\approx 1750$ detected with the Einstein\nTelescope. On the same line, spin-tidal couplings may be relevant if mildly\nhigh-spin $\\chi \\gtrsim 0.1$ neutron star binaries exist in nature.\n", - "method": "Rev.. D 98, 104046 (2018)].", - "id": "1807.08016" - }, - { - "abstract": " X-ray absorption fine structure (XAFS) and x-ray emission spectroscopy (XES)\nare advanced x-ray spectroscopies that impact a wide range of disciplines.\nHowever, unlike the majority of other spectroscopic methods, XAFS and XES are\naccompanied by an unusual access model, wherein; the dominant use of the\ntechnique is for premier research studies at world-class facilities, i.e.,\nsynchrotron x-ray light sources. In this paper we report the design and\nperformance of an improved spectrometer XAFS and XES based on the general\nconceptual design of Seidler, et al., Rev. Sci. Instrum. 2014. New developments\ninclude reduced mechanical degrees of freedom, much-increased flux, and a wider\nBragg angle range to enable extended x-ray absorption fine structure (EXAFS)\nfor the first time with this type of modern laboratory XAFS configuration. This\ninstrument enables a new class of routine applications that are incompatible\nwith the mission and access model of the synchrotron light sources. To\nillustrate this, we provide numerous examples of x-ray absorption near edge\nstructure (XANES), EXAFS, and XES results for a variety of problems and energy\nranges. Highlights include XAFS and XES measurements of battery electrode\nmaterials, EXAFS of Ni and V with full modeling of results to validate\nmonochromator performance, valence-to-core XES for 3d transition metal\ncompounds, and uranium XANES and XES for different oxidation states. Taken en\nmasse, these results further support the growing perspective that modern\nlaboratory-based XAFS and XES have the potential to develop a new branch of\nanalytical chemistry.\n", - "method": "Sci.. Instrum.. 2014.", - "id": "1807.08059" - }, - { - "abstract": " In a recent article by M. Bona and A. Vince, the authors introduced the\nconcept of an assembly tree for a graph. Assembly trees act as record keeping\ndevices for the construction of a given graph from its vertices. In this paper\nwe extend the work initiated by M. Bona and A. Vince as well as define and\nexamine a more generalized definition of assembly tree.\n", - "method": "Bona and A.", - "id": "1807.08079" - }, - { - "abstract": " The Alaskan Stream is the northern boundary current in the subarctic North\nPacific. This area is characterized by significant temperature, salinity and\ndensity differences between coastal and open-ocean waters and strong mesoscale\ndynamics. In this paper we demonstrate the transport pathways of Alaskan Stream\nwater in the eastern subarctic Pacific and the eastern Bering Sea from October\n1, 1994 to September 12, 2016 with the help of altimetry-based Lagrangian maps.\nA mesoscale eddy activity along the shelf-deep basin boundaries in the Alaskan\nStream region and the eastern Bering Sea is shown to be related with the wind\nstress curl in the northern North Pacific in winter. A significant correlation\nis found between the concentration of chlorophyll a in the Alaskan Stream area\nand eastern Bering Sea in August - September and the wind stress curl in the\nnorthern North Pacific in November - March. The mesoscale dynamics, forced by\nthe wind stress curl in winter, may determine not only lower-trophic-level\norganism biomass but also salmon abundance/catch in the study area.\n", - "method": "In this paper we demonstrate the transport pathways of Alaskan Stream\nwater in the eastern subarctic Pacific and the eastern Bering Sea from October\n1, 1994 to September 12, 2016 with the help of altimetry-based Lagrangian maps.", - "id": "1807.08094" - }, - { - "abstract": " A new convenient method of describing flat convex compact sets is proposed.\nIt generalizes classical trigonometric functions $\\sin$ and $\\cos$. Apparently,\nthis method may be very useful for explicit description of solutions of optimal\ncontrol problems with two-dimensional control. Using this method a series of\nsub-Finsler problems with two-dimensional control lying in an arbitrary convex\nset $\\Omega$ is investigated. Namely, problems on the Heisenberg, Engel, and\nCartan groups and also Grushin's and Martinet's cases are considered. A\nparticular attention is paid to the case when $\\Omega$ is a polygon.\n", - "method": "A new convenient method of describing flat convex compact sets is proposed.", - "id": "1807.08155" - }, - { - "abstract": " Runtime monitoring is commonly used to detect the violation of desired\nproperties in safety critical cyber-physical systems by observing its\nexecutions. Bauer et al. introduced an influential framework for monitoring\nLinear Temporal Logic (LTL) properties based on a three-valued semantics: the\nformula is already satisfied by the given prefix, it is already violated, or it\nis still undetermined, i.e., it can still be satisfied and violated by\nappropriate extensions. However, a wide range of formulas are not monitorable\nunder this approach, meaning that they have a prefix for which satisfaction and\nviolation will always remain undetermined no matter how it is extended. In\nparticular, Bauer et al. report that 44% of the formulas they consider in their\nexperiments fall into this category.\n Recently, a robust semantics for LTL was introduced to capture different\ndegrees by which a property can be violated. In this paper we introduce a\nrobust semantics for finite strings and show its potential in monitoring: every\nformula considered by Bauer et al. is monitorable under our approach.\nFurthermore, we discuss which properties that come naturally in LTL monitoring\n- such as the realizability of all truth values - can be transferred to the\nrobust setting. Lastly, we show that LTL formulas with robust semantics can be\nmonitored by deterministic automata and report on a prototype implementation.\n", - "method": "Bauer et al.. is monitorable under our approach.", - "id": "1807.08203" - }, - { - "abstract": " Subset Sum is a classical optimization problem taught to undergraduates as an\nexample of an NP-hard problem, which is amenable to dynamic programming,\nyielding polynomial running time if the input numbers are relatively small.\nFormally, given a set $S$ of $n$ positive integers and a target integer $t$,\nthe Subset Sum problem is to decide if there is a subset of $S$ that sums up to\n$t$. Dynamic programming yields an algorithm with running time $O(nt)$.\nRecently, the authors [SODA '17] improved the running time to\n$\\tilde{O}\\bigl(\\sqrt{n}t\\bigr)$, and it was further improved to\n$\\tilde{O}\\bigl(n+t\\bigr)$ by a somewhat involved randomized algorithm by\nBringmann [SODA '17], where $\\tilde{O}$ hides polylogarithmic factors.\n Here, we present a new and significantly simpler algorithm with running time\n$\\tilde{O}\\bigl(\\sqrt{n}t\\bigr)$. While not the fastest, we believe the new\nalgorithm and analysis are simple enough to be presented in an algorithms\nclass, as a striking example of a divide-and-conquer algorithm that uses FFT to\na problem that seems (at first) unrelated. In particular, the algorithm and its\nanalysis can be described in full detail in two pages (see pages 3-5).\n", - "method": "Dynamic programming yields an algorithm with running time $O(nt)$.", - "id": "1807.08248" - }, - { - "abstract": " The minimum stretch spaning tree problem for a grah G is to find a spaning\ntree T of G such as that the maximum distance in T between two adjacent\nvertices is minimized. The minimum value of this optimization problem gives\nrise to a grpah invariant {\\sigma} T(G) called the tree stretch of G. The\nproblem has been studied in the algorithmic aspects, such as NP-hardness and\nfixed-parameter solvability. This paper presents the exact values {\\sigma} T(G)\nof hamming graphs Kn1 * Kn2 * ... * Knd and the higer-dimensional grids Pn1 *\nPn2 * ... * Pnd.\n", - "method": "* Pnd.", - "id": "1807.08252" - }, - { - "abstract": " In human parsing, the pixel-wise classification loss has drawbacks in its\nlow-level local inconsistency and high-level semantic inconsistency. The\nintroduction of the adversarial network tackles the two problems using a single\ndiscriminator. However, the two types of parsing inconsistency are generated by\ndistinct mechanisms, so it is difficult for a single discriminator to solve\nthem both. To address the two kinds of inconsistencies, this paper proposes the\nMacro-Micro Adversarial Net (MMAN). It has two discriminators. One\ndiscriminator, Macro D, acts on the low-resolution label map and penalizes\nsemantic inconsistency, e.g., misplaced body parts. The other discriminator,\nMicro D, focuses on multiple patches of the high-resolution label map to\naddress the local inconsistency, e.g., blur and hole. Compared with traditional\nadversarial networks, MMAN not only enforces local and semantic consistency\nexplicitly, but also avoids the poor convergence problem of adversarial\nnetworks when handling high resolution images. In our experiment, we validate\nthat the two discriminators are complementary to each other in improving the\nhuman parsing accuracy. The proposed framework is capable of producing\ncompetitive parsing performance compared with the state-of-the-art methods,\ni.e., mIoU=46.81% and 59.91% on LIP and PASCAL-Person-Part, respectively. On a\nrelatively small dataset PPSS, our pre-trained model demonstrates impressive\ngeneralization ability. The code is publicly available at\nhttps://github.com/RoyalVane/MMAN.\n", - "method": "It has two discriminators.", - "id": "1807.08260" - }, - { - "abstract": " We show that a quantum particle in $\\mathbb{R}^d$, for $d \\geq 1$, subject to\na white-noise potential, moves super-ballistically in the sense that the mean\nsquare displacement $\\int \\|x\\|^2 \\langle \\rho(x,x,t) \\rangle ~dx$ grows like\n$t^{3}$ in any dimension. The white noise potential is Gaussian distributed\nwith an arbitrary spatial correlation function and a delta correlation function\nin time. This is a known result in one dimension (see refs. Fischer, Leschke,\nM\\\"uller and Javannar, Kumar}. The energy of the system is also shown to\nincrease linearly in time. We also prove that for the same white-noise\npotential model on the lattice $\\mathbb{Z}^d$, for $d \\geq 1$, the mean square\ndisplacement is diffusive growing like $t^{1}$. This behavior on the lattice is\nconsistent with the diffusive behavior observed for similar models in the\nlattice $\\mathbb{Z}^d$ with a time-dependent Markovian potential (see ref.\nKang, Schenker).\n", - "method": "This is a known result in one dimension (see refs.", - "id": "1807.08317" - }, - { - "abstract": " In this note, we examine the bundle picture of the pullback construction of\nLie algebroids. The notion of submersions by Lie algebroids is introduced,\nwhich leads to a new proof of the local normal form for lie algebroid\ntransversals of [Bursztyn et al., Crelle, 2017], and which we use to deduce\nthat Lie algebroids transversals concentrate all local cohomology.\n The locally trivial version of submersions by Lie algebroids $\\mathfrak{S}$\nis then discussed, and we show that this notion is equivalent to the existence\nof a complete Ehresmann connection for $\\mathfrak{S}$, extending the main\nresult in [del Hoyo, Indag. Math. 2016].\n Finally, we show that locally trivial version of submersions by Lie\nalgebroids gives rise to a system of local coefficients, which is an integral\npart of a version of the homotopy invariance of de Rham cohomology in the\ncontext of Lie algebroids, and we apply such local systems to extend the\nlocalization theorem of [Chen et al. 2006].\n", - "method": "Math.. 2016].. 2006].", - "id": "1807.08382" - }, - { - "abstract": " Observations of the black hole shadow of supermassive black holes, such as\nSagittarius A* at the center of our Milky Way galaxy, allow us to study the\nproperties of black holes and the nature of strong-field gravity. According to\nthe Kerr hypothesis, isolated, stationary, and axisymmetric astrophysical black\nholes are described by the Kerr metric. The Kerr hypothesis holds in General\nRelativity and in some modified gravity theories, but there are others in which\nit is violated. In principle, black hole shadow observations can be used to\ndetermine if the Kerr metric is the correct description for black holes, and in\nturn, they could be used to place constraints on modified gravity theories that\ndo not admit the Kerr solution. We here investigate whether black hole shadow\nobservations can constrain deviations from general relativity, focusing on two\nwell-motivated modified quadratic gravity theories:\nEinstein-dilaton-Gauss-Bonnet gravity and dynamical Chern-Simons gravity. We\nshow that current constraints on Einstein-dilaton-Gauss-Bonnet gravity are\nstronger than any that could be placed with black hole shadow observations of\nsupermassive black holes. We also show that the same holds for dynamical\nChern-Simons gravity through a systematic bias and a likelihood analysis when\nconsidering slowly-rotating supermassive black holes. However, observations of\nmore rapidly-rotating black holes, with dimensionless spins\n$|\\vec{J}|/M^{2}\\simeq0.5$, could be used to better constrain dynamical\nChern-Simons gravity.\n", - "method": "We here investigate whether black hole shadow\nobservations can constrain deviations from general relativity, focusing on two\nwell-motivated modified quadratic gravity theories:\nEinstein-dilaton-Gauss-Bonnet gravity and dynamical Chern-Simons gravity.", - "id": "1807.08422" - }, - { - "abstract": " We participated the Task 1: Lesion Segmentation. The paper describes our\nalgorithm and the final result of validation set for the ISIC Challenge 2018 -\nSkin Lesion Analysis Towards Melanoma Detection.\n", - "method": "We participated the Task 1: Lesion Segmentation.", - "id": "1807.08471" - }, - { - "abstract": " The first billion years in the evolution of the Universe mark the formation\nof the first stars, black holes and galaxies. The radiation from the first\ngalaxies plays an important role in determining the final state of primordial\ngas collapsing in a neighboring halo. This is due to the fact that the primary\ncoolant for primordial gas is molecular hydrogen, which can be dissociated into\natomic hydrogen by Lyman-Werner photons in the energy range $11.2 - 13.6$~eV.\nWhile cooling by molecular hydrogen leads to Pop. III star formation, cooling\nby atomic hydrogen can lead to the formation of a supermassive star (or\nquasi-star) which results in the formation of a massive $10^{4-5} M_\\odot$\nblack hole, or a direct collapse black hole. The spectrum of this radiation\nfield is critical in order to determine whether a primordial gas cloud forms a\nPop. III star or a very massive black hole. We will in the following explore\nthis scenario and discuss how the radiation spectrum influences the outcome of\nthe collapse.\n", - "method": "While cooling by molecular hydrogen leads to Pop.", - "id": "1807.08499" - }, - { - "abstract": " The solubility of a \"heavy\" alkane (solute) in a \"light\" alkane (solvent) is\ngenerally temperature dependent. Moreover, it is determined by the molar masses\nof the solute and solvent. In the current paper, published solubility data for\nbinary normal-alkane mixtures is reviewed (solid-liquid equilibrium). A total\nof 43 unique solute-solvent data-sets, obtained from a total of 24 published\nexperimental studies, are collected and presented in a systematic manner. Based\non thermodynamic considerations and the experimental data, it is demonstrated\nthat there is a log-linear relationship between the solubility and the\ntemperature in the dilute range. Linear regression is employed to 1) obtain\ndata-set-specific solubility-temperature best-fit parameters and 2) obtain a\ngeneral correlation between the solubility and the solvent and solute molar\nmasses and the temperature. Finally, it is demonstrated that the developed\ncorrelation carries predictive power even for multi-component mixtures by\nutilizing solvent and solute average molar masses.\n", - "method": "Linear regression is employed to 1) obtain\ndata-set-specific solubility-temperature best-fit parameters and 2) obtain a\ngeneral correlation between the solubility and the solvent and solute molar\nmasses and the temperature.", - "id": "1807.08514" - }, - { - "abstract": " In this paper, we investigate the properties of a random walk on the\nalternating group $A_n$ generated by $3$-cycles of the form $(i,n-1,n)$ and\n$(i,n,n-1)$. We call this the transpose top-$2$ with random shuffle. We find\nthe spectrum of the transition matrix of this shuffle. We show that the mixing\ntime is of order $\\left(n-\\frac{3}{2}\\right)\\log n$ and prove that there is a\ntotal variation cutoff for this shuffle.\n", - "method": "We call this the transpose top-$2$ with random shuffle.", - "id": "1807.08539" - }, - { - "abstract": " We formulate real-space dynamical mean-field theory within scattering theory.\nThereby the Friedel sum rule is derived for interacting lattice fermions at\nzero temperature.\n", - "method": "We formulate real-space dynamical mean-field theory within scattering theory.", - "id": "1807.08559" - }, - { - "abstract": " We propose an end-to-end deep learning method that learns to estimate\nemphysema extent from proportions of the diseased tissue. These proportions\nwere visually estimated by experts using a standard grading system, in which\ngrades correspond to intervals (label example: 1-5% of diseased tissue). The\nproposed architecture encodes the knowledge that the labels represent a\nvolumetric proportion. A custom loss is designed to learn with intervals. Thus,\nduring training, our network learns to segment the diseased tissue such that\nits proportions fit the ground truth intervals. Our architecture and loss\ncombined improve the performance substantially (8% ICC) compared to a more\nconventional regression network. We outperform traditional lung densitometry\nand two recently published methods for emphysema quantification by a large\nmargin (at least 7% AUC and 15% ICC), and achieve near-human-level performance.\nMoreover, our method generates emphysema segmentations that predict the spatial\ndistribution of emphysema at human level.\n", - "method": "These proportions\nwere visually estimated by experts using a standard grading system, in which\ngrades correspond to intervals (label example: 1-5% of diseased tissue).", - "id": "1807.08601" - }, - { - "abstract": " Two meromorphic functions $f$ and $g$ are said to share the set $S\\subset\n\\mathbb{C}\\cup\\{\\infty\\}$ with weight $l\\in\\mathbb{N}\\cup\\{0\\}\\cup\\{\\infty\\}$,\nif $E_{f}(S,l)=E_{g}(S,l)$ where $$E_{f}(S,l)=\\bigcup\\limits_{a \\in S}\\{(z,t)\n\\in \\mathbb{C}\\times\\mathbb{N}~ |~ f(z)=a ~\\text{with~ multiplicity}~ p\\},$$\nwhere $t=p$ if $p\\leq l$ and $t=p+1$ if $p>l$.\n In this paper, we improve and supplement the result of L. W. Liao and C. C.\nYang (On the cardinality of the unique range sets for meromorphic and entire\nfunctions, Indian J. Pure appl. Math., 31 (2000), no. 4, 431-440) by showing\nthat there exist a finite set $S$ with cardinality $\\geq 13$ such that\n$E_{f}(S,1)=E_{g}(S,1)$ implies $f\\equiv g$.\n", - "method": "W.. Liao and C.. Two meromorphic functions $f$ and $g$ are said to share the set $S\\subset\n\\mathbb{C}\\cup\\{\\infty\\}$ with weight $l\\in\\mathbb{N}\\cup\\{0\\}\\cup\\{\\infty\\}$,\nif $E_{f}(S,l)=E_{g}(S,l)$ where $$E_{f}(S,l)=\\bigcup\\limits_{a \\in S}\\{(z,t)\n\\in \\mathbb{C}\\times\\mathbb{N}~ |~ f(z)=a ~\\text{with~ multiplicity}~ p\\},$$\nwhere $t=p$ if $p\\leq l$ and $t=p+1$ if $p>l$.", - "id": "1807.08619" - }, - { - "abstract": " The von Neumann entropy is a key quantity in quantum information theory and,\nroughly speaking, quantifies the amount of quantum information contained in a\nstate when many identical and independent i.i.d. copies of the state are\navailable, in a regime that is often referred to as being asymptotic. In this\nwork, we provide a new operational characterization of the von Neumann entropy\nwhich neither requires an i.i.d. limit nor any explicit randomness. We do so by\nshowing that the von Neumann entropy fully characterizes single-shot state\ntransitions in unitary quantum mechanics, as long as one has access to a\ncatalyst - an ancillary system that can be re-used after the transition - and\nan environment which has the effect of dephasing in a preferred basis. Building\nupon these insights, we formulate and provide evidence for the catalytic\nentropy conjecture, which states that the above result holds true even in the\nabsence of decoherence. If true, this would prove an intimate connection\nbetween single-shot state transitions in unitary quantum mechanics and the von\nNeumann entropy. Our results add significant support to recent insights that,\ncontrary to common wisdom, the standard von Neumann entropy also characterizes\nsingle-shot situations and opens up the possibility for operational single-shot\ninterpretations of other standard entropic quantities. We discuss implications\nof these insights to readings of the third law of quantum thermodynamics and\nhint at potentially profound implications to holography.\n", - "method": "In this\nwork, we provide a new operational characterization of the von Neumann entropy\nwhich neither requires an i.i.d. limit nor any explicit randomness.", - "id": "1807.08773" - }, - { - "abstract": " Ground-layer adaptive optics (GLAO) systems offer the possibility of\nimproving the \"seeing\" of large ground-based telescopes and increasing the\nefficiency and sensitivity of observations over a wide field-of-view. We\nexplore the utility and feasibility of deploying a GLAO system at the W. M.\nKeck Observatory in order to feed existing and future multi-object\nspectrographs and wide-field imagers. We also briefly summarize science cases\nspanning exoplanets to high-redshift galaxy evolution that would benefit from a\nKeck GLAO system. Initial simulations indicate that a Keck GLAO system would\ndeliver a 1.5x and 2x improvement in FWHM at optical (500 nm) and infrared (1.5\nmicron), respectively. The infrared instrument, MOSFIRE, is ideally suited for\na Keck GLAO feed in that it has excellent image quality and is on the\ntelescope's optical axis. However, it lacks an atmospheric dispersion\ncompensator, which would limit the minimum usable slit size for long-exposure\nscience cases. Similarly, while LRIS and DEIMOS may be able to accept a GLAO\nfeed based on their internal image quality, they lack either an atmospheric\ndispersion compensator (DEIMOS) or flexure compensation (LRIS) to utilize\nnarrower slits matched to the GLAO image quality. However, some science cases\nneeding shorter exposures may still benefit from Keck GLAO and we will\ninvestigate the possibility of installing an ADC.\n", - "method": "M.", - "id": "1807.08832" - }, - { - "abstract": " We construct the lattice AdS geometry. The lattice AdS$_2$ geometry and\nAdS$_3$ geometry can be extended from the lattice AdS$_2$ induced metric, which\nprovided the lattice Schwarzian theory at the classical limit. Then we use the\nlattice embedding coordinates to rewrite the lattice AdS$_2$ geometry and\nAdS$_3$ geometry with the manifest isometry. The lattice AdS$_2$ geometry can\nbe obtained from the lattice AdS$_3$ geometry through the compactification\nwithout the lattice artifact. The lattice embedding coordinates can also be\nused in the higher dimensional AdS geometry. Because the lattice Schwarzian\ntheory does not suffer from the issue of the continuum limit, the lattice\nAdS$_2$ geometry can be obtained from the higher dimensional AdS geometry\nthrough the compactification, and the lattice AdS metric does not depend on the\nangular coordinates, we expect that the continuum limit should exist in the\nlattice Einstein gravity theory from this geometric lattice AdS geometry.\nFinally, we apply this lattice construction to construct the holographic tensor\nnetwork without the issue of a continuum limit.\n", - "method": "We construct the lattice AdS geometry.", - "id": "1807.08871" - }, - { - "abstract": " We consider spectra of $n$-by-$n$ irreducible tridiagonal matrices over a\nfield and of their $n-1$-by-$n-1$ trailing principal submatrices. The real\nsymmetric and complex Hermitian cases have been fully understood: it is\nnecessary and sufficient that the necessarily real eigenvalues are distinct and\nthose of the principal submatrix strictly interlace. So this case is very\nrestrictive.\n By contrast, for a general field, the requirements on the two spectra are\nmuch less restrictive. In particular, in the real or complex case, the\n$n$-by-$n$ characteristic polynomial is arbitrary (so that the algebraic\nmultiplicities may be anything in place of all 1's in the classical cases) and\nthat of the principal submatrix is the complement of a lower dimensional\nalgebraic set (and so relatively free). Explicit conditions are given.\n", - "method": "Explicit conditions are given.", - "id": "1807.08877" - }, - { - "abstract": " An ideal fusion method preserves the Spectral information in fused image and\nadds spatial information to it with no spectral distortion. Recently wavelet\nkalman filter method is proposed which uses ARSIS concept to fuses MS and PAN\nimages. This method is applied in a multiscale version, i.e. the variable index\nis scale instead of time. With the aim of fusion we present a more detailed\nstudy on this model and discuss about rationality of its assumptions such as\nfirst order markov model and Gaussian distribution of the posterior density.\nFinally, we propose a method using wavelet Kalman Particle filter to improve\nthe spectral and spatial quality of the fused image. We show that our model is\nmore consistent with natural MS and PAN images. Visual and statistical analyzes\nshow that the proposed algorithm clearly improves the fusion quality in terms\nof: correlation coefficient, ERGAS, UIQI, and Q4; compared to other methods\nincluding IHS, HMP, PCA, A`trous, udWI, udWPC, Adaptive IHS, Improved Adaptive\nPCA and wavelet kalman filter.\n", - "method": "This method is applied in a multiscale version, i.e. the variable index\nis scale instead of time.", - "id": "1807.08917" - }, - { - "abstract": " We apply the semi-discrete method, c.f. \\emph{N. Halidias and I.S. Stamatiou\n(2016), On the numerical solution of some non-linear stochastic differential\nequations using the semi-discrete method, Computational Methods in Applied\nMathematics, 16(1)}, to a class of non-colliding particle systems. The proposed\nnumerical scheme preserves the non-colliding property and strongly converges to\nthe exact solution.\n", - "method": "We apply the semi-discrete method, c.f. \\emph{N.. We apply the semi-discrete method, c.f. \\emph{N.. Halidias and I.S. Stamatiou\n(2016), On the numerical solution of some non-linear stochastic differential\nequations using the semi-discrete method, Computational Methods in Applied\nMathematics, 16(1)}, to a class of non-colliding particle systems.", - "id": "1807.08924" - }, - { - "abstract": " The change of the precession angle is studied analytically and numerically\nfor the integrable tops of Kovalevskaya and Goryachev-Chaplygin. Based on the\nknown results on the topology of Liouville foliations for these systems, we\nfind initial conditions for which the average change of the precession angle is\nzero or can be estimated asymptotically. Some more difficult cases are studied\nnumerically.\n", - "method": "The change of the precession angle is studied analytically and numerically\nfor the integrable tops of Kovalevskaya and Goryachev-Chaplygin.", - "id": "1807.09016" - }, - { - "abstract": " Standard random-effects meta-analysis methods perform poorly when applied to\nfew studies only. Such settings however are commonly encountered in practice.\nIt is unclear, whether or to what extent small-sample-size behaviour can be\nimproved by more sophisticated modeling.\n We consider several likelihood-based inference methods. Confidence intervals\nare based on normal or Student-t approximations. We extract an empirical data\nset of 40 meta-analyses from recent reviews published by the German Institute\nfor Quality and Efficiency in Health Care (IQWiG). Methods are then compared\nempirically as well as in a simulation study, considering odds-ratio and risk\nratio effect sizes.\n Empirically, a majority of the identified meta-analyses include only 2\nstudies. In the simulation study, coverage probability is, in the presence of\nheterogeneity and few studies, below the nominal level for all frequentist\nmethods based on normal approximation, in particular when sizes in\nmeta-analyses are not balanced, but improve when confidence intervals are\nadjusted. Bayesian methods result in better coverage than the frequentist\nmethods with normal approximation in all scenarios. Credible intervals are\nempirically and in the simulation study wider than unadjusted confidence\nintervals, but considerably narrower than adjusted ones. Confidence intervals\nbased on the generalized linear mixed models are in general, slightly narrower\nthan those from other frequentist methods. Certain methods turned out\nimpractical due to frequent numerical problems.\n In the presence of between-study heterogeneity, especially with unbalanced\nstudy sizes, caution is needed in applying meta-analytical methods to few\nstudies, as either coverage probabilities might be compromised, or intervals\nare inconclusively wide. Bayesian estimation with a sensibly chosen prior for\nbetween-trial heterogeneity may offer a promising compromise.\n", - "method": "We consider several likelihood-based inference methods.", - "id": "1807.09037" - }, - { - "abstract": " We study functions defined on the vertices of the Hamming graphs $H(n,q)$.\nThe adjacency matrix of $H(n,q)$ has $n+1$ distinct eigenvalues $n(q-1)-q\\cdot\ni$ with corresponding eigenspaces $U_{i}(n,q)$ for $0\\leq i\\leq n$. In this\nwork, we consider the problem of finding the minimum possible support (the\nnumber of nonzeros) of functions belonging to a direct sum $U_i(n,q)\\oplus\nU_{i+1}(n,q)\\oplus\\ldots\\oplus U_j(n,q)$ for $0\\leq i\\leq j\\leq n$. For the\ncase $n\\geq i+j$ and $q\\geq 3$ we find the minimum cardinality of the support\nof such functions and obtain a characterization of functions with the minimum\ncardinality of the support. In the case $n \\frac{n}{2}$,$\\,q\\ge 5$.\n", - "method": "We study functions defined on the vertices of the Hamming graphs $H(n,q)$.", - "id": "1807.09139" - }, - { - "abstract": " The nested Kingman coalescent describes the dynamics of particles (called\ngenes) contained in larger components (called species), where pairs of species\ncoalesce at constant rate and pairs of genes coalesce at constant rate provided\nthey lie within the same species. We prove that starting from $rn$ species, the\nempirical distribution of species masses (numbers of genes$/n$) at time $t/n$\nconverges as $n\\to\\infty$ to a solution of the deterministic\ncoagulation-transport equation $$\n \\partial_t d \\ = \\ \\partial_x ( \\psi d ) \\ + \\ a(t)\\left(d\\star d - d\n\\right), $$ where $\\psi(x) = cx^2$, $\\star$ denotes convolution and $a(t)=\n1/(t+\\delta)$ with $\\delta=2/r$. The most interesting case when $\\delta =0$\ncorresponds to an infinite initial number of species. This equation describes\nthe evolution of the distribution of species of mass $x$, where pairs of\nspecies can coalesce and each species' mass evolves like $\\dot x = -\\psi(x)$.\nWe provide two natural probabilistic solutions of the latter IPDE and address\nin detail the case when $\\delta=0$. The first solution is expressed in terms of\na branching particle system where particles carry masses behaving as\nindependent continuous-state branching processes. The second one is the law of\nthe solution to the following McKean-Vlasov equation $$ dx_t \\ = \\ - \\psi(x_t)\n\\,dt \\ + \\ v_t\\,\\Delta J_t $$ where $J$ is an inhomogeneous Poisson process\nwith rate $1/(t+\\delta)$ and $(v_t; t\\geq0)$ is a sequence of independent rvs\nsuch that $ {\\mathcal L}(v_t) = {\\mathcal L}(x_t)$. We show that there is a\nunique solution to this equation and we construct this solution with the help\nof a marked Brownian coalescent point process. When $\\psi(x)=x^\\gamma$, we show\nthe existence of a self-similar solution for the PDE which relates when\n$\\gamma=2$ to the speed of coming down from infinity of the nested Kingman\ncoalescent.\n", - "method": "We prove that starting from $rn$ species, the\nempirical distribution of species masses (numbers of genes$/n$) at time $t/n$\nconverges as $n\\to\\infty$ to a solution of the deterministic\ncoagulation-transport equation $$\n \\partial_t d \\ = \\ \\partial_x ( \\psi d ) \\ + \\ a(t)\\left(d\\star d - d\n\\right), $$ where $\\psi(x) = cx^2$, $\\star$ denotes convolution and $a(t)=\n1/(t+\\delta)$ with $\\delta=2/r$.", - "id": "1807.09153" - }, - { - "abstract": " CSIndexbr is a web-based system that provides meaningful,open,and transparent\ndata about Brazilian scientific production in Computer Science. Currently, the\nsystem collects full research papers published in the main track of selected\nconferences. The papers are retrieved from DBLP. In this article, we describe\nthe main features and resources provided by CSIndexbr. We also comment on how\nother researchers can use the data provided by the system to analyze the\nBrazilian production in Computer Science.\n", - "method": "The papers are retrieved from DBLP.", - "id": "1807.09266" - }, - { - "abstract": " We consider the information extraction framework known as document spanners,\nand study the problem of efficiently computing the results of the extraction\nfrom an input document, where the extraction task is described as a sequential\nvariable-set automaton (VA). We pose this problem in the setting of enumeration\nalgorithms, where we can first run a preprocessing phase and must then produce\nthe results with a small delay between any two consecutive results. Our goal is\nto have an algorithm which is tractable in combined complexity, i.e., in the\nsizes of the input document and the VA; while ensuring the best possible data\ncomplexity bounds in the input document size, i.e., constant delay in the\ndocument size. Several recent works at PODS'18 proposed such algorithms but\nwith linear delay in the document size or with an exponential dependency in\nsize of the (generally nondeterministic) input VA. In particular, Florenzano et\nal. suggest that our desired runtime guarantees cannot be met for general\nsequential VAs. We refute this and show that, given a nondeterministic\nsequential VA and an input document, we can enumerate the mappings of the VA on\nthe document with the following bounds: the preprocessing is linear in the\ndocument size and polynomial in the size of the VA, and the delay is\nindependent of the document and polynomial in the size of the VA. The resulting\nalgorithm thus achieves tractability in combined complexity and the best\npossible data complexity bounds. Moreover, it is rather easy to describe, in\nparticular for the restricted case of so-called extended VAs.\n", - "method": "Several recent works at PODS'18 proposed such algorithms but\nwith linear delay in the document size or with an exponential dependency in\nsize of the (generally nondeterministic) input VA.", - "id": "1807.09320" - }, - { - "abstract": " The Dark Energy Spectroscopic Instrument (DESI) is under construction to\nmeasure the expansion history of the Universe using the Baryon Acoustic\nOscillation technique. The spectra of 35 million galaxies and quasars over\n14000 sq deg will be measured during the life of the experiment. A new prime\nfocus corrector for the KPNO Mayall telescope will deliver light to 5000 fiber\noptic positioners. The fibers in turn feed ten broad-band spectrographs. To\nachieve this goal, it is crucial to guarantee that fiber positioners work\nproperly under the extremes of potential operating conditions, including the\nfull range of temperatures, high speed wind disturbance etc. Thermal testing\nprovides valuable insight into the functionality of the fiber positioners that\ncan be used to help mitigate poor performance at extreme temperatures and wind\ndisturbance test provide guidance to design of ventilation system. Here, we\ndescribe the thermal and wind disturbance tests for DESI fiber positioners and\nhow the test results helped improve the robustness of the positioners.\n", - "method": "The fibers in turn feed ten broad-band spectrographs.", - "id": "1807.09383" - }, - { - "abstract": " We study Cooper pairing in the Dirac composite fermion (CF) system. The\npresence of the mass term in the Dirac CF description (which may simulate\nLandau level mixing) i.e. breaking of particle-hole (PH) symmetry in this\nsystem, is a necessary condition for the existence of a PH Pfaffian-like\ntopological state. In the scope of RPA approximation and hydrodynamic approach,\nwe find some signatures of pairing at finite frequencies. Motivated by this\ninsight, we extend our analysis to the case of a different but still Dirac\nquasi-particle (CF) representation, appropriate in the presence of a mass term,\nand discuss the likelihood of PH Pfaffian pairing and Pfaffian pairing in\ngeneral. On the basis of gauge field effects, we find for small Dirac mass\nanti-Pfaffian or Pfaffian instability depending on the sign of mass, while for\nlarge mass (Landau level mixing), irrespective of its sign, PH Pfaffian-like\ninstability.\n", - "method": "We study Cooper pairing in the Dirac composite fermion (CF) system.", - "id": "1807.09428" - }, - { - "abstract": " This paper presents a user modeling pipeline to analyze discussions and\nopinions shared on social media regarding polarized political events (e.g.,\npublic polls). The pipeline follows a four-step methodology. First, social\nmedia posts and users metadata are crawled. Second, a filtering mechanism is\napplied to filter spammers and bot users. As a third step, demographics\ninformation is extracted out of the valid users, namely gender, age, ethnicity\nand location information. Finally, the political polarity of the users with\nrespect to the analyzed event is predicted. In the scope of this work, our\nproposed pipeline is applied to two referendum scenarios (independence of\nCatalonia in Spain and autonomy of Lombardy in Italy) in order to assess the\nperformance of the approach with respect to the capability of collecting\ncorrect insights on the demographics of social media users and of predicting\nthe poll results based on the opinions shared by the users. Experiments show\nthat the method was effective in predicting the political trends for the\nCatalonia case, but not for the Lombardy case. Among the various motivations\nfor this, we noticed that in general Twitter was more representative of the\nusers opposing the referendum than the ones in favor.\n", - "method": "The pipeline follows a four-step methodology.", - "id": "1807.09459" - }, - { - "abstract": " We describe some highlights in the theory of chaos, that started with\nPoincare (1899). Generic systems have both ordered and chaotic domains. Chaos\nappears mainly near un- stable periodic orbits. Large chaotic domains are due\nto resonance overlap. Two recent developments of the theory of chaos, refer (a)\nto the analytical formulae (Moser series) in the chaotic domains near unstable\nperiodic orbits, and (b) quantum chaos, described by the Bohmian theory of\norbits.\n", - "method": "Generic systems have both ordered and chaotic domains.", - "id": "1807.09492" - }, - { - "abstract": " The improvement on the Imaging Air Cherenkov Technique (IACT) led to the\ndiscovery of a new type of sources that can emit at very high energies: the\ngamma-ray binaries. Only six systems are part of this exclusive class. We\nsummarize the latest results from the observations performed with the MAGIC\ntelescopes on different systems as the gamma-ray binary LS I +61$^{\\circ}$ 303\nand the microquasars SS433, V404 Cygni and Cygnus X-1, which are considered\npotential VHE gamma-ray emitters. The binary system LS I +61$^{\\circ}$ 303 has\nbeen observed by MAGIC in a long-term monitoring campaign. We show the newest\nresults of our search for super-orbital variability also in context of\ncontemporaneous optical observations. Besides, we will present the results of\nthe only super-critical accretor known in our galaxy: SS 433. We will introduce\nthe VHE results achieved with MAGIC after 100 hours of observations on the\nmicroquasar Cygnus X-1 and report on the microquasar V404 Cyg, which has been\nobserved with MAGIC after it went through a series of exceptional X-ray\noutbursts in June 2015.\n", - "method": "Besides, we will present the results of\nthe only super-critical accretor known in our galaxy: SS 433.", - "id": "1807.09509" - }, - { - "abstract": " Both coherence and entanglement stem from the superposition principle,\ncapture quantumness of a physical system, and play a central role in quantum\nphysics. In a multipartite quantum system, coherence and quantum correlations\nare closely connected. In particular, it has been established that quantum\ncoherence of a bipartite state is an important resource for its conversion to\nentanglement [A. Streltsov {\\it et al.}, Phys. Rev. Lett. {\\bf 115}, 020403\n(2015)] and to quantum discord [J. Ma {\\it et al}., Phys. Rev. Lett. {\\bf 116},\n160407 (2016)]. We show here that there is a very close association between\npartial coherence introduced by Luo and Sun [S. Luo and Y. Sun, Phys. Rev. A\n{\\bf 96}, 022136 (2017)] and quantum correlations (quantified by quantum\ndiscord) in both directions. Furthermore, we propose families of coherence\nmeasures in terms of quantum correlations and quantum Fisher information.\n", - "method": "Luo and Y.. Sun, Phys.. Rev.", - "id": "1807.09565" - }, - { - "abstract": " (shorter version)The aim of this work is to search for planets around\nintermediate-mass stars in open clusters by using RV data obtained with HARPS\nfrom an extensive survey with more than 15 years of observations for a sample\nof 142 giant stars in 17 open clusters. We present the discovery of a periodic\nRV signal compatible with the presence of a planet candidate in the 1.15 Gyr\nopen cluster IC4651 orbiting the 2.06 M$_\\odot$ star No. 9122. If confirmed,\nthe planet candidate would have a minimum mass of 7.2 M$_{J}$ and a period of\n747 days. However, we also find that the FWHM of the CCF varies with a period\nclose to the RV, casting doubts on the planetary nature of the signal. We also\nprovide refined parameters for the previously discovered planet around NGC2423\nNo. 3 but show evidence that the BIS of the CCF is correlated with the RV\nduring some of the observing periods. This fact advises us that this might not\nbe a real planet and that the RV variations could be caused by stellar activity\nand/or pulsations. Finally, we show that the previously reported signal by a\nbrown dwarf around NGC4349 No. 127 is presumably produced by stellar activity\nmodulation. The long-term monitoring of several red giants in open clusters has\nallowed us to find periodic RV variations in several stars. However, we also\nshow that the follow-up of this kind of stars should last more than one orbital\nperiod to detect long-term signals of stellar origin. This work warns that\nalthough it is possible to detect planets around red giants, large-amplitude,\nlong-period RV modulations do exist in such stars that can mimic the presence\nof an orbiting planetary body. Therefore, we need to better understand how such\nRV modulations behave as stars evolve along the RGB and perform a detailed\nstudy of all the possible stellar-induced signals (e.g. spots, pulsations,\ngranulation) to comprehend the origin of RV variations.\n", - "method": "We present the discovery of a periodic\nRV signal compatible with the presence of a planet candidate in the 1.15 Gyr\nopen cluster IC4651 orbiting the 2.06 M$_\\odot$ star No. 9122.", - "id": "1807.09608" - }, - { - "abstract": " It is often difficult for a phased mission system (PMS) to be highly\nreliable, because this entails achieving high reliability in every phase of\noperation. Consequently, reliability analysis of such systems is of critical\nimportance. However, efficient and interpretable analysis of PMSs enabling\ngeneral component lifetime distributions, arbitrary structures, and the\npossibility that components skip phases has been an open problem.\n In this paper, we show that the survival signature can be used for\nreliability analysis of PMSs with similar types of component in each phase,\nproviding an alternative to the existing limited approaches in the literature.\nWe then develop new methodology addressing the full range of challenges above.\nThe new method retains the attractive survival signature property of separating\nthe system structure from the component lifetime distributions, simplifying\ncomputation, insight into, and inference for system reliability.\n", - "method": "We then develop new methodology addressing the full range of challenges above.", - "id": "1807.09616" - }, - { - "abstract": " The entanglement entropy of free fermions with a Fermi surface is known to\nobey a logarithmic scaling and violate the area law in all dimensions. Here, we\nwould like to see how temperature affects the logarithmic scaling behavior. To\nthis end, we compute the entanglement negativity of free fermions using the\nfermionic partial transpose developed in our earlier paper [Phys. Rev. B 95,\n165101 (2017)]. In one dimension, we analytically derive the leading order term\nin the finite-temperature entanglement negativity and show how the entanglement\nnegativity indicates a crossover from a quantum entangled state to a classical\nthermal state, where the entanglement is completely lost. We explain how the\none-dimensional result can be generalized to codimension-one Fermi surface of\narbitrary shape in higher dimensions. In both one and two dimensions, we check\nthat our analytical results agree with the numerical simulation of free\nfermions on a lattice.\n", - "method": "Rev.", - "id": "1807.09808" - }, - { - "abstract": " Contextual multi-armed bandit problems arise frequently in important\nindustrial applications. Existing solutions model the context either linearly,\nwhich enables uncertainty driven (principled) exploration, or non-linearly, by\nusing epsilon-greedy exploration policies. Here we present a deep learning\nframework for contextual multi-armed bandits that is both non-linear and\nenables principled exploration at the same time. We tackle the exploration vs.\nexploitation trade-off through Thompson sampling by exploiting the connection\nbetween inference time dropout and sampling from the posterior over the weights\nof a Bayesian neural network. In order to adjust the level of exploration\nautomatically as more data is made available to the model, the dropout rate is\nlearned rather than considered a hyperparameter. We demonstrate that our\napproach substantially reduces regret on two tasks (the UCI Mushroom task and\nthe Casino Parity task) when compared to 1) non-contextual bandits, 2)\nepsilon-greedy deep contextual bandits, and 3) fixed dropout rate deep\ncontextual bandits. Our approach is currently being applied to marketing\noptimization problems at HubSpot.\n", - "method": "We tackle the exploration vs.", - "id": "1807.09809" - }, - { - "abstract": " In this paper we make a systematic study of the multiplicity of the jumping\npoints associated to the mixed multiplier ideals of a family of ideals in a\ncomplex surface with rational singularities. In particular we study the\nbehaviour of the multiplicity by small perturbations of the jumping points. We\nalso introduce a Poincar\\'e series for mixed multiplier ideals and prove its\nrationality. Finally, we study the set of divisors that contribute to the\nlog-canonical wall.\n", - "method": "In particular we study the\nbehaviour of the multiplicity by small perturbations of the jumping points.", - "id": "1807.09839" - }, - { - "abstract": " We derive for the first time in the literature a rate of convergence in the\nhydrodynamic limit of the Kawasaki dynamics for a one-dimensional lattice\nsystem. We use an adaptation of the two-scale approach. The main difference to\nthe original two-scale approach is that the observables on the mesoscopic level\nare described by a projection onto splines of second order, and not by a\nprojection onto piecewise constant functions. This allows us to use a more\nnatural definition of the mesoscopic dynamics, which yields a better rate of\nconvergence than the original two-scale approach.\n", - "method": "We use an adaptation of the two-scale approach.", - "id": "1807.09850" - }, - { - "abstract": " Visual embeddings from Convolutional Neural Networks (CNN) trained on the\nImageNet dataset for the ILSVRC challenge have shown consistently good\nperformance for transfer learning and are widely used in several tasks,\nincluding image recommendation. However, some important questions have not yet\nbeen answered in order to use these embeddings for a larger scope of\nrecommendation domains: a) Do CNNs that perform better in ImageNet are also\nbetter for transfer learning in content-based image recommendation?, b) Does\nfine-tuning help to improve performance? and c) Which is the best way to\nperform the fine-tuning?\n In this paper we compare several CNN models pre-trained with ImageNet to\nevaluate their transfer learning performance to an artwork image recommendation\ntask. Our results indicate that models with better performance in the ImageNet\nchallenge do not always imply better transfer learning for recommendation tasks\n(e.g. NASNet vs. ResNet). Our results also show that fine-tuning can be helpful\neven with a small dataset, but not every fine-tuning works. Our results can\ninform other researchers and practitioners on how to train their CNNs for\nbetter transfer learning towards image recommendation systems.\n", - "method": "Our results indicate that models with better performance in the ImageNet\nchallenge do not always imply better transfer learning for recommendation tasks\n(e.g. NASNet vs.. ResNet).", - "id": "1807.09870" - }, - { - "abstract": " Platinum ditelluride has recently been characterized, based on angle-resolved\nphotoemission spectroscopy data and electronic band structure calculations, as\na possible representative of type-II Dirac semimetals. Here, we report on the\nmagnetotransport behavior (electrical resistivity, Hall effect) in this\ncompound, investigated on high-quality single-crystalline specimens. The\nmagnetoresistance (MR) of PtTe$_2$ is large (over $3000\\%$ at $T=1.8$ K in\n$B=9$ T) and unsaturated in strong fields in the entire temperature range\nstudied. The MR isotherms obey a Kohler's type scaling with the exponent $m$ =\n1.69, different from the case of ideal electron-hole compensation. In applied\nmagnetic fields, the resistivity shows a low-temperature plateau,\ncharacteristic of topological semimetals. In strong fields, well-resolved\nShubnikov - de Haas (SdH) oscillations with two principle frequencies were\nfound, and their analysis yielded charge mobilities of the order of\n$10^3\\,\\rm{cm^2V^{-1}s^{-1}}$ and rather small effective masses of charge\ncarriers, $0.11m_e$ and $0.21m_e$. However, the extracted Berry phases point to\ntrivial character of the electronic bands involved in the SdH oscillations. The\nHall effect data corroborated a multi-band character of the electrical\nconductivity in PtTe$_2$, with moderate charge compensation.\n", - "method": "Here, we report on the\nmagnetotransport behavior (electrical resistivity, Hall effect) in this\ncompound, investigated on high-quality single-crystalline specimens.", - "id": "1807.09876" - }, - { - "abstract": " In this article we study a mixed finite element formulation for solving the\nStokes problem with general surface forces that induce a jump of the normal\ntrace of the stress tensor, on an interface that splits the domain into two\nsubdomains. Equality of velocities is assumed at the interface. The interface\nconditions are taken into account with multipliers. A suitable Lagrangian\nfunctional is introduced for deriving a weak formulation of the problem. A\nspecificity of this work is the consideration of the interface with a\nfictitious domain approach. The latter is inspired by the XFEM approach in the\nsense that cut-off functions are used, but it is simpler to implement since no\nenrichment by singular functions is provided. In that context, getting\nconvergence for the dual variables defined on the interface is non-trivial. For\nthat purpose, an augmented Lagrangian technique stabilizes the convergence of\nthe multipliers, which is important because their value would determine the\ndynamics of the interface in an unsteady framework. Theoretical analysis is\nprovided, where we show that a discrete inf-sup condition, independent of the\nmesh size, is satisfied for the stabilized formulation. This guarantees optimal\nconvergence rates, that we observe with numerical tests. The capacity of the\nmethod is demonstrated with robustness tests, and with an unsteady model tested\nfor deformations of the interface that correspond to ellipsoidal shapes in\ndimension 2.\n", - "method": "Equality of velocities is assumed at the interface.", - "id": "1807.09894" - }, - { - "abstract": " A point on a smooth cubic is called a flex (or an inflection point) if the\ntangent line at that point intersects with the cubic with multiplicity 3. We\nprove the flexes on a smooth cubic over a global field satisfy the Hasse\nprinciple. Moreover, we study the Hasse principle allowing exceptional sets of\npositive density, following a suggestion by J.-P. Serre. We prove a smooth\ncubic over a global field has a flex over the base field if and only if it has\na flex locally at a set of places of Dirichlet density strictly larger than\n8/9.\n", - "method": "Serre.", - "id": "1807.09909" - }, - { - "abstract": " We present a time-dependent formulation of coupled cluster theory. This\ntheory allows for direct computation of the free energy of quantum systems at\nfinite temperature by imaginary time integration and is closely related to the\nthermal cluster cumulant theory of Mukherjee and co-workers. Our derivation\nhighlights the connection to perturbation theory and zero-temperature coupled\ncluster theory. We show explicitly how the finite-temperature coupled cluster\nsingles and doubles amplitude equations can be derived in analogy with the\nzero-temperature theory and how response properties can be efficiently computed\nusing a variational Lagrangian. We discuss the implementation for realistic\nsystems and showcase the potential utility of the method with calculations of\nthe exchange correlation energy of the uniform electron gas at warm dense\nmatter conditions.\n", - "method": "We present a time-dependent formulation of coupled cluster theory.", - "id": "1807.09961" - }, - { - "abstract": " In this paper, we derive in a novel approach the possible textures of\nneutrino mass matrix that can lead us to maximal atmospheric mixing angle and\nDirac CP phase which are consistent with the current neutrino oscillation data.\nA total of eleven textures are thus found. Interestingly, the specific texture\ngiven by the $\\mu$-$\\tau$ reflection symmetry can be reproduced from one of the\nobtained textures. For these textures, some neutrino mass sum rules which\nrelate the neutrino masses and Majorana CP phases will emerge.\n", - "method": "A total of eleven textures are thus found.", - "id": "1807.10031" - }, - { - "abstract": " We introduce a notion of differential of a Sobolev map between metric spaces.\nThe differential is given in the framework of tangent and cotangent modules of\nmetric measure spaces, developed by the first author. We prove that our notion\nis consistent with Kirchheim's metric differential when the source is a\nEuclidean space, and with the abstract differential provided by the first\nauthor when the target is $\\mathbb{R}$.\n", - "method": "We introduce a notion of differential of a Sobolev map between metric spaces.", - "id": "1807.10063" - }, - { - "abstract": " Discovering whether words are semantically related and identifying the\nspecific semantic relation that holds between them is of crucial importance for\nNLP as it is essential for tasks like query expansion in IR. Within this\ncontext, different methodologies have been proposed that either exclusively\nfocus on a single lexical relation (e.g. hypernymy vs. random) or learn\nspecific classifiers capable of identifying multiple semantic relations (e.g.\nhypernymy vs. synonymy vs. random). In this paper, we propose another way to\nlook at the problem that relies on the multi-task learning paradigm. In\nparticular, we want to study whether the learning process of a given semantic\nrelation (e.g. hypernymy) can be improved by the concurrent learning of another\nsemantic relation (e.g. co-hyponymy). Within this context, we particularly\nexamine the benefits of semi-supervised learning where the training of a\nprediction function is performed over few labeled data jointly with many\nunlabeled ones. Preliminary results based on simple learning strategies and\nstate-of-the-art distributional feature representations show that concurrent\nlearning can lead to improvements in a vast majority of tested situations.\n", - "method": "synonymy vs.. random).", - "id": "1807.10076" - }, - { - "abstract": " We show the global well-posedness for the two-dimensional Zakharov-Kuznetsov\nequation in $H^{s}({\\mathbb{R}^2})$ when $\\frac{11}{13}30 MeV) fluence\nexceeding $5*10^{10}$ and $10^{11}$ cm$^2$ are expected on timescales of a\nthousand and million years, respectively. Conclusions. We conclude that the\nmean flux of solar energetic particles hardly depends on the level of solar\nactivity, in contrast to the solar modulation of galactic cosmic rays. This\nputs new observational constraints on solar physics and becomes important for\nassessing radiation hazards for the planned space missions.\n", - "method": "Methods.", - "id": "1807.10153" - }, - { - "abstract": " This paper concerns computation of optimal policies in which the one-step\nreward function contains a cost term that models Kullback-Leibler divergence\nwith respect to nominal dynamics. This technique was introduced by Todorov in\n2007, where it was shown under general conditions that the solution to the\naverage-reward optimality equations reduce to a simple eigenvector problem.\nSince then many authors have sought to apply this technique to control problems\nand models of bounded rationality in economics.\n A crucial assumption is that the input process is essentially unconstrained.\nFor example, if the nominal dynamics include randomness from nature (e.g., the\nimpact of wind on a moving vehicle), then the optimal control solution does not\nrespect the exogenous nature of this disturbance.\n This paper introduces a technique to solve a more general class of\naction-constrained MDPs. The main idea is to solve an entire parameterized\nfamily of MDPs, in which the parameter is a scalar weighting the one-step\nreward function. The approach is new and practical even in the original\nunconstrained formulation.\n", - "method": "A crucial assumption is that the input process is essentially unconstrained.", - "id": "1807.10244" - }, - { - "abstract": " \"Topological Quantum Chemistry (TQC) links the chemical and symmetry\nstructure of a given material with its topological properties. This field\ntabulates the data of the 10398 real-space atomic limits of materials, and\nsolves the compatibility relations of electronic bands in momentum space. A\nmaterial that is not an atomic limit or whose bands do not satisfy the\ncompatibility relations, is a topological insulator/semimetal. We use TQC to\nfind the topological stoichiometric non-magnetic, \"high-quality'' materials in\nthe world. We develop several code additions to VASP which can compute all\ncharacters of all symmetries at all high-symmetry points in the Brillouin Zone\n(BZ). Using TQC we then develop codes to check which materials in ICSD are\ntopological. Out of 26938 stoichiometric materials in our filtered ICSD\ndatabase, we find 2861 topological insulators (TI) and 2936 topological\nsemimetals (2505 and 2560 non-f electron, respectively). Our method is uniquely\ncapable to show that none of the TI's found exhibit fragile topology. We\npartition the topological materials in different physical classes. For the\nmajority of the 5797 \"high-quality'' topological material, we compute: the\ntopological class (equivalence classes of TQC elementary band representations\n-- equivalent to the topological index), the symmetry(ies) that protects the\ntopological class, the representations at high symmetry points and the direct\ngap (for insulators), and the topological index. For topological semimetals we\nthen compute whether the system becomes a topological insulator (whose\nindex/class we compute) upon breaking symmetries -- useful for experiments.\n2152 more TI's are obtained in this way. For almost all 5065 non-f-electron\ntopological materials, we provide the electronic band structures, allowing the\nidentification of quantitative properties (gaps, velocities). Remarkably, our\nexhaustive results show that a large proportion ( ~ 24% !) of all materials in\nnature are topological (confirmed by calculations of \"low-quality'' materials).\nWe confirm the topology of several new materials by Wilson loop calculations.\nWe added an open-source code and end-user button on the Bilbao Crystallographic\nServer (BCS) which checks the topology of any material. We comment on the\nchemistry of each compound and sample part of the \"low-quality'' ICSD data to\nfind more materials.\"\n", - "method": "We confirm the topology of several new materials by Wilson loop calculations.", - "id": "1807.10271" - }, - { - "abstract": " Noise level frequently set the basic limit on the smallest signal. New noise\nreduction technology and amplifiers voltage-noise density, yet still offer high\nspeed, high accuracy, and low power solution. Low noise amplifiers always play\na significant role in RF technology. Hence in this paper, high linear low noise\namplifier (LNA) using cascade self-biased multiple gated transistors (MGTR) is\npresented. The proposed system is covering 0.9 to 2.4 GHz applications. To\nverify the functionality of the proposed LNA as a bottleneck of RF technology,\na cascade LNA without MGTR is implemented and synthesized. The comparison has\nbeen done with the single-gate LNA. From the synthesized result, proposed LNA\nobtained 10 dBm third-order input intercept point (IIP3) in compare with\nsingle-gate LNA at 9dB gain. The proposed LNA is implemented in 90 nm CMOS\ntechnology and reported 13 dBm IIP3, 1.9 dB NF and 9 dB gain while consuming\n7.9 mW from 2 V supply.\n", - "method": "The proposed system is covering 0.9 to 2.4 GHz applications.", - "id": "1807.10306" - }, - { - "abstract": " We consider Heisenberg groups equipped with a sub-Finsler metric. Using\nmethods of optimal control theory we prove that in this geometric setting the\ninfinite geodesics are horizontal lines under the assumption that the\nsub-Finsler metric is defined by a strictly convex norm. This answers a\nquestion posed in [5] and has applications in the characterisation of isometric\nembeddings into Heisenberg groups.\n", - "method": "We consider Heisenberg groups equipped with a sub-Finsler metric.", - "id": "1807.10369" - }, - { - "abstract": " A recent paper by Ge et al. performs a series of experiments with two full\nspectral fitting codes, pPXF and starlight, finding that the two yield\nconsistent results when the input spectrum is not heavily reddened. For E(B-V)\n> 0.2, however, they claim starlight leads to severe biases in the derived\nproperties. Counterintuitively, and at odds with previous simulations, they\nfind that this behaviour worsens significantly as the signal-to-noise ratio of\nthe input spectrum increases. This communication shows that this is entirely\ndue to an A_V < 1 mag condition imposed while initializing the Markov chains in\nthe code. This choice is normally irrelevant in real-life galaxy work but can\nbecome critical in artificial experiments. Alleviating this usually harmless\ninitialization constraint changes the Ge et al. results completely, as was\nexplained to the authors before their publication. We replicate their spectral\nfitting experiments, finding much smaller biases. Furthermore both bias and\nscatter in the derived properties all converge as S/N increases, as one would\nexpect. We also show how the very output of the code provides ways of\ndiagnosing anomalies in the fits. The code behaviour has been documented in\ncareful and extensive experiments in the literature, but the biased analysis of\nGe et al. is just not representative of starlight at all.\n", - "method": "A recent paper by Ge et al.", - "id": "1807.10423" - }, - { - "abstract": " Factorial experiments in research on memory, language, and in other areas are\noften analyzed using analysis of variance (ANOVA). However, for effects with\nmore than one numerator degrees of freedom, e.g., for experimental factors with\nmore than two levels, the ANOVA omnibus F-test is not informative about the\nsource of a main effect or interaction. Because researchers typically have\nspecific hypotheses about which condition means differ from each other, a\npriori contrasts (i.e., comparisons planned before the sample means are known)\nbetween specific conditions or combinations of conditions are the appropriate\nway to represent such hypotheses in the statistical model. Many researchers\nhave pointed out that contrasts should be \"tested instead of, rather than as a\nsupplement to, the ordinary `omnibus' F test\" (Hays, 1973, p. 601). In this\ntutorial, we explain the mathematics underlying different kinds of contrasts\n(i.e., treatment, sum, repeated, polynomial, custom, nested, interaction\ncontrasts), discuss their properties, and demonstrate how they are applied in\nthe R System for Statistical Computing (R Core Team, 2018). In this context, we\nexplain the generalized inverse which is needed to compute the coefficients for\ncontrasts that test hypotheses that are not covered by the default set of\ncontrasts. A detailed understanding of contrast coding is crucial for\nsuccessful and correct specification in linear models (including linear mixed\nmodels). Contrasts defined a priori yield far more useful confirmatory tests of\nexperimental hypotheses than standard omnibus F-test. Reproducible code is\navailable from https://osf.io/7ukf6/.\n", - "method": "601).", - "id": "1807.10451" - }, - { - "abstract": " The approximate period recovery problem asks to compute all\n$\\textit{approximate word-periods}$ of a given word $S$ of length $n$: all\nprimitive words $P$ ($|P|=p$) which have a periodic extension at edit distance\nsmaller than $\\tau_p$ from $S$, where $\\tau_p = \\lfloor\n\\frac{n}{(3.75+\\epsilon)\\cdot p} \\rfloor$ for some $\\epsilon>0$. Here, the set\nof periodic extensions of $P$ consists of all finite prefixes of $P^\\infty$.\n We improve the time complexity of the fastest known algorithm for this\nproblem of Amir et al. [Theor. Comput. Sci., 2018] from $O(n^{4/3})$ to $O(n\n\\log n)$. Our tool is a fast algorithm for Approximate Pattern Matching in\nPeriodic Text. We consider only verification for the period recovery problem\nwhen the candidate approximate word-period $P$ is explicitly given up to cyclic\nrotation; the algorithm of Amir et al. reduces the general problem in $O(n)$\ntime to a logarithmic number of such more specific instances.\n", - "method": "[Theor.. Comput.", - "id": "1807.10483" - }, - { - "abstract": " Automatic grading is not a new approach but the need to adapt the latest\ntechnology to automatic grading has become very important. As the technology\nhas rapidly became more powerful on scoring exams and essays, especially from\nthe 1990s onwards, partially or wholly automated grading systems using\ncomputational methods have evolved and have become a major area of research. In\nparticular, the demand of scoring of natural language responses has created a\nneed for tools that can be applied to automatically grade these responses. In\nthis paper, we focus on the concept of automatic grading of short answer\nquestions such as are typical in the UK GCSE system, and providing useful\nfeedback on their answers to students. We present experimental results on a\ndataset provided from the introductory computer science class in the University\nof North Texas. We first apply standard data mining techniques to the corpus of\nstudent answers for the purpose of measuring similarity between the student\nanswers and the model answer. This is based on the number of common words. We\nthen evaluate the relation between these similarities and marks awarded by\nscorers. We then consider an approach that groups student answers into\nclusters. Each cluster would be awarded the same mark, and the same feedback\ngiven to each answer in a cluster. In this manner, we demonstrate that clusters\nindicate the groups of students who are awarded the same or the similar scores.\nWords in each cluster are compared to show that clusters are constructed based\non how many and which words of the model answer have been used. The main\nnovelty in this paper is that we design a model to predict marks based on the\nsimilarities between the student answers and the model answer.\n", - "method": "This is based on the number of common words.", - "id": "1807.10543" - }, - { - "abstract": " Context. Classical novae are explosive phenomena that take place in stellar\nbinary systems. They are powered by mass transfer from a low-mass main sequence\nstar onto a white dwarf (either CO or ONe). The material accumulates for 10+4 -\n10+5 yr until ignition under degenerate conditions, resulting in a\nthermonuclear runaway. The nuclear energy released produces peak temperatures\nof about 0.1 - 0.4 GK. During these events, 10-7 - 10-3 Msun enriched in\nintermediate-mass elements (with respect to solar abundances) are ejected into\nthe interstellar medium. However, the origin of the large metallicity\nenhancements and the inhomogeneous distribution of chemical species observed in\nhigh-resolution spectra of ejected nova shells is not fully understood. Aims.\nRecent multidimensional simulations have demonstrated that Kelvin-Helmholtz\ninstabilities that operate at the core-envelope interface can naturally produce\nself-enrichment of the accreted envelope with material from the underlying\nwhite dwarf at levels that agree with observations. However, such\nmultidimensional simulations have been performed for a small number of cases,\nand much of the parameter space remains unexplored. Methods. Here we\ninvestigate the dredge-up, driven by Kelvin-Helmholtz instabilities, for white\ndwarf masses in the range 0.8-1.25 Msun and different core compositions\n(CO-rich and ONe-rich substrates). We present a set of five numerical\nsimulations performed in two dimensions aimed at analyzing the possible impact\nof the white dwarf mass (and composition) on the metallicity enhancement and on\nthe explosion characteristics. Results. We observe greater mixing (about 30 per\ncent higher when measured in the same conditions), at the time we stop the\nsimulations, and more energetic outbursts for ONe-rich substrates than for\nCO-rich substrates and for more massive white dwarfs.\n", - "method": "Methods.", - "id": "1807.10646" - }, - { - "abstract": " The fast neutrons produced by a calibrated 241Am-Be source were detected by\ntwo different Cs2LiYCl6:Ce (CLYC) scintillator detectors. The two cylindrical\ncrystals (1\"x1\" in size) were enriched with more than 99% of 7Li (C7LYC) and\nwith about 95% of 6Li (C6LYC), respectively. Both crystals can detect fast\nneutrons whereas only C6LYC can also detect thermal neutrons, due to the\npresence of 6Li. The measurement was performed at the L.A.S.A. Laboratory of\nINFN and University of Milano (Italy). To identify the neutron events, the\nPulse-Shape-Discrimination technique was used. A value for the detection\nefficiency of the 241Am-Be emitted neutrons, with energy up to 10 MeV, was\ndeduced.\n", - "method": "Both crystals can detect fast\nneutrons whereas only C6LYC can also detect thermal neutrons, due to the\npresence of 6Li. The measurement was performed at the L.A.S.A. Laboratory of\nINFN and University of Milano (Italy).. The fast neutrons produced by a calibrated 241Am-Be source were detected by\ntwo different Cs2LiYCl6:Ce (CLYC) scintillator detectors.. Both crystals can detect fast\nneutrons whereas only C6LYC can also detect thermal neutrons, due to the\npresence of 6Li. The measurement was performed at the L.A.S.A. Laboratory of\nINFN and University of Milano (Italy).. The fast neutrons produced by a calibrated 241Am-Be source were detected by\ntwo different Cs2LiYCl6:Ce (CLYC) scintillator detectors.", - "id": "1807.10759" - }, - { - "abstract": " We consider the Haar functions $h_I$ on dyadic intervals. We show that if\n$p>\\frac23$ and $E\\subset[0,1]$ then the set of all functions\n$\\|h_I1_E\\|_2^{-1}h_I1_E$ with $|I\\cap E|\\geq p|I|$ is a Riesz sequence. For\n$p\\leq\\frac23$ we provide a counterexample.\n", - "method": "We consider the Haar functions $h_I$ on dyadic intervals.", - "id": "1807.10809" - }, - { - "abstract": " With the imminent construction of an electron-ion collider in the USA, both\nsaturation physics and nuclear shadowing physics will be important for the same\nprocesses for the first time. In particular for exclusive production of vector\nmesons, such as the J/$\\psi$. In this paper we investigate the underlying\nassumptions commonly made when phenomenologically investigating this process\nfrom a shadowing and saturation perspective respectively. We use the bSat model\nwhich is commonly used to describe saturation physics, and a calculation by\nM.G. Ryskin which is commonly used when describing nuclear shadowing at leading\ntwist. It is expected that the bSat model is equivalent to Ryskin's result in\nthe hard scattering and non-relativistic limits. By explicitly taking these\nlimits we show that the bSat model does indeed become equivalent to Ryskin's\nresult. However, two more approximations are needed for this result. Firstly,\nthe factorisation and renormalisation scales in the bSat model has to be\nindependent of the dipole radius, and secondly, we need to omit a term in the\noverlap between the vector meson and virtual photon wave functions. We show\nthat for the typical dipole radius of the J/$\\psi$, the different\napproximations off-set each other and the bSat model agrees very well with\nRyskin's calculation in hard-scattering and non-relativistic limits.\n", - "method": "We use the bSat model\nwhich is commonly used to describe saturation physics, and a calculation by\nM.G. Ryskin which is commonly used when describing nuclear shadowing at leading\ntwist.", - "id": "1807.10888" - }, - { - "abstract": " We prove the existence of Ulrich bundles on any Brauer--Severi variety. In\nsome cases, the minimal possible rank of the obtained Ulrich bundles equals the\nperiod of the Brauer--Severi variety. Moreover, we find a formula for the rank\nof an Ulrich bundle involving the period of the considered Brauer--Severi\nvariety $X$, at least if $\\mathrm{dim}(X)=p-1$ for an odd prime $p$. This\nformula implies that the rank of any Ulrich bundle on such a Brauer--Severi\nvariety $X$ must be a multiple of the period.\n", - "method": "We prove the existence of Ulrich bundles on any Brauer--Severi variety.", - "id": "1807.10919" - }, - { - "abstract": " Accurate synthesis of a full 3D MR image containing tumours from available\nMRI (e.g. to replace an image that is currently unavailable or corrupted) would\nprovide a clinician as well as downstream inference methods with important\ncomplementary information for disease analysis. In this paper, we present an\nend-to-end 3D convolution neural network that takes a set of acquired MR image\nsequences (e.g. T1, T2, T1ce) as input and concurrently performs (1) regression\nof the missing full resolution 3D MRI (e.g. FLAIR) and (2) segmentation of the\ntumour into subtypes (e.g. enhancement, core). The hypothesis is that this\nwould focus the network to perform accurate synthesis in the area of the\ntumour. Experiments on the BraTS 2015 and 2017 datasets [1] show that: (1) the\nproposed method gives better performance than state-of-the-art methods in terms\nof established global evaluation metrics (e.g. PSNR), (2) replacing real MR\nvolumes with the synthesized MRI does not lead to significant degradation in\ntumour and sub-structure segmentation accuracy. The system further provides\nuncertainty estimates based on Monte Carlo (MC) dropout [11] for the\nsynthesized volume at each voxel, permitting quantification of the system's\nconfidence in the output at each location.\n", - "method": "In this paper, we present an\nend-to-end 3D convolution neural network that takes a set of acquired MR image\nsequences (e.g. T1, T2, T1ce) as input and concurrently performs (1) regression\nof the missing full resolution 3D MRI (e.g. FLAIR) and (2) segmentation of the\ntumour into subtypes (e.g. enhancement, core).", - "id": "1807.10972" - }, - { - "abstract": " Recently using entanglement as resource to distinguish orthogonal product\nstates by local operations and classical communication (LOCC) has been studied\nintensively. Zhang. et al. presented protocols to use entanglement to\ndistinguish certain classes of orthogonal product states in\n$\\mathbb{C}^m\\otimes \\mathbb{C}^n$\\cite{Zhang016}. In this paper, we study\nlocal distinguishability of multipartite orthogonal product states and provide\na practical solution. Our method relies upon a special class of locally\nindistinguishable multipartite product states introduced by Wang et. al. to\nbuild a protocol to distinguishes perfectly multipartitie quantum states by\nLOCC using an entangled state as a resource for implementing quantum\nmeasurements.\n", - "method": "et al.", - "id": "1807.10996" - }, - { - "abstract": " We consider collision of particles in a wormhole near its throat. Particles\ncome from the opposite mouths. If the lapse function is small enough there, the\nenergy $E$ of debris at infinity grows unbounded, so we are faced with the\nso-called super-Penrose process. This requires the existence of the ergoregion,\nso a wormhole should be rotating.\n", - "method": "We consider collision of particles in a wormhole near its throat.", - "id": "1807.11033" - }, - { - "abstract": " Let $f$ be a polynomial system consisting of $n$ polynomials $f_1,\\cdots,\nf_n$ in $n$ variables $x_1,\\cdots, x_n$, with coefficients in $\\mathbb{Q}$ and\nlet $\\langle f\\rangle$ be the ideal generated by $f$. Such a polynomial system,\nwhich has as many equations as variables is called a square system. It may be\nzero-dimensional, i.e the system of equations $f = 0$ has finitely many complex\nsolutions, or equivalently the dimension of the quotient algebra $A =\n\\mathbb{Q}[x]/\\langle f\\rangle$ is finite. In this case, the companion matrices\nof $f$ are defined as the matrices of the endomorphisms of $A$, called\nmultiplication maps, $x_j : \\left\\vert \\begin{array}{c} h \\mapsto x_jh\n\\end{array} \\right.$, written in some basis of $A$. We present a practical and\nefficient method to compute the companion matrices of $f$ in the case when the\nsystem is zero-dimensional. When it is not zero-dimensional, then the method\nworks as well and still produces matrices having properties similar to the\nzero-dimensional case. The whole method consists in matrix calculations. An\nexperiment illustrates the method's effectiveness.\n", - "method": "Such a polynomial system,\nwhich has as many equations as variables is called a square system.", - "id": "1807.11088" - }, - { - "abstract": " In this paper, we will show that the sheaf of p-adic vanishing cycles of a\nlog smooth family over a DVR of mixed characteristic is generated by Milnor\nsymboles. A key ingredient is a computation (due to K. Kato) on the graded\nquotients of a multi-indexed filtration on the sheaf concerned, which has been\nused in several papers of the first author.\n", - "method": "A key ingredient is a computation (due to K.", - "id": "1807.11128" - }, - { - "abstract": " We construct Andrews-Gordon type evidently positive series as generating\nfunctions for the partitions satisfying the difference conditions imposed by\nCapparelli's identities and G\\\"{o}llnitz-Gordon identities. The construction\ninvolves building base partitions, and move parts around subject to certain\nrules. Some missing cases are also considered, but no new identities are given.\n", - "method": "Some missing cases are also considered, but no new identities are given.", - "id": "1807.11189" - }, - { - "abstract": " We study the application of active learning techniques to the translation of\nunbounded data streams via interactive neural machine translation. The main\nidea is to select, from an unbounded stream of source sentences, those worth to\nbe supervised by a human agent. The user will interactively translate those\nsamples. Once validated, these data is useful for adapting the neural machine\ntranslation model.\n We propose two novel methods for selecting the samples to be validated. We\nexploit the information from the attention mechanism of a neural machine\ntranslation system. Our experiments show that the inclusion of active learning\ntechniques into this pipeline allows to reduce the effort required during the\nprocess, while increasing the quality of the translation system. Moreover, it\nenables to balance the human effort required for achieving a certain\ntranslation quality. Moreover, our neural system outperforms classical\napproaches by a large margin.\n", - "method": "We propose two novel methods for selecting the samples to be validated.", - "id": "1807.11243" - }, - { - "abstract": " Shape analysis and compuational anatomy both make use of sophisticated tools\nfrom infinite-dimensional differential manifolds and Riemannian geometry on\nspaces of functions. While comprehensive references for the mathematical\nfoundations exist, it is sometimes difficult to gain an overview how\ndifferential geometry and functional analysis interact in a given problem. This\npaper aims to provide a roadmap to the unitiated to the world of\ninfinite-dimensional Riemannian manifolds, spaces of mappings and Sobolev\nmetrics: all tools used in computational anatomy and shape analysis.\n", - "method": "This\npaper aims to provide a roadmap to the unitiated to the world of\ninfinite-dimensional Riemannian manifolds, spaces of mappings and Sobolev\nmetrics: all tools used in computational anatomy and shape analysis.", - "id": "1807.11290" - }, - { - "abstract": " We investigate the formation of cluster crystals with multiply occupied\nlattice sites on a spherical surface in systems of ultra-soft particles\ninteracting via repulsive, bounded pair potentials. Not all interactions of\nthis kind lead to clustering: we generalize the criterion devised in C.N. Likos\net al., Phys. Rev. E, 2001, 63, 031206 to spherical systems in order to\ndistinguish between cluster forming systems and fluids which display reentrant\nmelting. We use both DFT and Monte Carlo simulations to characterize the\nbehavior of the system, and obtain semi-quantitative agreement between the two.\nFurthermore, we study the effect of topological frustration on the system due\nto the sphere curvature by comparing the properties of disclinations, i.e.,\nclusters with fewer than six neighbors, and non-defective clusters.\nDisclinations are shown to be less stable, contain fewer particles, and be\ncloser to their neighbors than other lattice points: these properties are\nexplained on the basis of geometric and energetic considerations.\n", - "method": "Not all interactions of\nthis kind lead to clustering: we generalize the criterion devised in C.N. Likos\net al., Phys.. Rev.", - "id": "1807.11303" - }, - { - "abstract": " Although deep networks have recently emerged as the model of choice for many\ncomputer vision problems, in order to yield good results they often require\ntime-consuming architecture search. To combat the complexity of design choices,\nprior work has adopted the principle of modularized design which consists in\ndefining the network in terms of a composition of topologically identical or\nsimilar building blocks (a.k.a. modules). This reduces architecture search to\nthe problem of determining the number of modules to compose and how to connect\nsuch modules. Again, for reasons of design complexity and training cost,\nprevious approaches have relied on simple rules of connectivity, e.g.,\nconnecting each module to only the immediately preceding module or perhaps to\nall of the previous ones. Such simple connectivity rules are unlikely to yield\nthe optimal architecture for the given problem.\n In this work we remove these predefined choices and propose an algorithm to\nlearn the connections between modules in the network. Instead of being chosen a\npriori by the human designer, the connectivity is learned simultaneously with\nthe weights of the network by optimizing the loss function of the end task\nusing a modified version of gradient descent. We demonstrate our connectivity\nlearning method on the problem of multi-class image classification using two\npopular architectures: ResNet and ResNeXt. Experiments on four different\ndatasets show that connectivity learning using our approach yields consistently\nhigher accuracy compared to relying on traditional predefined rules of\nconnectivity. Furthermore, in certain settings it leads to significant savings\nin number of parameters.\n", - "method": "To combat the complexity of design choices,\nprior work has adopted the principle of modularized design which consists in\ndefining the network in terms of a composition of topologically identical or\nsimilar building blocks (a.k.a. modules).", - "id": "1807.11473" - }, - { - "abstract": " Unmanned aerial vehicles (UAVs) have gained a lot of popularity in diverse\nwireless communication fields. They can act as high-altitude flying relays to\nsupport communications between ground nodes due to their ability to provide\nline-of-sight links. With the flourishing Internet of Things, several types of\nnew applications are emerging. In this paper, we focus on bandwidth hungry and\ndelay-tolerant applications where multiple pairs of transceivers require the\nsupport of UAVs to complete their transmissions. To do so, the UAVs have the\npossibility to employ two different bands namely the typical microwave and the\nhigh-rate millimeter wave bands. In this paper, we develop a generic framework\nto assign UAVs to supported transceivers and optimize their trajectories such\nthat a weighted function of the total service time is minimized. Taking into\naccount both the communication time needed to relay the message and the flying\ntime of the UAVs, a mixed non-linear programming problem aiming at finding the\nstops at which the UAVs hover to forward the data to the receivers is\nformulated. An iterative approach is then developed to solve the problem.\nFirst, a mixed linear programming problem is optimally solved to determine the\npath of each available UAV. Then, a hierarchical iterative search is executed\nto enhance the UAV stops' locations and reduce the service time. The behavior\nof the UAVs and the benefits of the proposed framework are showcased for\nselected scenarios.\n", - "method": "An iterative approach is then developed to solve the problem.", - "id": "1807.11578" - }, - { - "abstract": " CONTEXT. 2MASS is the reference survey in the NIR part of the spectrum given\nits whole-sky coverage, large dynamic range, and proven calibration uniformity.\nHowever, previous studies disagree in the value of the zero points (ZPs) for\nits three bands JHK at the hundredth of a magnitude level. The disagreement\nshould become more noticeable now that Gaia provides whole-sky optical\nphotometry calibrated below that level. AIMS. We want to establish the value of\nthe 2MASS ZPs based on NICMOS/HST spectrophotometry of the CALSPEC standard\nstars and test it with the help of Gaia DR2 parallaxes. METHODS. We have\ncomputed the synthetic JHK photometry for a sample of stars using the HST\nCALSPEC spectroscopic standards and compared it with their 2MASS magnitudes to\nevaluate the ZPs. We have tested our results by analysing a sample of FGK\ndwarfs with excellent 2MASS photometry and accurate Gaia DR2 parallaxes.}\nRESULTS. The Vega ZPs for 2MASS J, H, and K are found to be -0.025$\\pm$0.005\nmag, 0.004$\\pm$0.005 mag, and -0.015$\\pm$0.005 mag, respectively. The analysis\nof the FGK sample indicates that the new ZPs are more accurate than previous\nones.\n", - "method": "METHODS.", - "id": "1807.11686" - }, - { - "abstract": " Many complex vacuum systems include edge-welded bellows. Their simulation in\nthe molecular flow regime with a Test Particle Monte Carlo (TPMC) code, such as\nMolFlow+, can take considerable amounts of computing power and time. Therefore\nwe investigated the change of the transmission probability of a bellow compared\nto a cylindrical tube in TPMC simulations. The results were used to develop an\nempirical model to simplify the geometry of a bellow in a TPMC model by\nreplacing it with a cylindrical tube of extended length, consequently\ncompensating the reduced conductance of the bellow. The simulated transmission\nprobabilities of a variety of different bellow lengths have been used to fit\nthe parameters of the model. Each bellow geometry simulated with MolFlow+\ncomprised two cylindrical tubes, one at each end, and the central bellow\nsection. The geometry of each bellow is described by two quantities. The first\none is the total length of the geometry model normalized to the inner diameter\nof the bellow. The second one is the fraction of the bellow section with regard\nto the total length. Simulating a tube instead of a long bellow reduced the\nsimulation time by a factor of up to 1000, while the error introduced through\nthe replacement of the bellow with a cylindrical tube of modified length was in\nmost cases negligible.\n", - "method": "The geometry of each bellow is described by two quantities.", - "id": "1807.11691" - }, - { - "abstract": " Let $G$ be a finite group and let $H$ be a proper subgroup of $G$ of minimal\nindex. By applying an old result of Y. Berkovich, we provide a polynomial\nalgorithm for computing $|G : H|$ for a permutation group $G$. Moreover, we\nfind $H$ explicitly if $G$ is given by a Cayley table. As a corollary, we get\nan algorithm for testing whether a finite permutation group acts on a tree or\nnot.\n", - "method": "By applying an old result of Y.", - "id": "1807.11773" - }, - { - "abstract": " This paper presents a study of causality in a reversible, concurrent setting.\nThere exist various notions of causality in pi-calculus, which differ in the\ntreatment of parallel extrusions of the same name. In this paper we present a\nuniform framework for reversible pi-calculi that is parametric with respect to\na data structure that stores information about an extrusion of a name.\nDifferent data structures yield different approaches to the parallel extrusion\nproblem. We map three well-known causal semantics into our framework. We show\nthat the (parametric) reversibility induced by our framework is causally-\nconsistent and prove a causal correspondence between an appropriate instance of\nthe framework and Boreale and Sangiorgi's causal semantics.\n", - "method": "We map three well-known causal semantics into our framework.", - "id": "1807.11800" - }, - { - "abstract": " We give a structure theorem for inside factorial domains. As an example we\nstudy the monoid of nonnegative integer solutions of equations of the form\n$a_1x_1+\\cdots +a_{r-1}x_{r-1}=a_rx_r$, with $a_1,\\ldots,a_r$ positive\nintegers. This set is isomorphic to a simplicial full affine semigroup, and\nthus it can be described in terms of its extremal rays and the Ap\\'ery sets\nwith respect to the extremal rays.\n", - "method": "We give a structure theorem for inside factorial domains.", - "id": "1807.11885" - }, - { - "abstract": " Formulation of the Lagrangian approach is presented for studying features of\nmotions of stellar bodies with non-zero rest mass in the vicinity of\nfast-spinning black holes. The structure of Lagrangian is discussed. The\ngeneral method is applied to description of body motion in the Kerr model of\nspace-time to transition to the problem of tidal disruption of elastic bodies\nby strong gravitational field.\n", - "method": "The structure of Lagrangian is discussed.", - "id": "1808.00044" - }, - { - "abstract": " Since the pioneering works of the founder of membrane mimetic chemistry Janos\nH. Fendler it is known that a number of atomic or molecular clusters and films\n(including nanoscale ones) are capable of mimicking the membrane functions.\nMembrane mimetic materials can be either soft matter or solid state materials.\nConducting films (including those with magnetic properties) and semiconductors\nare also known to possess membrane mimetic properties. If we consider the agent\nexchange through the membrane in the operator form, the chemical composition of\nthe membranes and their models, as well as the difference between the atomic\nand molecular clusters or layers become not so essential, and hence, membrane\nmimetic chemistry of nano- and mesostructures do not differ significantly\nwithin the agent-based approach. This invited review containing several parts\nreflects the main aspects of the author's report at the conference \"Graphene: a\nmolecule and 2D-crystal\" (September 8-12, 2015, Novosibirsk, Russia) and\nconsiders various aspects of the similarity between the graphene\nnanostructures, membranes and bionic membrane-like nanomaterials.\n", - "method": "Membrane mimetic materials can be either soft matter or solid state materials.", - "id": "1808.00109" - }, - { - "abstract": " Side-effect modulation (SEM) has the potential to be a significant source of\ninterference in future visible light communication (VLC) systems. SEM is a\nvariation in the intensity of the light emitted by a luminaire and is usually a\nside-effect caused by the power supply used to drive the luminaires. For LED\nluminaires powered by a switched mode power supply, the SEM can be at much\nhigher frequencies than that emitted by conventional incandescent or\nfluorescent lighting. It has been shown that the SEM caused by commercially\navailable LED luminaires is often periodic and of low power. In this paper, we\ninvestigate the impact of typical forms of SEM on the performance of optical\nOFDM VLC systems; both ACO-OFDM and DCO-OFDM are considered. Our results show\nthat even low levels of SEM power can significantly degrade the bit-error-rate\nperformance. To solve this problem, an SEM mitigation scheme is described. The\nmitigation scheme is decision-directed and is based on estimating and\nsubtracting the fundamental component of the SEM from the received signal. We\ndescribe two forms of the algorithm; one uses blind estimation while the other\nuses pilot-assisted estimation based on a training sequence. Decision errors,\nresulting in decision noise, limit the performance of the blind estimator even\nwhen estimation is based on very long signals. However, the pilot system can\nachieve more accurate estimations, thus better performance. Results are first\npresented for typical SEM waveforms for the case where the fundamental\nfrequency of the SEM is known. The algorithms are then extended to include a\nfrequency estimation step and the mitigation algorithm is shown also to be\neffective in this case.\n", - "method": "To solve this problem, an SEM mitigation scheme is described.", - "id": "1808.00147" - }, - { - "abstract": " Data analytics and data science play a significant role in nowadays society.\nIn the context of Smart Grids (SG), the collection of vast amounts of data has\nseen the emergence of a plethora of data analysis approaches. In this paper, we\nconduct a Systematic Mapping Study (SMS) aimed at getting insights about\ndifferent facets of SG data analysis: application sub-domains (e.g., power load\ncontrol), aspects covered (e.g., forecasting), used techniques (e.g.,\nclustering), tool-support, research methods (e.g., experiments/simulations),\nreplicability/reproducibility of research. The final goal is to provide a view\nof the current status of research. Overall, we found that each sub-domain has\nits peculiarities in terms of techniques, approaches and research methodologies\napplied. Simulations and experiments play a crucial role in many areas. The\nreplicability of studies is limited concerning the provided implemented\nalgorithms, and to a lower extent due to the usage of private datasets.\n", - "method": "Simulations and experiments play a crucial role in many areas.", - "id": "1808.00156" - }, - { - "abstract": " We study the efficiency of a simple quantum dot heat engine at maximum power.\nIn contrast to the quasi-statically operated Carnot engine whose efficiency\nreaches the theoretical maximum, recent research on more realistic engines\noperated in a finite time has revealed other classes of efficiencies such as\nthe Curzon-Ahlborn efficiency maximizing the power. Such a power-maximizing\nefficiency has been argued to be always the half of the maximum efficiency up\nto the linear order near equilibrium under the tight-coupling condition between\nthermodynamic fluxes. We show, however, that this universality may break down\nfor the quantum dot heat engine, depending on the constraint imposed on the\nengine control parameters (local optimization), even though the tight-coupling\ncondition remains satisfied.It is shown that this deviation is critically\nrelated to the applicability of the linear irreversible thermodynamics.\n", - "method": "We study the efficiency of a simple quantum dot heat engine at maximum power.", - "id": "1808.00219" - }, - { - "abstract": " Context. Giant planets open gaps in their protoplanetary and subsequently\nsuffer so-called type II migration. Schematically, planets are thought to be\ntightly locked within their surrounding disks, and forced to follow the viscous\nadvection of gas onto the central star. This fundamental principle however has\nrecently been questioned, as migrating planets were shown to decouple from the\ngas' radial drift. Aims. In this framework, we question whether the\ntraditionally used linear scaling of migration rate of a giant planet with the\ndisk's viscosity still holds. Additionally, we assess the role of\norbit-crossing material as part of the decoupling mechanism. Methods. We have\nperformed 2D (r, {\\theta}) numerical simulations of point-mass planets embedded\nin locally isothermal {\\alpha}-disks in steady-state accretion, with various\nvalues of {\\alpha}. Arbitrary planetary accretion rates were used as a means to\ndiminish or nullify orbit-crossing flows. Results. We confirm that the\nmigration rate of a gap-opening planet is indeed proportional to the disk's\nviscosity, but is not equal to the gas drift speed in the unperturbed disk. We\nshow that the role of gap-crossing flows is in fact negligible. Conclusions.\nFrom these observations, we propose a new paradigm for type II migration : a\ngiant planet feels a torque from the disk that promotes its migration, while\nthe gap profile relative to the planet is restored on a viscous timescale, thus\nlimiting the planet migration rate to be proportional to the disk's viscosity.\nHence, in disks with low viscosity in the planet region, type II migration\nshould still be very slow. Key words. protoplanetary disks; planet-disk\ninteractions; planets and satellites: formation\n", - "method": "Methods.", - "id": "1808.00381" - }, - { - "abstract": " Bounds for $\\max\\{m,\\tilde{m}\\}$ subject to $m,\\tilde{m} \\in\n\\mathbb{Z}\\cap[1,p)$, $p$ prime, $z$ indivisible by $p$, $m\\tilde{m}\\equiv\nz\\bmod p$ and $m$ belonging to some fixed Beatty sequence $\\{ \\lfloor\nn\\alpha+\\beta \\rfloor : n\\in\\mathbb{N} \\}$ are obtained, assuming certain\nconditions on $\\alpha$. The proof uses a method due to Banks and Shparlinski.\nAs an intermediate step, bounds for the discrete periodic autocorrelation of\nthe finite sequence $0,\\, \\operatorname{e}_p(y\\overline{1}),\n\\operatorname{e}_p(y\\overline{2}), \\ldots,\n\\operatorname{e}_p(y(\\overline{p-1}))$ on average are obtained, where\n$\\operatorname{e}_p(t) = \\exp(2\\pi i t/p)$ and $m\\overline{m} \\equiv 1\\bmod p$.\nThe latter is accomplished by adapting a method due to Kloosterman.\n", - "method": "The proof uses a method due to Banks and Shparlinski.. The latter is accomplished by adapting a method due to Kloosterman.", - "id": "1808.00413" - }, - { - "abstract": " In this paper we significantly extend Serre's table on the vanishing\nproperties of Fourier coefficients of odd powers of the Dedekind eta function.\nWe address several conjectures of Cohen and Str\\\"omberg and give a partial\nanswer to a question of Ono. In the even-power case, we extend Lehmer's\nconjecture on the coefficients of the discriminant function $\\Delta$ to all\nnon-CM-forms. All our results are supported with numerical data. For example\nall Fourier coefficients $a_9(n)$ of the $9$-th power of the Dedekind eta\nfunction are non-vanishing for $n \\leq 10^{10}$. We also relate the\nnon-vanishing of the Fourier coefficients of $\\Delta^2$ to Maeda's conjecture.\n", - "method": "All our results are supported with numerical data.", - "id": "1808.00431" - }, - { - "abstract": " Improved communication systems, shrinking battery sizes and the price drop of\ntracking devices have led to an increasing availability of trajectory tracking\ndata. These data are often analyzed to understand animal behavior.\n In this work, we propose a new model for interpreting the animal movent as a\nmixture of characteristic patterns, that we interpret as different behaviors.\nThe probability that the animal is behaving according to a specific pattern, at\neach time instant, is non-parametrically estimated using the Logistic-Gaussian\nprocess. Owing to a new formalization and the way we specify the\ncoregionalization matrix of the associated multivariate Gaussian process, our\nmodel is invariant with respect to the choice of the reference element and of\nthe ordering of the probability vector components. We fit the model under a\nBayesian framework, and show that the Markov chain Monte Carlo algorithm we\npropose is straightforward to implement.\n We perform a simulation study with the aim of showing the ability of the\nestimation procedure to retrieve the model parameters. We also test the\nperformance of the information criterion we used to select the number of\nbehaviors. The model is then applied to a real dataset where a wolf has been\nobserved before and after procreation. The results are easy to interpret, and\nclear differences emerge in the two phases.\n", - "method": "These data are often analyzed to understand animal behavior.", - "id": "1808.00436" - }, - { - "abstract": " The seminal work by Mackey et al. in 2006 (reference [21] of the article)\nintroduced vector spaces of matrix pencils, with the property that almost all\nthe pencils in the spaces are strong linearizations of a given square regular\nmatrix polynomial. This work was subsequently extended by De Ter\\'an et al. in\n2009 (reference [5] of the article) to include the case of square singular\nmatrix polynomials. We extend this work to non-square matrix polynomials by\nproposing similar vector spaces of rectangular matrix pencils that are equal to\nthe ones introduced by Mackey et al. when the polynomial is square. Moreover,\nthe properties of these vector spaces are similar to those in the article by De\nTer\\'an et al. for the singular case. In particular, the complete eigenvalue\nproblem associated with the matrix polynomial can be solved by using almost\nevery matrix pencil from these spaces. Further, almost every pencil in these\nspaces can be trimmed to form many smaller pencils that are strong\nlinearizations of the matrix polynomial which readily solve the complete\neigenvalue problem for the polynomial. These linearizations are easier to\nconstruct and are often smaller than the Fiedler linearizations introduced by\nDe Ter\\'an et al. in 2012 (reference [7] of the article). Further, the global\nbackward error analysis by Dopico et al. in 2016 (reference [10] of the\narticle) applied to these linearizations, shows that they provide a wide choice\nof linearizations with respect to which the complete polynomial eigenvalue\nproblem can be solved in a globally backward stable manner.\n", - "method": "when the polynomial is square.", - "id": "1808.00517" - }, - { - "abstract": " We use the recently-proposed \\emph{compressible cell} Ising-like model [Phys.\nRev. Lett. \\textbf{120}, 120603 (2018)] to estimate the ratio between thermal\nexpansivity and specific heat (the Gr\\\"uneisen parameter $\\Gamma$) in\nsupercooled water. Near the critical pressure and temperature, $\\Gamma$\nincreases. The $\\Gamma$ value diverges near the pressure-induced finite-$T$\ncritical end-point [Phys. Rev. Lett. \\textbf{104}, 245701 (2010)] and quantum\ncritical points [Phys. Rev. Lett. \\textbf{91}, 066404 (2003)], which indicates\nthat two energy scales are governing the system. This enhanced behavior of\n$\\Gamma$ is caused by the coexistence of high- and low-density liquids [Science\n\\textbf{358}, 1543 (2017)]. Our findings support the proposed liquid-liquid\ncritical point in supercooled water in the No-Man's Land regime, and indicates\npossible applications of this model to other systems.\n", - "method": "Rev.. Lett.", - "id": "1808.00536" - } -] \ No newline at end of file