Datasets:

prithivMLmods

/

Content-Articles

like 8

Chaotic behavior in Casimir oscillators: A case study for phase change materials

Casimir forces between material surfaces at close proximity of less than 200 nm can lead to increased chaotic behavior of actuating devices depending on the strength of the Casimir interaction. We investigate these phenomena for phase change materials in torsional oscillators, where the amorphous to crystalline phase transitions lead to transitions between high and low Casimir force and torque states respectively, without material compositions. For a conservative system bifurcation curve and Poincare maps analysis show the absence of chaotic behavior but with the crystalline phase (high force/torque state) favoring more unstable behavior and stiction. However, for a non-conservative system chaotic behavior can occur introducing significant risk for stiction, which is again more pronounced for the crystalline phase. The latter illustrates the more general scenario that stronger Casimir forces and torques increase the possibility for chaotic behavior. The latter is making impossible to predict whether stiction or stable actuation will occur on a long term basis, and it is setting limitations in the design of micro/nano devices operating at short range nanoscale separations.

Confluence in Probabilistic Rewriting

Driven by the interest of reasoning about probabilistic programming languages, we set out to study a notion of unicity of normal forms for them. To provide a tractable proof method for it, we define a property of distribution confluence which is shown to imply the desired uniqueness (even for infinite sequences of reduction) and further properties. We then carry over several criteria from the classical case, such as Newman's lemma, to simplify proving confluence in concrete languages. Using these criteria, we obtain simple proofs of confluence for $\lambda_1$, an affine probabilistic $\lambda$-calculus, and for Q$^*$, a quantum programming language for which a related property has already been proven in the literature.

Structure preserving schemes for mean-field equations of collective behavior

In this paper we consider the development of numerical schemes for mean-field equations describing the collective behavior of a large group of interacting agents. The schemes are based on a generalization of the classical Chang-Cooper approach and are capable to preserve the main structural properties of the systems, namely nonnegativity of the solution, physical conservation laws, entropy dissipation and stationary solutions. In particular, the methods here derived are second order accurate in transient regimes whereas they can reach arbitrary accuracy asymptotically for large times. Several examples are reported to show the generality of the approach.

Localization optoacoustic tomography

Localization-based imaging has revolutionized fluorescence optical microscopy and has also enabled unprecedented ultrasound images of microvascular structures in deep tissues. Herein, we introduce a new concept of localization optoacoustic tomography (LOAT) that employs rapid sequential acquisition of three-dimensional optoacoustic images from flowing absorbing particles. We show that the new method enables breaking through the spatial resolution barrier of acoustic diffraction while further enhancing the visibility of structures under limited-view tomographic conditions. Given the intrinsic sensitivity of optoacoustics to multiple hemodynamic and oxygenation parameters, LOAT may enable new level of performance in studying functional and anatomical alterations of microcirculation.

Reduced-basis approach to many-body localization

Within the standard model of many-body localization, i.e., the disordered chain of spinless fermions, we investigate how the interaction affects the many-body states in the basis of noninteracting localized Anderson states. From this starting point we follow the approach that uses a reduced basis of many-body states. Together with an extrapolation to the full basis, it proves to be efficient for the evaluation of the stiffnesses of local observables, which remain finite within the non-ergodic regime and represent the hallmark of the many-body localization (MBL). The method enables a larger span of system sizes and, within the MBL regime, allows for a more careful analysis of the size scaling of calculated quantities. On the other hand, the survival stiffness as the representative of non--local quantities, reveals limitations of the reduced-basis approach.

Electron-correlation study of Y III-Tc VII ions using a relativistic coupled-cluster theory

Spectroscopic properties, useful for plasma diagnostics and astrophysics, of a few rubidium-like ions are studied here. We choose one of the simplest, but correlationally challenging series where $d-$ and $f-$ orbitals are present in the core and/or valence shells with $4d$ $^2D_{3/2}$ as the ground state. We study different correlation characteristics of this series and make precise calculations of electronic structure and rates of electromagnetic transitions. Our calculated lifetimes and transition rates are compared with other available experimental and theoretical values. Radiative rates of vacuum ultra-violet electromagnetic transitions of the long lived Tc$^{6+}$ ion, useful in several areas of physics and chemistry, are estimated. To the best of our knowledge, there is no literature for most of these transitions.

Align and Copy: UZH at SIGMORPHON 2017 Shared Task for Morphological Reinflection

This paper presents the submissions by the University of Zurich to the SIGMORPHON 2017 shared task on morphological reinflection. The task is to predict the inflected form given a lemma and a set of morpho-syntactic features. We focus on neural network approaches that can tackle the task in a limited-resource setting. As the transduction of the lemma into the inflected form is dominated by copying over lemma characters, we propose two recurrent neural network architectures with hard monotonic attention that are strong at copying and, yet, substantially different in how they achieve this. The first approach is an encoder-decoder model with a copy mechanism. The second approach is a neural state-transition system over a set of explicit edit actions, including a designated COPY action. We experiment with character alignment and find that naive, greedy alignment consistently produces strong results for some languages. Our best system combination is the overall winner of the SIGMORPHON 2017 Shared Task 1 without external resources. At a setting with 100 training samples, both our approaches, as ensembles of models, outperform the next best competitor.

Krylov methods for low-rank commuting generalized Sylvester equations

We consider generalizations of the Sylvester matrix equation, consisting of the sum of a Sylvester operator and a linear operator $\Pi$ with a particular structure. More precisely, the commutator of the matrix coefficients of the operator $\Pi$ and the Sylvester operator coefficients are assumed to be matrices with low rank. We show (under certain additional conditions) low-rank approximability of this problem, i.e., the solution to this matrix equation can be approximated with a low-rank matrix. Projection methods have successfully been used to solve other matrix equations with low-rank approximability. We propose a new projection method for this class of matrix equations. The choice of subspace is a crucial ingredient for any projection method for matrix equations. Our method is based on an adaption and extension of the extended Krylov subspace method for Sylvester equations. A constructive choice of the starting vector/block is derived from the low-rank commutators. We illustrate the effectiveness of our method by solving large-scale matrix equations arising from applications in control theory and the discretization of PDEs. The advantages of our approach in comparison to other methods are also illustrated.

Atomic-scale origin of dynamic viscoelastic response and creep in disordered solids

Viscoelasticity has been described since the time of Maxwell as an interpolation of purely viscous and purely elastic response, but its microscopic atomic-level mechanism in solids has remained elusive. We studied three model disordered solids: a random lattice, the bond-depleted fcc lattice, and the fcc lattice with vacancies. Within the harmonic approximation for central-force lattices, we applied sum-rules for viscoelastic response derived on the basis of non-affine atomic motions. The latter motions are a direct result of local structural disorder, and in particular, of the lack of inversion-symmetry in disordered lattices. By defining a suitable quantitative and general atomic-level measure of nonaffinity and inversion-symmetry, we show that the viscoelastic responses of all three systems collapse onto a master curve upon normalizing by the overall strength of inversion-symmetry breaking in each system. Close to the isostatic point for central-force lattices, power-law creep $G(t)\sim t^{-1/2}$ emerges as a consequence of the interplay between soft vibrational modes and non-affine dynamics, and various analytical scalings, supported by numerical calculations, are predicted by the theory.

On-sky closed loop correction of atmospheric dispersion for high-contrast coronagraphy and astrometry

Adaptive optic (AO) systems delivering high levels of wavefront correction are now common at observatories. One of the main limitations to image quality after wavefront correction comes from atmospheric refraction. An Atmospheric dispersion compensator (ADC) is employed to correct for atmospheric refraction. The correction is applied based on a look-up table consisting of dispersion values as a function of telescope elevation angle. The look-up table based correction of atmospheric dispersion results in imperfect compensation leading to the presence of residual dispersion in the point-spread function (PSF) and is insufficient when sub-milliarcsecond precision is required. The presence of residual dispersion can limit the achievable contrast while employing high-performance coronagraphs or can compromise high-precision astrometric measurements. In this paper, we present the first on-sky closed-loop correction of atmospheric dispersion by directly using science path images. The concept behind the measurement of dispersion utilizes the chromatic scaling of focal plane speckles. An adaptive speckle grid generated with a deformable mirror (DM) that has a sufficiently large number of actuators is used to accurately measure the residual dispersion and subsequently correct it by driving the ADC. We have demonstrated with the Subaru Coronagraphic Extreme AO (SCExAO) system on-sky closed-loop correction of residual dispersion to < 1 mas across H-band. This work will aid in the direct detection of habitable exoplanets with upcoming extremely large telescopes (ELTs) and also provide a diagnostic tool to test the performance of instruments which require sub-milliarcsecond correction.

Experimental observations and modelling of intrinsic rotation reversals in tokamaks

The progress made in understanding spontaneous toroidal rotation reversals in tokamaks is reviewed and current ideas to solve this ten-year-old puzzle are explored. The paper includes a summarial synthesis of the experimental observations in AUG, C-Mod, KSTAR, MAST and TCV tokamaks, reasons why turbulent momentum transport is thought to be responsible for the reversals, a review of the theory of turbulent momentum transport and suggestions for future investigations.

Where, When, and How mmWave is Used in 5G and Beyon

Wireless engineers and business planners commonly raise the question on where, when, and how millimeter-wave (mmWave) will be used in 5G and beyond. Since the next generation network is not just a new radio access standard, but instead an integration of networks for vertical markets with diverse applications, answers to the question depend on scenarios and use cases to be deployed. This paper gives four 5G mmWave deployment examples and describes in chronological order the scenarios and use cases of their probable deployment, including expected system architectures and hardware prototypes. The paper starts with 28 GHz outdoor backhauling for fixed wireless access and moving hotspots, which will be demonstrated at the PyeongChang winter Olympic games in 2018. The second deployment example is a 60 GHz unlicensed indoor access system at the Tokyo-Narita airport, which is combined with Mobile Edge Computing (MEC) to enable ultra-high speed content download with low latency. The third example is mmWave mesh network to be used as a micro Radio Access Network ({\mu}-RAN), for cost-effective backhauling of small-cell Base Stations (BSs) in dense urban scenarios. The last example is mmWave based Vehicular-to-Vehicular (V2V) and Vehicular-to-Everything (V2X) communications system, which enables automated driving by exchanging High Definition (HD) dynamic map information between cars and Roadside Units (RSUs). For 5G and beyond, mmWave and MEC will play important roles for a diverse set of applications that require both ultra-high data rate and low latency communications.

A simple script language for choreography of multiple, synchronizing non-anthropomorphic robots

The scripting language described in this document is (in the first place) intended to be used on robots developed by Anja M{\o}lle Lindelof and Henning Christiansen as part of a research project about robots performing on stage. The target robots are expected to appear as familiar domestic objects that take their own life, so to speak, and perhaps perform together with human players, creating at illusion of a communication between them. In the current version, these robots' common behaviour is determined uniquely by a script written in the language described here -- the only possible autonomy for the robots is action to correct dynamically for inaccuracies that arise during a performance. The present work is preliminary and has not been compared to properly to other research work in this area, and the testing is still limited. A first implementation on small Lego Mindstorms based robots is under development by Mads Saustrup Fox as part of his master thesis work.

Stability results for abstract evolution equations with intermittent time-delay feedback

We consider abstract evolution equations with on-off time delay feedback. Without the time delay term, the model is described by an exponentially stable semigroup. We show that, under appropriate conditions involving the delay term, the system remains asymptotically stable. Under additional assumptions exponential stability results are also obtained. Concrete examples illustrating the abstract results are finally given.

Solar wind turbulent cascade from MHD to sub-ion scales: large-size 3D hybrid particle-in-cell simulations

Spectral properties of the turbulent cascade from fluid to kinetic scales in collisionless plasmas are investigated by means of large-size three-dimensional (3D) hybrid (fluid electrons, kinetic protons) particle-in-cell simulations. Initially isotropic Alfvènic fluctuations rapidly develop a strongly anisotropic turbulent cascade, mainly in the direction perpendicular to the ambient magnetic field. The omnidirectional magnetic field spectrum shows a double power-law behavior over almost two decades in wavenumber, with a Kolmogorov-like index at large scales, a spectral break around ion scales, and a steepening at sub-ion scales. Power laws are also observed in the spectra of the ion bulk velocity, density, and electric field, both at magnetohydrodynamic (MHD) and at kinetic scales. Despite the complex structure, the omnidirectional spectra of all fields at ion and sub-ion scales are in remarkable quantitative agreement with those of a two-dimensional (2D) simulation with similar physical parameters. This provides a partial, a-posteriori validation of the 2D approximation at kinetic scales. Conversely, at MHD scales, the spectra of the density and of the velocity (and, consequently, of the electric field) exhibit differences between the 2D and 3D cases. Although they can be partly ascribed to the lower spatial resolution, the main reason is likely the larger importance of compressible effects in a full geometry. Our findings are also in remarkable quantitative agreement with solar wind observations.

Blind source separation of tensor-valued time series

The blind source separation model for multivariate time series generally assumes that the observed series is a linear transformation of an unobserved series with temporally uncorrelated or independent components. Given the observations, the objective is to find a linear transformation that recovers the latent series. Several methods for accomplishing this exist and three particular ones are the classic SOBI and the recently proposed generalized FOBI (gFOBI) and generalized JADE (gJADE), each based on the use of joint lagged moments. In this paper we generalize the methodologies behind these algorithms for tensor-valued time series. We assume that our data consists of a tensor observed at each time point and that the observations are linear transformations of latent tensors we wish to estimate. The tensorial generalizations are shown to have particularly elegant forms and we show that each of them is Fisher consistent and orthogonal equivariant. Comparing the new methods with the original ones in various settings shows that the tensorial extensions are superior to both their vector-valued counterparts and to two existing tensorial dimension reduction methods for i.i.d. data. Finally, applications to fMRI-data and video processing show that the methods are capable of extracting relevant information from noisy high-dimensional data.

Mixed Cages

We introduce the notion of a $[z, r; g]$-mixed cage. A $[z, r; g]$-mixed cage is a mixed graph $G$, $z$-regular by arcs, $r$-regular by edges, with girth $g$ and minimum order. In this paper we prove the existence of $[z, r ;g]$-mixed cages and exhibit families of mixed cages for some specific values. We also give lower and upper bounds for some choices of $z, r$ and $g$. In particular we present the first results on $[z,r;g]$- mixed cages for $z=1$ and any $r\geq 1$ and $g\geq 3$, and for any $z\geq 1$, $r=1$ and $g=4$.

Single-beam dielectric-microsphere trapping with optical heterodyne detection

A technique to levitate and measure the three-dimensional position of micrometer-sized dielectric spheres with heterodyne detection is presented. The two radial degrees of freedom are measured by interfering light transmitted through the microsphere with a reference wavefront, while the axial degree of freedom is measured from the phase of the light reflected from the surface of the microsphere. This method pairs the simplicity and accessibility of single beam optical traps to a measurement of displacement that is intrinsically calibrated by the wavelength of the trapping light and has exceptional immunity to stray light. A theoretical shot noise limit of $1.3\times10^{-13}\,\text{m}/\sqrt{\text{Hz}}$ for the radial degrees of freedom, and $3.0\times10^{-15} \, \text{m}/\sqrt{\text{Hz}}$ for the axial degree of freedom can be obtained in the system described. The measured acceleration noise in the radial direction is $7.5\times10^{-5} \, (\text{m/s}^2)/\sqrt{\text{Hz}}$.

An Empirical Analysis of Traceability in the Monero Blockchain

Monero is a privacy-centric cryptocurrency that allows users to obscure their transactions by including chaff coins, called "mixins," along with the actual coins they spend. In this paper, we empirically evaluate two weaknesses in Monero's mixin sampling strategy. First, about 62% of transaction inputs with one or more mixins are vulnerable to "chain-reaction" analysis -- that is, the real input can be deduced by elimination. Second, Monero mixins are sampled in such a way that they can be easily distinguished from the real coins by their age distribution; in short, the real input is usually the "newest" input. We estimate that this heuristic can be used to guess the real input with 80% accuracy over all transactions with 1 or more mixins. Next, we turn to the Monero ecosystem and study the importance of mining pools and the former anonymous marketplace AlphaBay on the transaction volume. We find that after removing mining pool activity, there remains a large amount of potentially privacy-sensitive transactions that are affected by these weaknesses. We propose and evaluate two countermeasures that can improve the privacy of future transactions.

Automatic Face Image Quality Prediction

Face image quality can be defined as a measure of the utility of a face image to automatic face recognition. In this work, we propose (and compare) two methods for automatic face image quality based on target face quality values from (i) human assessments of face image quality (matcher-independent), and (ii) quality values computed from similarity scores (matcher-dependent). A support vector regression model trained on face features extracted using a deep convolutional neural network (ConvNet) is used to predict the quality of a face image. The proposed methods are evaluated on two unconstrained face image databases, LFW and IJB-A, which both contain facial variations with multiple quality factors. Evaluation of the proposed automatic face image quality measures shows we are able to reduce the FNMR at 1% FMR by at least 13% for two face matchers (a COTS matcher and a ConvNet matcher) by using the proposed face quality to select subsets of face images and video frames for matching templates (i.e., multiple faces per subject) in the IJB-A protocol. To our knowledge, this is the first work to utilize human assessments of face image quality in designing a predictor of unconstrained face quality that is shown to be effective in cross-database evaluation.

Stochastic Block Models are a Discrete Surface Tension

Networks, which represent agents and interactions between them, arise in myriad applications throughout the sciences, engineering, and even the humanities. To understand large-scale structure in a network, a common task is to cluster a network's nodes into sets called "communities" such that there are dense connections within communities but sparse connections between them. A popular and statistically principled method to perform such clustering is to use a family of generative models known as stochastic block models (SBMs). In this paper, we show that maximum likelihood estimation in an SBM is a network analog of a well-known continuum surface-tension problem that arises from an application in metallurgy. To illustrate the utility of this bridge, we implement network analogs of three surface-tension algorithms, with which we successfully recover planted community structure in synthetic networks and which yield fascinating insights on empirical networks from the field of hyperspectral video segmentation.

Robust d-wave pairing symmetry in multi-orbital cobalt high temperature superconductors

The pairing symmetry of the newly proposed cobalt high temperature (high-$T_c$) superconductors formed by vertex shared cation-anion tetrahedral complexes is studied by the methods of mean field, random phase approximation (RPA) and functional renormalization group (FRG) analysis. The results of all these methods show that the $d_{x^2-y^2}$ pairing symmetry is robustly favored near half filling. The RPA and FRG methods, which are valid in weak interaction regions, predict that the superconducting state is also strongly orbital selective, namely the $d_{x^2-y^2}$ orbital that has the largest density near half filling among the three $t_{2g}$ orbitals dominates superconducting pairing. These results suggest that the new materials, if synthesized, can provide indisputable test to high-$T_c$ pairing mechanism and the validity of different theoretical methods.

Thermodynamic properties of Ba$_2$CoSi$_2$O$_6$Cl$_2$ in strong magnetic field: Realization of flat-band physics in a highly frustrated quantum magnet

The search for flat-band solid-state realizations is a crucial issue to verify or to challenge theoretical predictions for quantum many-body flat-band systems. For frustrated quantum magnets flat bands lead to various unconventional properties related to the existence of localized many-magnon states. The recently synthesized magnetic compound Ba$_2$CoSi$_2$O$_6$Cl$_2$ seems to be an almost perfect candidate to observe these features in experiments. We develop a theory for Ba$_2$CoSi$_2$O$_6$Cl$_2$ by adapting the localized-magnon concept to this compound. We first show that our theory describes the known experimental facts and then we propose new experimental studies to detect a field-driven phase transition related to a Wigner-crystal-like ordering of localized magnons at low temperatures.

Fisher Information and Natural Gradient Learning of Random Deep Networks

A deep neural network is a hierarchical nonlinear model transforming input signals to output signals. Its input-output relation is considered to be stochastic, being described for a given input by a parameterized conditional probability distribution of outputs. The space of parameters consisting of weights and biases is a Riemannian manifold, where the metric is defined by the Fisher information matrix. The natural gradient method uses the steepest descent direction in a Riemannian manifold, so it is effective in learning, avoiding plateaus. It requires inversion of the Fisher information matrix, however, which is practically impossible when the matrix has a huge number of dimensions. Many methods for approximating the natural gradient have therefore been introduced. The present paper uses statistical neurodynamical method to reveal the properties of the Fisher information matrix in a net of random connections under the mean field approximation. We prove that the Fisher information matrix is unit-wise block diagonal supplemented by small order terms of off-block-diagonal elements, which provides a justification for the quasi-diagonal natural gradient method by Y. Ollivier. A unitwise block-diagonal Fisher metrix reduces to the tensor product of the Fisher information matrices of single units. We further prove that the Fisher information matrix of a single unit has a simple reduced form, a sum of a diagonal matrix and a rank 2 matrix of weight-bias correlations. We obtain the inverse of Fisher information explicitly. We then have an explicit form of the natural gradient, without relying on the numerical matrix inversion, which drastically speeds up stochastic gradient learning.

Dynamic of plumes and scaling during the melting of a Phase Change Material heated from below

We identify and describe the main dynamic regimes occurring during the melting of the PCM n-octadecane in horizontal layers of several sizes heated from below. This configuration allows to cover a wide range of effective Rayleigh numbers on the liquid PCM phase, up to $\sim 10^9$, without changing any external parameter control. We identify four different regimes as time evolves: (i) the conductive regime, (ii) linear regime, (iii) coarsening regime and (iv) turbulent regime. The first two regimes appear at all domain sizes. However the third and fourth regime require a minimum advance of the solid/liquid interface to develop, and we observe them only for large enough domains. The transition to turbulence takes places after a secondary instability that forces the coarsening of the thermal plumes. Each one of the melting regimes creates a distinct solid/liquid front that characterizes the internal state of the melting process. We observe that most of the magnitudes of the melting process are ruled by power laws, although not all of them. Thus the number of plumes, some regimes of the Rayleigh number as a function of time, the number of plumes after the primary and secondary instability, the thermal and kinetic boundary layers follow simple power laws. In particular, we find that the Nusselt number scales with the Rayleigh number as $Nu \sim Ra^{0.29}$ in the turbulent regime, consistent with theories and experiments on Rayleigh-Bénard convection that show an exponent $2/7$.

Parallel mean curvature surfaces in four-dimensional homogeneous spaces

We survey different classification results for surfaces with parallel mean curvature immersed into some Riemannian homogeneous four-manifolds, including real and complex space forms, and product spaces. We provide a common framework for this problem, with special attention to the existence of holomorphic quadratic differentials on such surfaces. The case of spheres with parallel mean curvature is also explained in detail, as well as the state-of-the-art advances in the general problem.

Recent implementations, applications, and extensions of the Locally Optimal Block Preconditioned Conjugate Gradient method (LOBPCG)

Since introduction [A. Knyazev, Toward the optimal preconditioned eigensolver: Locally optimal block preconditioned conjugate gradient method, SISC (2001) DOI:10.1137/S1064827500366124] and efficient parallel implementation [A. Knyazev et al., Block locally optimal preconditioned eigenvalue xolvers (BLOPEX) in HYPRE and PETSc, SISC (2007) DOI:10.1137/060661624], LOBPCG has been used is a wide range of applications in mechanics, material sciences, and data sciences. We review its recent implementations and applications, as well as extensions of the local optimality idea beyond standard eigenvalue problems.

An evaluation of cosmological models from expansion and growth of structure measurements

We compare a large suite of theoretical cosmological models to observational data from the cosmic microwave background, baryon acoustic oscillation measurements of expansion, Type Ia SNe measurements of expansion, redshift space distortion measurements of the growth of structure, and the local Hubble constant. Our theoretical models include parametrizations of dark energy as well as physical models of dark energy and modified gravity. We determine the constraints on the model parameters, incorporating the redshift space distortion data directly in the analysis. To determine whether models can be ruled out, we evaluate the $p$ value (the probability under the model of obtaining data as bad or worse than the observed data). In our comparison, we find the well known tension of H$_0$ with the other data; no model resolves this tension successfully. Among the models we consider, the large scale growth of structure data does not affect the modified gravity models as a category particularly differently than dark energy models; it matters for some modified gravity models but not others, and the same is true for dark energy models. We compute predicted observables for each model under current observational constraints, and identify models for which future observational constraints will be particularly informative.

The variety of $2$-dimensional algebras over an algebraically closed field

The work is devoted to the variety of $2$-dimensional algebras over an algebraically closed field. Firstly, we classify such algebras modulo isomorphism. Then we describe the degenerations and the closures of principal algebra series in the variety under consideration. Finally, we apply our results to obtain analogous descriptions for the subvarieties of flexible, and bicommutative algebras. In particular, we describe rigid algebras and irreducible components for these subvarieties.

A Variational Observation Model of 3D Object for Probabilistic Semantic SLAM

We present a Bayesian object observation model for complete probabilistic semantic SLAM. Recent studies on object detection and feature extraction have become important for scene understanding and 3D mapping. However, 3D shape of the object is too complex to formulate the probabilistic observation model; therefore, performing the Bayesian inference of the object-oriented features as well as their pose is less considered. Besides, when the robot equipped with an RGB mono camera only observes the projected single view of an object, a significant amount of the 3D shape information is abandoned. Due to these limitations, semantic SLAM and viewpoint-independent loop closure using volumetric 3D object shape is challenging. In order to enable the complete formulation of probabilistic semantic SLAM, we approximate the observation model of a 3D object with a tractable distribution. We also estimate the variational likelihood from the 2D image of the object to exploit its observed single view. In order to evaluate the proposed method, we perform pose and feature estimation, and demonstrate that the automatic loop closure works seamlessly without additional loop detector in various environments.

Generation of optical frequency combs via four-wave mixing processes for low- and medium-resolution astronomy

We investigate the generation of optical frequency combs through a cascade of four-wave mixing processes in nonlinear fibres with optimised parameters. The initial optical field consists of two continuous-wave lasers with frequency separation larger than 40 GHz (312.7 pm at 1531 nm). It propagates through three nonlinear fibres. The first fibre serves to pulse shape the initial sinusoidal-square pulse, while a strong pulse compression down to sub-100 fs takes place in the second fibre which is an amplifying erbium-doped fibre. The last stage is a low-dispersion highly nonlinear fibre where the frequency comb bandwidth is increased and the line intensity is equalised. We model this system using the generalised nonlinear Schrödinger equation and investigate it in terms of fibre lengths, fibre dispersion, laser frequency separation and input powers with the aim to minimise the frequency comb noise. With the support of the numerical results, a frequency comb is experimentally generated, first in the near infra-red and then it is frequency-doubled into the visible spectral range. Using a MUSE-type spectrograph, we evaluate the comb performance for astronomical wavelength calibration in terms of equidistancy of the comb lines and their stability.

Cross-referencing Social Media and Public Surveillance Camera Data for Disaster Response

Physical media (like surveillance cameras) and social media (like Instagram and Twitter) may both be useful in attaining on-the-ground information during an emergency or disaster situation. However, the intersection and reliability of both surveillance cameras and social media during a natural disaster are not fully understood. To address this gap, we tested whether social media is of utility when physical surveillance cameras went off-line during Hurricane Irma in 2017. Specifically, we collected and compared geo-tagged Instagram and Twitter posts in the state of Florida during times and in areas where public surveillance cameras went off-line. We report social media content and frequency and content to determine the utility for emergency managers or first responders during a natural disaster.

Nonlinear Dynamics of a Viscous Bubbly Fluid

A physical model of a three-dimensional flow of a viscous bubbly fluid in an intermediate regime between bubble formation and breakage is presented. The model is based on mechanics and thermodynamics of a single bubble coupled to the dynamics of a viscous fluid as a whole, and takes into account multiple physical effects, including gravity, viscosity, and surface tension. Dimensionless versions of the resulting nonlinear model are obtained, and values of dimensionless parameters are estimated for typical magma flows in horizontal subaerial lava fields and vertical volcanic conduits. Exact solutions of the resulting system of nonlinear equations corresponding to equilibrium flows and traveling waves are analyzed in the one-dimensional setting. Generalized Su-Gardner-type perturbation analysis is employed to study approximate solutions of the model in the long-wave ansatz. Simplified nonlinear partial differential equations (PDE) satisfied by the leading terms of the perturbation solutions are systematically derived. It is shown that for specific classes of perturbations, approximate solutions of the bubbly fluid model arise from solutions of the classical diffusion, Burgers, variable-coefficient Burgers, and Korteweg-de Vries equations.

ConvNet-Based Localization of Anatomical Structures in 3D Medical Images

Localization of anatomical structures is a prerequisite for many tasks in medical image analysis. We propose a method for automatic localization of one or more anatomical structures in 3D medical images through detection of their presence in 2D image slices using a convolutional neural network (ConvNet). A single ConvNet is trained to detect presence of the anatomical structure of interest in axial, coronal, and sagittal slices extracted from a 3D image. To allow the ConvNet to analyze slices of different sizes, spatial pyramid pooling is applied. After detection, 3D bounding boxes are created by combining the output of the ConvNet in all slices. In the experiments 200 chest CT, 100 cardiac CT angiography (CTA), and 100 abdomen CT scans were used. The heart, ascending aorta, aortic arch, and descending aorta were localized in chest CT scans, the left cardiac ventricle in cardiac CTA scans, and the liver in abdomen CT scans. Localization was evaluated using the distances between automatically and manually defined reference bounding box centroids and walls. The best results were achieved in localization of structures with clearly defined boundaries (e.g. aortic arch) and the worst when the structure boundary was not clearly visible (e.g. liver). The method was more robust and accurate in localization multiple structures.

Gradient estimates for heat kernels and harmonic functions

Let $(X,d,\mu)$ be a doubling metric measure space endowed with a Dirichlet form $\E$ deriving from a "carré du champ". Assume that $(X,d,\mu,\E)$ supports a scale-invariant $L^2$-Poincaré inequality. In this article, we study the following properties of harmonic functions, heat kernels and Riesz transforms for $p\in (2,\infty]$: (i) $(G_p)$: $L^p$-estimate for the gradient of the associated heat semigroup; (ii) $(RH_p)$: $L^p$-reverse Hölder inequality for the gradients of harmonic functions; (iii) $(R_p)$: $L^p$-boundedness of the Riesz transform ($p<\infty$); (iv) $(GBE)$: a generalised Bakry-Émery condition. We show that, for $p\in (2,\infty)$, (i), (ii) (iii) are equivalent, while for $p=\infty$, (i), (ii), (iv) are equivalent. Moreover, some of these equivalences still hold under weaker conditions than the $L^2$-Poincaré inequality. Our result gives a characterisation of Li-Yau's gradient estimate of heat kernels for $p=\infty$, while for $p\in (2,\infty)$ it is a substantial improvement as well as a generalisation of earlier results by Auscher-Coulhon-Duong-Hofmann [7] and Auscher-Coulhon [6]. Applications to isoperimetric inequalities and Sobolev inequalities are given. Our results apply to Riemannian and sub-Riemannian manifolds as well as to non-smooth spaces, and to degenerate elliptic/parabolic equations in these settings.

The Deep Underground Neutrino Experiment -- DUNE: the precision era of neutrino physics

The last decade was remarkable for neutrino physics. In particular, the phenomenon of neutrino flavor oscillations has been firmly established by a series of independent measurements. All parameters of the neutrino mixing are now known and we have elements to plan a judicious exploration of new scenarios that are opened by these recent advances. With precise measurements, we can test the 3-neutrino paradigm, neutrino mass hierarchy and CP asymmetry in the lepton sector. The future long-baseline experiments are considered to be a fundamental tool to deepen our knowledge of electroweak interactions. The Deep Underground Neutrino Experiment -- DUNE will detect a broad-band neutrino beam from Fermilab in an underground massive Liquid Argon Time-Projection Chamber at an L/E of about $10^3$ km / GeV to reach good sensitivity for CP-phase measurements and the determination of the mass hierarchy. The dimensions and the depth of the Far Detector also create an excellent opportunity to look for rare signals like proton decay to study violation of baryonic number, as well as supernova neutrino bursts, broadening the scope of the experiment to astrophysics and associated impacts in cosmology. In this presentation, we will discuss the physics motivations and the main experimental features of the DUNE project required to reach its scientific goals.

Variations of $q$-Garnier system

We study several variants of q-Garnier system corresponding to various directions of discrete time evolutions. We also investigate a relation between the $q$-Garnier system and Suzuki's higher order $q$-Painlev/'e system by using a duality of the $q$-KP system.

Quantum Quench dynamics in Non-local Luttinger Model: Rigorous Results

We investigate, in the Luttinger model with fixed box potential, the time evolution of an inhomogeneous state prepared as a localized fermion added to the noninteracting ground state. We proved that, if the state is evolved with the interacting Hamiltonian, the averaged density has two peaks moving in opposite directions, with a constant but renormalized velocity. We also proved that a dynamical `Landau quasi-particle weight' appears in the oscillating part of the averaged density, asymptotically vanishing with large time. The results are proved with the Mattis-Lieb diagonalization method. A simpler proof with the exact Bosonization formulas is also provided.

Dynamics of Relaxed Inflation

The cosmological relaxation of the electroweak scale has been proposed as a mechanism to address the hierarchy problem of the Standard Model. A field, the relaxion, rolls down its potential and, in doing so, scans the squared mass parameter of the Higgs, relaxing it to a parametrically small value. In this work, we promote the relaxion to an inflaton. We couple it to Abelian gauge bosons, thereby introducing the necessary dissipation mechanism which slows down the field in the last stages. We describe a novel reheating mechanism, which relies on the gauge-boson production leading to strong electromagnetic fields, and proceeds via the vacuum production of electron-positron pairs through the Schwinger effect. We refer to this mechanism as Schwinger reheating. We discuss the cosmological dynamics of the model and the phenomenological constraints from CMB and other experiments. We find that a cutoff close to the Planck scale may be achieved. In its minimal form, the model does not generate sufficient curvature perturbations and additional ingredients, such as a curvaton field, are needed.

Clustering is semidefinitely not that hard: Nonnegative SDP for manifold disentangling

In solving hard computational problems, semidefinite program (SDP) relaxations often play an important role because they come with a guarantee of optimality. Here, we focus on a popular semidefinite relaxation of K-means clustering which yields the same solution as the non-convex original formulation for well segregated datasets. We report an unexpected finding: when data contains (greater than zero-dimensional) manifolds, the SDP solution captures such geometrical structures. Unlike traditional manifold embedding techniques, our approach does not rely on manually defining a kernel but rather enforces locality via a nonnegativity constraint. We thus call our approach NOnnegative MAnifold Disentangling, or NOMAD. To build an intuitive understanding of its manifold learning capabilities, we develop a theoretical analysis of NOMAD on idealized datasets. While NOMAD is convex and the globally optimal solution can be found by generic SDP solvers with polynomial time complexity, they are too slow for modern datasets. To address this problem, we analyze a non-convex heuristic and present a new, convex and yet efficient, algorithm, based on the conditional gradient method. Our results render NOMAD a versatile, understandable, and powerful tool for manifold learning.

Dispersive optical detection of magnetic Feshbach resonances in ultracold gases

Magnetically tunable Feshbach resonances in ultracold atomic systems are chiefly identified and characterized through time consuming atom loss spectroscopy. We describe an off-resonant dispersive optical probing technique to rapidly locate Feshbach resonances and demonstrate the method by locating four resonances of $^{87}$Rb, between the $|\rm{F} = 1, \rm{m_F}=1 \rangle$ and $|\rm{F} = 2, \rm{m_F}=0 \rangle$ states. Despite the loss features being $\lesssim0.1$ G wide, we require only 21 experimental runs to explore a magnetic field range >18 G, where $1~\rm{G}=10^{-4}$ T. The resonances consist of two known s-wave features in the vicinity of 9 G and 18 G and two previously unobserved p-wave features near 5 G and 10 G. We further utilize the dispersive approach to directly characterize the two-body loss dynamics for each Feshbach resonance.

Unveiling Bias Compensation in Turbo-Based Algorithms for (Discrete) Compressed Sensing

In Compressed Sensing, a real-valued sparse vector has to be recovered from an underdetermined system of linear equations. In many applications, however, the elements of the sparse vector are drawn from a finite set. Adapted algorithms incorporating this additional knowledge are required for the discrete-valued setup. In this paper, turbo-based algorithms for both cases are elucidated and analyzed from a communications engineering perspective, leading to a deeper understanding of the algorithm. In particular, we gain the intriguing insight that the calculation of extrinsic values is equal to the unbiasing of a biased estimate and present an improved algorithm.

Accelerated Linear Convergence of Stochastic Momentum Methods in Wasserstein Distances

Momentum methods such as Polyak's heavy ball (HB) method, Nesterov's accelerated gradient (AG) as well as accelerated projected gradient (APG) method have been commonly used in machine learning practice, but their performance is quite sensitive to noise in the gradients. We study these methods under a first-order stochastic oracle model where noisy estimates of the gradients are available. For strongly convex problems, we show that the distribution of the iterates of AG converges with the accelerated $O(\sqrt{\kappa}\log(1/\varepsilon))$ linear rate to a ball of radius $\varepsilon$ centered at a unique invariant distribution in the 1-Wasserstein metric where $\kappa$ is the condition number as long as the noise variance is smaller than an explicit upper bound we can provide. Our analysis also certifies linear convergence rates as a function of the stepsize, momentum parameter and the noise variance; recovering the accelerated rates in the noiseless case and quantifying the level of noise that can be tolerated to achieve a given performance. In the special case of strongly convex quadratic objectives, we can show accelerated linear rates in the $p$-Wasserstein metric for any $p\geq 1$ with improved sensitivity to noise for both AG and HB through a non-asymptotic analysis under some additional assumptions on the noise structure. Our analysis for HB and AG also leads to improved non-asymptotic convergence bounds in suboptimality for both deterministic and stochastic settings which is of independent interest. To the best of our knowledge, these are the first linear convergence results for stochastic momentum methods under the stochastic oracle model. We also extend our results to the APG method and weakly convex functions showing accelerated rates when the noise magnitude is sufficiently small.

Compressed sensing with sparse corruptions: Fault-tolerant sparse collocation approximations

The recovery of approximately sparse or compressible coefficients in a Polynomial Chaos Expansion is a common goal in modern parametric uncertainty quantification (UQ). However, relatively little effort in UQ has been directed toward theoretical and computational strategies for addressing the sparse corruptions problem, where a small number of measurements are highly corrupted. Such a situation has become pertinent today since modern computational frameworks are sufficiently complex with many interdependent components that may introduce hardware and software failures, some of which can be difficult to detect and result in a highly polluted simulation result. In this paper we present a novel compressive sampling-based theoretical analysis for a regularized $\ell^1$ minimization algorithm that aims to recover sparse expansion coefficients in the presence of measurement corruptions. Our recovery results are uniform, and prescribe algorithmic regularization parameters in terms of a user-defined a priori estimate on the ratio of measurements that are believed to be corrupted. We also propose an iteratively reweighted optimization algorithm that automatically refines the value of the regularization parameter, and empirically produces superior results. Our numerical results test our framework on several medium-to-high dimensional examples of solutions to parameterized differential equations, and demonstrate the effectiveness of our approach.

Learning to Remember Rare Events

Despite recent advances, memory-augmented deep neural networks are still limited when it comes to life-long and one-shot learning, especially in remembering rare events. We present a large-scale life-long memory module for use in deep learning. The module exploits fast nearest-neighbor algorithms for efficiency and thus scales to large memory sizes. Except for the nearest-neighbor query, the module is fully differentiable and trained end-to-end with no extra supervision. It operates in a life-long manner, i.e., without the need to reset it during training. Our memory module can be easily added to any part of a supervised neural network. To show its versatility we add it to a number of networks, from simple convolutional ones tested on image classification to deep sequence-to-sequence and recurrent-convolutional models. In all cases, the enhanced network gains the ability to remember and do life-long one-shot learning. Our module remembers training examples shown many thousands of steps in the past and it can successfully generalize from them. We set new state-of-the-art for one-shot learning on the Omniglot dataset and demonstrate, for the first time, life-long one-shot learning in recurrent neural networks on a large-scale machine translation task.

Stacking-dependent electronic structure of trilayer graphene resolved by nanospot angle-resolved photoemission spectroscopy

The crystallographic stacking order in multilayer graphene plays an important role in determining its electronic structure. In trilayer graphene, rhombohedral stacking (ABC) is particularly intriguing, exhibiting a flat band with an electric-field tunable band gap. Such electronic structure is distinct from simple hexagonal stacking (AAA) or typical Bernal stacking (ABA), and is promising for nanoscale electronics, optoelectronics applications. So far clean experimental electronic spectra on the first two stackings are missing because the samples are usually too small in size (um or nm scale) to be resolved by conventional angle-resolved photoemission spectroscopy (ARPES). Here by using ARPES with nanospot beam size (NanoARPES), we provide direct experimental evidence for the coexistence of three different stackings of trilayer graphene and reveal their distinctive electronic structures directly. By fitting the experimental data, we provide important experimental band parameters for describing the electronic structure of trilayer graphene with different stackings.

Interaction energy between vortices of vector fields on Riemannian surfaces

We study a variational Ginzburg-Landau type model depending on a small parameter $\epsilon>0$ for (tangent) vector fields on a $2$-dimensional Riemannian surface. As $\epsilon\to 0$, the vector fields tend to be of unit length and will have singular points of a (non-zero) index, called vortices. Our main result determines the interaction energy between these vortices as a $\Gamma$-limit (at the second order) as $\epsilon\to 0$.

Multimodal Nonlinear Microscope based on a Compact Fiber-format Laser Source

We present a multimodal non-linear optical (NLO) laser-scanning microscope, based on a compact fiber-format excitation laser and integrating coherent anti-Stokes Raman scattering (CARS), stimulated Raman scattering (SRS) and two-photon-excitation fluorescence (TPEF) on a single platform. We demonstrate its capabilities in simultaneously acquiring CARS and SRS images of a blend of 6-{\mu}m poly(methyl methacrylate) beads and 3-{\mu}m polystyrene beads. We then apply it to visualize cell walls and chloroplast of an unprocessed fresh leaf of Elodea aquatic plant via SRS and TPEF modalities, respectively. The presented NLO microscope, developed in house using off-the-shelf components, offers full accessibility to the optical path and ensures its easy re-configurability and flexibility.

The Diffuse Light of the Universe - On the microwave background before and after its discovery: open questions

In 1965, the discovery of a new type of uniform radiation, located between radiowaves and infrared light, was accidental. Known today as Cosmic Microwave background (CMB), this diffuse radiation is commonly interpreted as a fossil light released in an early hot and dense universe and constitutes today the main 'pilar' of the big bang cosmology. Considerable efforts have been devoted to derive fundamental cosmological parameters from the characteristics of this radiation that led to a surprising universe that is shaped by at least three major unknown components: inflation, dark matter and dark energy. This is an important weakness of the present consensus cosmological model that justifies raising several questions on the CMB interpretation. Can we consider its cosmological nature as undisputable? Do other possible interpretations exist in the context of other cosmological theories or simply as a result of other physical mechanisms that could account for it? In an effort to questioning the validity of scientific hypotheses and the under-determination of theories compared to observations, we examine here the difficulties that still exist on the interpretation of this diffuse radiation and explore other proposed tracks to explain its origin. We discuss previous historical concepts of diffuse radiation before and after the CMB discovery and underline the limit of our present understanding.

Classification of $5$-Dimensional Complex Nilpotent Leibniz Algebras

Leibniz algebras are certain generalization of Lie algebras. In this paper we give the classification of $5-$dimensional complex non-Lie nilpotent Leibniz algebras. We use the canonical forms for the congruence classes of matrices of bilinear forms to classify the case $\dim(A^2)=3$ and $\dim(Leib(A))=1$ which can be applied to higher dimensions. The remaining cases are classified via direct method.

Haptics of Screwing and Unscrewing for its Application in Smart Factories for Disassembly

Reconstruction of skilled humans sensation and control system often leads to a development of robust control for the robots. We are developing an unscrewing robot for the automated disassembly which requires a comprehensive control system, but unscrewing experiments with robots are often limited to several conditions. On the contrary, humans typically have a broad range of screwing experiences and sensations throughout their lives, and we conducted an experiment to find these haptic patterns. Results show that people apply axial force to the screws to avoid screwdriver slippage (cam-outs), which is one of the key problems during screwing and unscrewing, and this axial force is proportional to the torque which is required for screwing. We have found that type of the screw head influences the amount of axial force applied. Using this knowledge an unscrewing robot for the smart disassembly factory RecyBot is developed, and experiments confirm the optimality of the strategy, used by humans. Finally, a methodology for robust unscrewing algorithm design is presented as a generalization of the findings. It can seriously speed up the development of the screwing and unscrewing robots and tools.

$G$-invariant Szegö kernel asymptotics and CR reduction

Let $(X, T^{1,0}X)$ be a compact connected orientable CR manifold of dimension $2n+1$ with non-degenerate Levi curvature. Assume that $X$ admits a connected compact Lie group action $G$. Under certain natural assumptions about the group action $G$, we show that the $G$-invariant Szegö kernel for $(0,q)$ forms is a complex Fourier integral operator, smoothing away $\mu^{-1}(0)$ and there is a precise description of the singularity near $\mu^{-1}(0)$, where $\mu$ denotes the CR moment map. We apply our result to the case when $X$ admits a transversal CR $S^1$ action and deduce an asymptotic expansion for the $m$-th Fourier component of the $G$-invariant Szegö kernel for $(0,q)$ forms as $m \to+\infty$. As an application, we show that if $m$ large enough, quantization commutes with reduction.

Self-similar resistive circuits as fractal-like structures

In the present work we explore resistive circuits where the individual resistors are arranged in fractal-like patterns. These circuits have some of the characteristics typically found in geometric fractals, namely self-similarity and scale invariance. Considering resistive circuits as graphs, we propose a definition of self-similar circuits which mimics a self-similar fractal. General properties of the resistive circuits generated by this approach are investigated, and interesting examples are commented in detail. Specifically, we consider self-similar resistive series, tree-like resistive networks and Sierpinski's configurations with resistors.

DynaPhoPy: A code for extracting phonon quasiparticles from molecular dynamics simulations

We have developed a computational code, DynaPhoPy, that allow us to extract the microscopic anharmonic phonon properties from molecular dynamics (MD) simulations using the normal-mode-decomposition technique as presented by Sun et al. [T. Sun, D. Zhang, R. Wentzcovitch, 2014]. Using this code we calculated the quasiparticle phonon frequencies and linewidths of crystalline silicon at different temperatures using both of first-principles and the Tersoff empirical potential approaches. In this work we show the dependence of these properties on the temperature using both approaches and compare them with reported experimental data obtained by Raman spectroscopy [M. Balkanski, R. Wallis, E. Haro, 1983 and R. Tsu, J. G. Hernandez, 1982].

Lower Bounds for Two-Sample Structural Change Detection in Ising and Gaussian Models

The change detection problem is to determine if the Markov network structures of two Markov random fields differ from one another given two sets of samples drawn from the respective underlying distributions. We study the trade-off between the sample sizes and the reliability of change detection, measured as a minimax risk, for the important cases of the Ising models and the Gaussian Markov random fields restricted to the models which have network structures with $p$ nodes and degree at most $d$, and obtain information-theoretic lower bounds for reliable change detection over these models. We show that for the Ising model, $\Omega\left(\frac{d^2}{(\log d)^2}\log p\right)$ samples are required from each dataset to detect even the sparsest possible changes, and that for the Gaussian, $\Omega\left( \gamma^{-2} \log(p)\right)$ samples are required from each dataset to detect change, where $\gamma$ is the smallest ratio of off-diagonal to diagonal terms in the precision matrices of the distributions. These bounds are compared to the corresponding results in structure learning, and closely match them under mild conditions on the model parameters. Thus, our change detection bounds inherit partial tightness from the structure learning schemes in previous literature, demonstrating that in certain parameter regimes, the naive structure learning based approach to change detection is minimax optimal up to constant factors.

How Usable are Rust Cryptography APIs?

Context: Poor usability of cryptographic APIs is a severe source of vulnerabilities. Aim: We wanted to find out what kind of cryptographic libraries are present in Rust and how usable they are. Method: We explored Rust's cryptographic libraries through a systematic search, conducted an exploratory study on the major libraries and a controlled experiment on two of these libraries with 28 student participants. Results: Only half of the major libraries explicitly focus on usability and misuse resistance, which is reflected in their current APIs. We found that participants were more successful using rust-crypto which we considered less usable than ring before the experiment. Conclusion: We discuss API design insights and make recommendations for the design of crypto libraries in Rust regarding the detail and structure of the documentation, higher-level APIs as wrappers for the existing low-level libraries, and selected, good-quality example code to improve the emerging cryptographic libraries of Rust.

Augmenting End-to-End Dialog Systems with Commonsense Knowledge

Building dialog agents that can converse naturally with humans is a challenging yet intriguing problem of artificial intelligence. In open-domain human-computer conversation, where the conversational agent is expected to respond to human responses in an interesting and engaging way, commonsense knowledge has to be integrated into the model effectively. In this paper, we investigate the impact of providing commonsense knowledge about the concepts covered in the dialog. Our model represents the first attempt to integrating a large commonsense knowledge base into end-to-end conversational models. In the retrieval-based scenario, we propose the Tri-LSTM model to jointly take into account message and commonsense for selecting an appropriate response. Our experiments suggest that the knowledge-augmented models are superior to their knowledge-free counterparts in automatic evaluation.

A framework for Multi-A(rmed)/B(andit) testing with online FDR control

We propose an alternative framework to existing setups for controlling false alarms when multiple A/B tests are run over time. This setup arises in many practical applications, e.g. when pharmaceutical companies test new treatment options against control pills for different diseases, or when internet companies test their default webpages versus various alternatives over time. Our framework proposes to replace a sequence of A/B tests by a sequence of best-arm MAB instances, which can be continuously monitored by the data scientist. When interleaving the MAB tests with an an online false discovery rate (FDR) algorithm, we can obtain the best of both worlds: low sample complexity and any time online FDR control. Our main contributions are: (i) to propose reasonable definitions of a null hypothesis for MAB instances; (ii) to demonstrate how one can derive an always-valid sequential p-value that allows continuous monitoring of each MAB test; and (iii) to show that using rejection thresholds of online-FDR algorithms as the confidence levels for the MAB algorithms results in both sample-optimality, high power and low FDR at any point in time. We run extensive simulations to verify our claims, and also report results on real data collected from the New Yorker Cartoon Caption contest.

Statistical Speech Enhancement Based on Probabilistic Integration of Variational Autoencoder and Non-Negative Matrix Factorization

This paper presents a statistical method of single-channel speech enhancement that uses a variational autoencoder (VAE) as a prior distribution on clean speech. A standard approach to speech enhancement is to train a deep neural network (DNN) to take noisy speech as input and output clean speech. Although this supervised approach requires a very large amount of pair data for training, it is not robust against unknown environments. Another approach is to use non-negative matrix factorization (NMF) based on basis spectra trained on clean speech in advance and those adapted to noise on the fly. This semi-supervised approach, however, causes considerable signal distortion in enhanced speech due to the unrealistic assumption that speech spectrograms are linear combinations of the basis spectra. Replacing the poor linear generative model of clean speech in NMF with a VAE---a powerful nonlinear deep generative model---trained on clean speech, we formulate a unified probabilistic generative model of noisy speech. Given noisy speech as observed data, we can sample clean speech from its posterior distribution. The proposed method outperformed the conventional DNN-based method in unseen noisy environments.

BPGrad: Towards Global Optimality in Deep Learning via Branch and Pruning

Understanding the global optimality in deep learning (DL) has been attracting more and more attention recently. Conventional DL solvers, however, have not been developed intentionally to seek for such global optimality. In this paper we propose a novel approximation algorithm, BPGrad, towards optimizing deep models globally via branch and pruning. Our BPGrad algorithm is based on the assumption of Lipschitz continuity in DL, and as a result it can adaptively determine the step size for current gradient given the history of previous updates, wherein theoretically no smaller steps can achieve the global optimality. We prove that, by repeating such branch-and-pruning procedure, we can locate the global optimality within finite iterations. Empirically an efficient solver based on BPGrad for DL is proposed as well, and it outperforms conventional DL solvers such as Adagrad, Adadelta, RMSProp, and Adam in the tasks of object recognition, detection, and segmentation.

Lagrangian Flow Network approach to an open flow model

Concepts and tools from network theory, the so-called Lagrangian Flow Network framework, have been successfully used to obtain a coarse-grained description of transport by closed fluid flows. Here we explore the application of this methodology to open chaotic flows, and check it with numerical results for a model open flow, namely a jet with a localized wave perturbation. We find that network nodes with high values of out-degree and of finite-time entropy in the forward-in-time direction identify the location of the chaotic saddle and its stable manifold, whereas nodes with high in-degree and backwards finite-time entropy highlight the location of the saddle and its unstable manifold. The cyclic clustering coefficient, associated to the presence of periodic orbits, takes non-vanishing values at the location of the saddle itself.

Some new gradient estimates for two nonlinear parabolic equations under Ricci flow

In this paper, by maximum principle and cutoff function, we investigate gradient estimates for positive solutions to two nonlinear parabolic equations under Ricci flow. The related Harnack inequalities are deduced. An result about positive solutions on closed manifolds under Ricci flow is abtained. As applications, gradient estimates and Harnack inequalities for positive solutions to the heat equation under Ricci flow are derived. These results in the paper can be regard as generalizing the gradient estimates of Li-Yau, J. Y. Li, Hamilton and Li-Xu to the Ricci flow. Our results also improve the estimates of S. P. Liu and J. Sun to the nonlinear parabolic equation under Ricci flow.

On the pointwise iteration-complexity of a dynamic regularized ADMM with over-relaxation stepsize

In this paper, we extend the improved pointwise iteration-complexity result of a dynamic regularized alternating direction method of multipliers (ADMM) for a new stepsize domain. In this complexity analysis, the stepsize parameter can even be chosen in the interval $(0,2)$ instead of interval $(0,(1+\sqrt{5})/2)$. As usual, our analysis is established by interpreting this ADMM variant as an instance of a hybrid proximal extragradient framework applied to a specific monotone inclusion problem.

DeepBrain: Functional Representation of Neural In-Situ Hybridization Images for Gene Ontology Classification Using Deep Convolutional Autoencoders

This paper presents a novel deep learning-based method for learning a functional representation of mammalian neural images. The method uses a deep convolutional denoising autoencoder (CDAE) for generating an invariant, compact representation of in situ hybridization (ISH) images. While most existing methods for bio-imaging analysis were not developed to handle images with highly complex anatomical structures, the results presented in this paper show that functional representation extracted by CDAE can help learn features of functional gene ontology categories for their classification in a highly accurate manner. Using this CDAE representation, our method outperforms the previous state-of-the-art classification rate, by improving the average AUC from 0.92 to 0.98, i.e., achieving 75% reduction in error. The method operates on input images that were downsampled significantly with respect to the original ones to make it computationally feasible.

Nonlocal heat equations in the Heisenberg group

We study the following nonlocal diffusion equation in the Heisenberg group $\mathbb{H}_n$, \[ u_t(z,s,t)=J\ast u(z,s,t)-u(z,s,t), \] where $\ast$ denote convolution product and $J$ satisfies appropriated hypothesis. For the Cauchy problem we obtain that the asymptotic behavior of the solutions is the same form that the one for the heat equation in the Heisenberg group. To obtain this result we use the spherical transform related to the pair $(U(n),\mathbb{H}_n)$. Finally we prove that solutions of properly rescaled nonlocal Dirichlet problem converge uniformly to the solution of the corresponding Dirichlet problem for the classical heat equation in the Heisenberg group.

Motion Planning Networks

Fast and efficient motion planning algorithms are crucial for many state-of-the-art robotics applications such as self-driving cars. Existing motion planning methods such as RRT*, A*, and D*, become ineffective as their computational complexity increases exponentially with the dimensionality of the motion planning problem. To address this issue, we present a neural network-based novel planning algorithm which generates end-to-end collision-free paths irrespective of the obstacles' geometry. The proposed method, called MPNet (Motion Planning Network), comprises of a Contractive Autoencoder which encodes the given workspaces directly from a point cloud measurement, and a deep feedforward neural network which takes the workspace encoding, start and goal configuration, and generates end-to-end feasible motion trajectories for the robot to follow. We evaluate MPNet on multiple planning problems such as planning of a point-mass robot, rigid-body, and 7 DOF Baxter robot manipulators in various 2D and 3D environments. The results show that MPNet is not only consistently computationally efficient in all 2D and 3D environments but also show remarkable generalization to completely unseen environments. The results also show that computation time of MPNet consistently remains less than 1 second which is significantly lower than existing state-of-the-art motion planning algorithms. Furthermore, through transfer learning, the MPNet trained in one scenario (e.g., indoor living places) can also quickly adapt to new scenarios (e.g., factory floors) with a little amount of data.

Query Expansion Techniques for Information Retrieval: a Survey

With the ever increasing size of web, relevant information extraction on the Internet with a query formed by a few keywords has become a big challenge. To overcome this, query expansion (QE) plays a crucial role in improving the Internet searches, where the user's initial query is reformulated to a new query by adding new meaningful terms with similar significance. QE -- as part of information retrieval (IR) -- has long attracted researchers' attention. It has also become very influential in the field of personalized social document, Question Answering over Linked Data (QALD), and, Text Retrieval Conference (TREC) and REAL sets. This paper surveys QE techniques in IR from 1960 to 2017 with respect to core techniques, data sources used, weighting and ranking methodologies, user participation and applications (of QE techniques) -- bringing out similarities and differences.

Method for Aspect-Based Sentiment Annotation Using Rhetorical Analysis

This paper fills a gap in aspect-based sentiment analysis and aims to present a new method for preparing and analysing texts concerning opinion and generating user-friendly descriptive reports in natural language. We present a comprehensive set of techniques derived from Rhetorical Structure Theory and sentiment analysis to extract aspects from textual opinions and then build an abstractive summary of a set of opinions. Moreover, we propose aspect-aspect graphs to evaluate the importance of aspects and to filter out unimportant ones from the summary. Additionally, the paper presents a prototype solution of data flow with interesting and valuable results. The proposed method's results proved the high accuracy of aspect detection when applied to the gold standard dataset.

Commutativity of integral quasi-arithmetic means on measure spaces

Let $(X, \mathscr{L}, \lambda)$ and $(Y, \mathscr{M}, \mu)$ be finite measure spaces for which there exist $A \in \mathscr{L}$ and $B \in \mathscr{M}$ with $0 < \lambda(A) < \lambda(X)$ and $0 < \mu(B) < \mu(Y)$, and let $I\subseteq \mathbf{R}$ be a non-empty interval. We prove that, if $f$ and $g$ are continuous bijections $I \to \mathbf{R}^+$, then the equation $$ f^{-1}\!\left(\int_X f\!\left(g^{-1}\!\left(\int_Y g \circ h\;d\mu\right)\right)d \lambda\right)\! = g^{-1}\!\left(\int_Y g\!\left(f^{-1}\!\left(\int_X f \circ h\;d\lambda\right)\right)d \mu\right)$$ is satisfied by every $\mathscr{L} \otimes \mathscr{M}$-measurable simple function $h: X \times Y \to I$ if and only if $f=c g$ for some $c \in \mathbf{R}^+$ (it is easy to see that the equation is well posed). An analogous, but essentially different, result, with $f$ and $g$ replaced by continuous injections $I \to \mathbf R$ and $\lambda(X)=\mu(Y)=1$, was recently obtained in [Indag. Math. 27 (2016), 945-953].

Moderate Deviation for Random Elliptic PDEs with Small Noise

Partial differential equations with random inputs have become popular models to characterize physical systems with uncertainty coming from, e.g., imprecise measurement and intrinsic randomness. In this paper, we perform asymptotic rare event analysis for such elliptic PDEs with random inputs. In particular, we consider the asymptotic regime that the noise level converges to zero suggesting that the system uncertainty is low, but does exists. We develop sharp approximations of the probability of a large class of rare events.

Music Transformer

Music relies heavily on repetition to build structure and meaning. Self-reference occurs on multiple timescales, from motifs to phrases to reusing of entire sections of music, such as in pieces with ABA structure. The Transformer (Vaswani et al., 2017), a sequence model based on self-attention, has achieved compelling results in many generation tasks that require maintaining long-range coherence. This suggests that self-attention might also be well-suited to modeling music. In musical composition and performance, however, relative timing is critically important. Existing approaches for representing relative positional information in the Transformer modulate attention based on pairwise distance (Shaw et al., 2018). This is impractical for long sequences such as musical compositions since their memory complexity for intermediate relative information is quadratic in the sequence length. We propose an algorithm that reduces their intermediate memory requirement to linear in the sequence length. This enables us to demonstrate that a Transformer with our modified relative attention mechanism can generate minute-long compositions (thousands of steps, four times the length modeled in Oore et al., 2018) with compelling structure, generate continuations that coherently elaborate on a given motif, and in a seq2seq setup generate accompaniments conditioned on melodies. We evaluate the Transformer with our relative attention mechanism on two datasets, JSB Chorales and Piano-e-Competition, and obtain state-of-the-art results on the latter.

Residual-Based Detections and Unified Architecture for Massive MIMO Uplink

Massive multiple-input multiple-output (M-MIMO) technique brings better energy efficiency and coverage but higher computational complexity than small-scale MIMO. For linear detections such as minimum mean square error (MMSE), prohibitive complexity lies in solving large-scale linear equations. For a better trade-off between bit-error-rate (BER) performance and computational complexity, iterative linear algorithms like conjugate gradient (CG) have been applied and have shown their feasibility in recent years. In this paper, residual-based detection (RBD) algorithms are proposed for M-MIMO detection, including minimal residual (MINRES) algorithm, generalized minimal residual (GMRES) algorithm, and conjugate residual (CR) algorithm. RBD algorithms focus on the minimization of residual norm per iteration, whereas most existing algorithms focus on the approximation of exact signal. Numerical results have shown that, for $64$-QAM $128\times 8$ MIMO, RBD algorithms are only $0.13$ dB away from the exact matrix inversion method when BER$=10^{-4}$. Stability of RBD algorithms has also been verified in various correlation conditions. Complexity comparison has shown that, CR algorithm require $87\%$ less complexity than the traditional method for $128\times 60$ MIMO. The unified hardware architecture is proposed with flexibility, which guarantees a low-complexity implementation for a family of RBD M-MIMO detectors.

Sparsity constrained split feasibility for dose-volume constraints in inverse planning of intensity-modulated photon or proton therapy

A split feasibility formulation for the inverse problem of intensity-modulated radiation therapy (IMRT) treatment planning with dose-volume constraints (DVCs) included in the planning algorithm is presented. It involves a new type of sparsity constraint that enables the inclusion of a percentage-violation constraint in the model problem and its handling by continuous (as opposed to integer) methods. We propose an iterative algorithmic framework for solving such a problem by applying the feasibility-seeking CQ-algorithm of Byrne combined with the automatic relaxation method (ARM) that uses cyclic projections. Detailed implementation instructions are furnished. Functionality of the algorithm was demonstrated through the creation of an intensity-modulated proton therapy plan for a simple 2D C-shaped geometry and also for a realistic base-of-skull chordoma treatment site. Monte Carlo simulations of proton pencil beams of varying energy were conducted to obtain dose distributions for the 2D test case. A research release of the Pinnacle3 proton treatment planning system was used to extract pencil beam doses for a clinical base-of-skull chordoma case. In both cases the beamlet doses were calculated to satisfy dose-volume constraints according to our new algorithm. Examination of the dose-volume histograms following inverse planning with our algorithm demonstrated that it performed as intended. The application of our proposed algorithm to dose-volume constraint inverse planning was successfully demonstrated. Comparison with optimized dose distributions from the research release of the Pinnacle3 treatment planning system showed the algorithm could achieve equivalent or superior results.

On 2-Verma modules for quantum $\mathfrak{sl}_2$

In this paper we study the superalgebra $A_n$, introduced by the authors in previous work on categorification of Verma modules for quantum $\mathfrak{sl}_2$. The superalgebra $A_n$ is akin to the nilHecke algebra, and shares similar properties. In particular, we prove a uniqueness result about 2-Verma modules on $\Bbbk$-linear 2-categories.

On the Number of Single-Peaked Narcissistic or Single-Crossing Narcissistic Preference Profiles

We investigate preference profiles for a set $\mathcal{V}$ of voters, where each voter $i$ has a preference order $\succ_i$ on a finite set $A$ of alternatives (that is, a linear order on $A$) such that for each two alternatives $a,b\in A$, voter $i$ prefers $a$ to $b$ if $a\succ_i b$. Such a profile is narcissistic if each alternative $a$ is preferred the most by at least one voter. It is single-peaked if there is a linear order $\triangleright^{\text{sp}}$ on the alternatives such that each voter's preferences on the alternatives along the order $\triangleright^{\text{sp}}$ are either strictly increasing, or strictly decreasing, or first strictly increasing and then strictly decreasing. It is single-crossing if there is a linear order $\triangleright^{\text{sc}}$ on the voters such that each pair of alternatives divides the order $\triangleright^{\text{sc}}$ into at most two suborders, where in each suborder, all voters have the same linear order on this pair. We show that for $n$ voters and $n$ alternatives,the number of single-peaked narcissistic profiles is $\prod_{i=2}^{n-1} \binom{n-1}{i-1}$ while the number of single-crossing narcissistic profiles is $2^{\binom{n-1}{2}}$.

Planet formation and disk-planet interactions

This review is based on lectures given at the 45th Saas-Fee Advanced Course 'From Protoplanetary Disks to Planet Formation' held in March 2015 in Les Diablerets, Switzerland. Starting with an overview of the main characterictics of the Solar System and extrasolar planets, we describe the planet formation process in terms of the sequential accretion scenario. First the growth processes of dust particles to planetesimals and subsequently to terrestrial planets or planetary cores are presented. This is followed by the formation process of the giant planets either by core accretion or gravitational instability. Finally, the dynamical evolution of the orbital elements as driven by disk-planet interaction and the overall evolution of multi-object systems is presented.

Cherednik algebras and Calogero-Moser cells

Using the representation theory of Cherednik algebras at $t=0$ and a Galois covering of the Calogero-Moser space, we define the notions of left, right and two-sided Calogero-Moser cells for any finite complex reflection group. To each Caloger-Moser two-sided cell is associated a Calogero-Moser family, while to each Calogero-Moser left cell is associated a Calogero-Moser cellular representation. We study properties of these objects and we conjecture that, whenever the reflection group is real (i.e. is a Coxeter group), these notions coincide with the one of Kazhdan-Lusztig left, right and two-sided cells, Kazhdan-Lusztig families and Kazhdan-Lusztig cellular representations.

Khovanov complexes of rational tangles

We show that the Khovanov complex of a rational tangle has a very simple representative whose backbone of non-zero morphisms forms a zig-zag. Furthermore, this minimal complex can be computed quickly by an inductive algorithm. (For example, we calculate $Kh(8_2)$ by hand.) We find that the bigradings of the subobjects in these minimal complexes can be described by matrix actions, which after a change of basis is the reduced Burau representation of $B_3$.

Asymmetric Connectedness of Fears in the U.S. Financial Sector

We study how shocks to the forward-looking expectations of investors buying call and put options transmit across the financial system. We introduce a new contagion measure, called asymmetric fear connectedness (AFC), which captures the information related to "fear" on the two sides of the options market and can be used as a forward-looking systemic risk monitoring tool. The decomposed connectedness measures provide timely predictive information for near-future macroeconomic conditions and uncertainty indicators, and they contain additional valuable information that is not included in the aggregate connectedness measure. The role of a positive/negative "fear" transmitter/receiver emerges clearly when we focus more closely on idiosyncratic events for financial institutions. We identify banks that are predominantly positive/negative receivers of "fear", as well as banks that positively/negatively transmit "fear" in the financial system.

Neural Network Based Speaker Classification and Verification Systems with Enhanced Features

This work presents a novel framework based on feed-forward neural network for text-independent speaker classification and verification, two related systems of speaker recognition. With optimized features and model training, it achieves 100% classification rate in classification and less than 6% Equal Error Rate (ERR), using merely about 1 second and 5 seconds of data respectively. Features with stricter Voice Active Detection (VAD) than the regular one for speech recognition ensure extracting stronger voiced portion for speaker recognition, speaker-level mean and variance normalization helps to eliminate the discrepancy between samples from the same speaker. Both are proven to improve the system performance. In building the neural network speaker classifier, the network structure parameters are optimized with grid search and dynamically reduced regularization parameters are used to avoid training terminated in local minimum. It enables the training goes further with lower cost. In speaker verification, performance is improved with prediction score normalization, which rewards the speaker identity indices with distinct peaks and penalizes the weak ones with high scores but more competitors, and speaker-specific thresholding, which significantly reduces ERR in the ROC curve. TIMIT corpus with 8K sampling rate is used here. First 200 male speakers are used to train and test the classification performance. The testing files of them are used as in-domain registered speakers, while data from the remaining 126 male speakers are used as out-of-domain speakers, i.e. imposters in speaker verification.

Slimness of graphs

Slimness of a graph measures the local deviation of its metric from a tree metric. In a graph $G=(V,E)$, a geodesic triangle $\bigtriangleup(x,y,z)$ with $x, y, z\in V$ is the union $P(x,y) \cup P(x,z) \cup P(y,z)$ of three shortest paths connecting these vertices. A geodesic triangle $\bigtriangleup(x,y,z)$ is called $\delta$-slim if for any vertex $u\in V$ on any side $P(x,y)$ the distance from $u$ to $P(x,z) \cup P(y,z)$ is at most $\delta$, i.e. each path is contained in the union of the $\delta$-neighborhoods of two others. A graph $G$ is called $\delta$-slim, if all geodesic triangles in $G$ are $\delta$-slim. The smallest value $\delta$ for which $G$ is $\delta$-slim is called the slimness of $G$. In this paper, using the layering partition technique, we obtain sharp bounds on slimness of such families of graphs as (1) graphs with cluster-diameter $\Delta(G)$ of a layering partition of $G$, (2) graphs with tree-length $\lambda$, (3) graphs with tree-breadth $\rho$, (4) $k$-chordal graphs, AT-free graphs and HHD-free graphs. Additionally, we show that the slimness of every 4-chordal graph is at most 2 and characterize those 4-chordal graphs for which the slimness of every of its induced subgraph is at most 1.

SirenAttack: Generating Adversarial Audio for End-to-End Acoustic Systems

Despite their immense popularity, deep learning-based acoustic systems are inherently vulnerable to adversarial attacks, wherein maliciously crafted audios trigger target systems to misbehave. In this paper, we present SirenAttack, a new class of attacks to generate adversarial audios. Compared with existing attacks, SirenAttack highlights with a set of significant features: (i) versatile -- it is able to deceive a range of end-to-end acoustic systems under both white-box and black-box settings; (ii) effective -- it is able to generate adversarial audios that can be recognized as specific phrases by target acoustic systems; and (iii) stealthy -- it is able to generate adversarial audios indistinguishable from their benign counterparts to human perception. We empirically evaluate SirenAttack on a set of state-of-the-art deep learning-based acoustic systems (including speech command recognition, speaker recognition and sound event classification), with results showing the versatility, effectiveness, and stealthiness of SirenAttack. For instance, it achieves 99.45% attack success rate on the IEMOCAP dataset against the ResNet18 model, while the generated adversarial audios are also misinterpreted by multiple popular ASR platforms, including Google Cloud Speech, Microsoft Bing Voice, and IBM Speech-to-Text. We further evaluate three potential defense methods to mitigate such attacks, including adversarial training, audio downsampling, and moving average filtering, which leads to promising directions for further research.

k-server via multiscale entropic regularization

We present an $O((\log k)^2)$-competitive randomized algorithm for the $k$-server problem on hierarchically separated trees (HSTs). This is the first $o(k)$-competitive randomized algorithm for which the competitive ratio is independent of the size of the underlying HST. Our algorithm is designed in the framework of online mirror descent where the mirror map is a multiscale entropy. When combined with Bartal's static HST embedding reduction, this leads to an $O((\log k)^2 \log n)$-competitive algorithm on any $n$-point metric space. We give a new dynamic HST embedding that yields an $O((\log k)^3 \log \Delta)$-competitive algorithm on any metric space where the ratio of the largest to smallest non-zero distance is at most $\Delta$.

Some Large Sample Results for the Method of Regularized Estimators

We present a general framework for studying regularized estimators; i.e., estimation problems wherein "plug-in" type estimators are either ill-defined or ill-behaved. We derive primitive conditions that imply consistency and asymptotic linear representation for regularized estimators, allowing for slower than $\sqrt{n}$ estimators as well as infinite dimensional parameters. We also provide data-driven methods for choosing tuning parameters that, under some conditions, achieve the aforementioned results. We illustrate the scope of our approach by studying a wide range of applications, revisiting known results and deriving new ones.

Coregionalised Locomotion Envelopes - A Qualitative Approach

'Sharing of statistical strength' is a phrase often employed in machine learning and signal processing. In sensor networks, for example, missing signals from certain sensors may be predicted by exploiting their correlation with observed signals acquired from other sensors. For humans, our hands move synchronously with our legs, and we can exploit these implicit correlations for predicting new poses and for generating new natural-looking walking sequences. We can also go much further and exploit this form of transfer learning, to develop new control schemas for robust control of rehabilitation robots. In this short paper we introduce coregionalised locomotion envelopes - a method for multi-dimensional manifold regression, on human locomotion variates. Herein we render a qualitative description of this method.

Optimal Invariant Tests in an Instrumental Variables Regression With Heteroskedastic and Autocorrelated Errors

This paper uses model symmetries in the instrumental variable (IV) regression to derive an invariant test for the causal structural parameter. Contrary to popular belief, we show there exist model symmetries when equation errors are heteroskedastic and autocorrelated (HAC). Our theory is consistent with existing results for the homoskedastic model (Andrews, Moreira and Stock(2006} and Chamberlain (2007}), but in general uses information on the structural parameter beyond the Anderson-Rubin, score, and rank statistics. This suggests that tests based only the Anderson-Rubin and score statistics discard information on the causal parameter of interest. We apply our theory to construct designs in which these tests indeed have power arbitrarily close to size. Other tests, including other adaptations to the CLR test, do not suffer the same deficiencies. Finally, we use the model symmetries to propose novel weighted-average power tests for the HAC-IV model.

Longitudinal electric field: from Maxwell equation to non-locality in time and space

In this paper we use the classical electrodynamics to show that the Lorenz gauge can be incompatible with some particular solutions of the d Alembert equations for electromagnetic potentials. In its turn, the d Alembert equations for the elec- tromagnetic potentials is the result of application of the Lorenz gauge to general equations for the potentials. The last ones is the straightforward consequence of Maxwell equations. Since the d Alembert equations and the electromagnetic poten- tials are necessary for quantum electrodynamics formulation, one should oblige to satisfy these equations also in classical case. The solution of d Alembert equations, which modifies longitudinal electric field is found. The requirement of this modifi- cation follows from the necessity to satisfy the physical condition of impossibility of instantaneous transferring of interaction in space.

Estimating linear functionals of a sparse family of Poisson means

Assume that we observe a sample of size n composed of p-dimensional signals, each signal having independent entries drawn from a scaled Poisson distribution with an unknown intensity. We are interested in estimating the sum of the n unknown intensity vectors, under the assumption that most of them coincide with a given 'background' signal. The number s of p-dimensional signals different from the background signal plays the role of sparsity and the goal is to leverage this sparsity assumption in order to improve the quality of estimation as compared to the naive estimator that computes the sum of the observed signals. We first introduce the group hard thresholding estimator and analyze its mean squared error measured by the squared Euclidean norm. We establish a nonasymptotic upper bound showing that the risk is at most of the order of {\sigma}^2(sp + s^2sqrt(p)) log^3/2(np). We then establish lower bounds on the minimax risk over a properly defined class of collections of s-sparse signals. These lower bounds match with the upper bound, up to logarithmic terms, when the dimension p is fixed or of larger order than s^2. In the case where the dimension p increases but remains of smaller order than s^2, our results show a gap between the lower and the upper bounds, which can be up to order sqrt(p).

Fabrication of porous microrings via laser printing and ion-beam post-etching

Pulsed-laser dry printing of noble-metal microrings with a tunable internal porous structure, which can be revealed via an ion-beam etching post-procedure, was demonstrated. Abundance and average size of the pores inside the microrings were shown to be tuned in a wide range by varying incident pulse energy and a nitrogen doping level controlled in the process of magnetron deposition of the gold film in the appropriate gaseous environment. The fabricated porous microrings were shown to provide many-fold near-field enhancement of incident electromagnetic fields, which was confirmed by mapping of the characteristic Raman band of a nanometer-thick covering layer of Rhodamine 6G dye molecules and supporting finite-difference time-domain calculations. The proposed laser printing/ion-beam etching approach is demonstrated to be a unique tool aimed at designing and fabricating multifunctional plasmonic structures and metasurfaces for spectroscopic bioidentification based on surface-enhanced infrared absorption, Raman scattering and photoluminescence detection schemes.

Performance Measurements of Supercomputing and Cloud Storage Solutions

Increasing amounts of data from varied sources, particularly in the fields of machine learning and graph analytics, are causing storage requirements to grow rapidly. A variety of technologies exist for storing and sharing these data, ranging from parallel file systems used by supercomputers to distributed block storage systems found in clouds. Relatively few comparative measurements exist to inform decisions about which storage systems are best suited for particular tasks. This work provides these measurements for two of the most popular storage technologies: Lustre and Amazon S3. Lustre is an open-source, high performance, parallel file system used by many of the largest supercomputers in the world. Amazon's Simple Storage Service, or S3, is part of the Amazon Web Services offering, and offers a scalable, distributed option to store and retrieve data from anywhere on the Internet. Parallel processing is essential for achieving high performance on modern storage systems. The performance tests used span the gamut of parallel I/O scenarios, ranging from single-client, single-node Amazon S3 and Lustre performance to a large-scale, multi-client test designed to demonstrate the capabilities of a modern storage appliance under heavy load. These results show that, when parallel I/O is used correctly (i.e., many simultaneous read or write processes), full network bandwidth performance is achievable and ranged from 10 gigabits/s over a 10 GigE S3 connection to 0.35 terabits/s using Lustre on a 1200 port 10 GigE switch. These results demonstrate that S3 is well-suited to sharing vast quantities of data over the Internet, while Lustre is well-suited to processing large quantities of data locally.

Multi-Objective Event-triggered Consensus of Linear Multi-agent Systems

This paper proposes a distributed consensus algorithm for linear event-based heterogeneous multi-agent systems (MAS). The proposed scheme is event-triggered in the sense that an agent selectively transmits its information within its local neighbourhood based on a directed network topology under the fulfillment of certain conditions. Using the Lyapunov stability theorem, the system constraints and event-triggering condition are expressed in terms of several linear matrix inequalities (LMIs) to derive the consensus parameters. The objective is to design the transmission threshold and minimum-norm heterogeneous control gains which collectively ensure an exponential consensus convergence rate for the closed-loop systems. The LMI computed control gains are robust to uncertainty with some deviation from their nominal values allowed. The practicability of the proposed event-based framework is further studied by proving the Zeno behaviour exclusion. Numerical simulations quantify the advantages of our event-triggered consensus approach in second-order, linear and heterogeneous multi-agent systems.

Private and Secure Coordination of Match-Making for Heavy-Duty Vehicle Platooning

A secure and private framework for inter-agent communication and coordination is developed. This allows an agent, in our case a fleet owner, to ask questions or submit queries in an encrypted fashion using semi-homomorphic encryption. The submitted query can be about the interest of the other fleet owners for using a road at a specific time of the day, for instance, for the purpose of collaborative vehicle platooning. The other agents can then provide appropriate responses without knowing the content of the questions or the queries. Strong privacy and security guarantees are provided for the agent who is submitting the queries. It is also shown that the amount of the information that this agent can extract from the other agent is bounded. In fact, with submitting one query, a sophisticated agent can at most extract the answer to two queries. This secure communication platform is used subsequently to develop a distributed coordination mechanisms among fleet owners.

Kernel Robust Bias-Aware Prediction under Covariate Shift

Under covariate shift, training (source) data and testing (target) data differ in input space distribution, but share the same conditional label distribution. This poses a challenging machine learning task. Robust Bias-Aware (RBA) prediction provides the conditional label distribution that is robust to the worstcase logarithmic loss for the target distribution while matching feature expectation constraints from the source distribution. However, employing RBA with insufficient feature constraints may result in high certainty predictions for much of the source data, while leaving too much uncertainty for target data predictions. To overcome this issue, we extend the representer theorem to the RBA setting, enabling minimization of regularized expected target risk by a reweighted kernel expectation under the source distribution. By applying kernel methods, we establish consistency guarantees and demonstrate better performance of the RBA classifier than competing methods on synthetically biased UCI datasets as well as datasets that have natural covariate shift.

Degeneration in VAE: in the Light of Fisher Information Loss

While enormous progress has been made to Variational Autoencoder (VAE) in recent years, similar to other deep networks, VAE with deep networks suffers from the problem of degeneration, which seriously weakens the correlation between the input and the corresponding latent codes, deviating from the goal of the representation learning. To investigate how degeneration affects VAE from a theoretical perspective, we illustrate the information transmission in VAE and analyze the intermediate layers of the encoders/decoders. Specifically, we propose a Fisher Information measure for the layer-wise analysis. With such measure, we demonstrate that information loss is ineluctable in feed-forward networks and causes the degeneration in VAE. We show that skip connections in VAE enable the preservation of information without changing the model architecture. We call this class of VAE equipped with skip connections as SCVAE and perform a range of experiments to show its advantages in information preservation and degeneration mitigation.

Room-temperature spin transport in n-Ge probed by four-terminal nonlocal measurements

We demonsrtate electrical spin injection and detection in $n$-type Ge ($n$-Ge) at room temperature using four-terminal nonlocal spin-valve and Hanle-effect measurements in lateral spin-valve (LSV) devices with Heusler-alloy Schottky tunnel contacts. The spin diffusion length ($\lambda$$_{\rm Ge}$) of the Ge layer used ($n \sim$ 1 $\times$ 10$^{19}$ cm$^{-3}$) at 296 K is estimated to be $\sim$ 0.44 $\pm$ 0.02 $\mu$m. Room-temperature spin signals can be observed reproducibly at the low bias voltage range ($\le$ 0.7 V) for LSVs with relatively low resistance-area product ($RA$) values ($\le$ 1 k$\Omega$$\mu$m$^{2}$). This means that the Schottky tunnel contacts used here are more suitable than ferromagnet/MgO tunnel contacts ($RA \ge$ 100 k$\Omega$$\mu$m$^{2}$) for developing Ge spintronic applications.

Synthesizing Neural Network Controllers with Probabilistic Model based Reinforcement Learning

We present an algorithm for rapidly learning controllers for robotics systems. The algorithm follows the model-based reinforcement learning paradigm, and improves upon existing algorithms; namely Probabilistic learning in Control (PILCO) and a sample-based version of PILCO with neural network dynamics (Deep-PILCO). We propose training a neural network dynamics model using variational dropout with truncated Log-Normal noise. This allows us to obtain a dynamics model with calibrated uncertainty, which can be used to simulate controller executions via rollouts. We also describe set of techniques, inspired by viewing PILCO as a recurrent neural network model, that are crucial to improve the convergence of the method. We test our method on a variety of benchmark tasks, demonstrating data-efficiency that is competitive with PILCO, while being able to optimize complex neural network controllers. Finally, we assess the performance of the algorithm for learning motor controllers for a six legged autonomous underwater vehicle. This demonstrates the potential of the algorithm for scaling up the dimensionality and dataset sizes, in more complex control tasks.

Stealthy Deception Attacks Against SCADA Systems

SCADA protocols for Industrial Control Systems (ICS) are vulnerable to network attacks such as session hijacking. Hence, research focuses on network anomaly detection based on meta--data (message sizes, timing, command sequence), or on the state values of the physical process. In this work we present a class of semantic network-based attacks against SCADA systems that are undetectable by the above mentioned anomaly detection. After hijacking the communication channels between the Human Machine Interface (HMI) and Programmable Logic Controllers (PLCs), our attacks cause the HMI to present a fake view of the industrial process, deceiving the human operator into taking manual actions. Our most advanced attack also manipulates the messages generated by the operator's actions, reversing their semantic meaning while causing the HMI to present a view that is consistent with the attempted human actions. The attacks are totaly stealthy because the message sizes and timing, the command sequences, and the data values of the ICS's state all remain legitimate. We implemented and tested several attack scenarios in the test lab of our local electric company, against a real HMI and real PLCs, separated by a commercial-grade firewall. We developed a real-time security assessment tool, that can simultaneously manipulate the communication to multiple PLCs and cause the HMI to display a coherent system--wide fake view. Our tool is configured with message-manipulating rules written in an ICS Attack Markup Language (IAML) we designed, which may be of independent interest. Our semantic attacks all successfully fooled the operator and brought the system to states of blackout and possible equipment damage.

Beta-rhythm oscillations and synchronization transition in network models of Izhikevich neurons: effect of topology and synaptic type

Despite their significant functional roles, beta-band oscillations are least understood. Synchronization in neuronal networks have attracted much attention in recent years with the main focus on transition type. Whether one obtains explosive transition or a continuous transition is an important feature of the neuronal network which can depend on network structure as well as synaptic types. In this study we consider the effect of synaptic interaction (electrical and chemical) as well as structural connectivity on synchronization transition in network models of Izhikevich neurons which spike regularly with beta rhythms. We find a wide range of behavior including continuous transition, explosive transition, as well as lack of global order. The stronger electrical synapses are more conducive to synchronization and can even lead to explosive synchronization. The key network element which determines the order of transition is found to be the clustering coefficient and not the small world effect, or the existence of hubs in a network. These results are in contrast to previous results which use phase oscillator models such as the Kuramoto model. Furthermore, we show that the patterns of synchronization changes when one goes to the gamma band. We attribute such a change to the change in the refractory period of Izhikevich neurons which changes significantly with frequency.

Renyi Differential Privacy

We propose a natural relaxation of differential privacy based on the Renyi divergence. Closely related notions have appeared in several recent papers that analyzed composition of differentially private mechanisms. We argue that the useful analytical tool can be used as a privacy definition, compactly and accurately representing guarantees on the tails of the privacy loss. We demonstrate that the new definition shares many important properties with the standard definition of differential privacy, while additionally allowing tighter analysis of composite heterogeneous mechanisms.

Enhancing Quality for VVC Compressed Videos by Jointly Exploiting Spatial Details and Temporal Structure

In this paper, we propose a quality enhancement network for Versatile Video Coding (VVC) compressed videos by jointly exploiting spatial details and temporal structure (SDTS). The network consists of a temporal structure prediction subnet and a spatial detail enhancement subnet. The former subnet is used to estimate and compensate the temporal motion across frames, and the spatial detail subnet is used to reduce the compression artifacts and enhance the reconstruction quality of the VVC compressed video. Experimental results demonstrate the effectiveness of our SDTS-based approach. It offers over 7.82$\%$ BD-rate saving on the common test video sequences and achieves the state-of-the-art performance.