For Convolutional Neural Network-based object detection, there is a typical dilemma: the spatial information is well kept in the shallow layers which unfortunately do not have enough semantic information, while the deep layers have a high semantic concept but lost a lot of spatial information, resulting in serious information imbalance. To acquire enough semantic information for shallow layers, Feature Pyramid Networks (FPN) is used to build a top-down propagated path. In this paper, except for top-down combining of information for shallow layers, we propose a novel network called Image Pyramid Guidance Network (IPG-Net) to make sure both the spatial information and semantic information are abundant for each layer. Our IPG-Net has two main parts: the image pyramid guidance transformation module and the image pyramid guidance fusion module. Our main idea is to introduce the image pyramid guidance into the backbone stream to solve the information imbalance problem, which alleviates the vanishment of the small object features. This IPG transformation module promises even in the deepest stage of the backbone, there is enough spatial information for bounding box regression and classification. Furthermore, we designed an effective fusion module to fuse the features from the image pyramid and features from the backbone stream. We have tried to apply this novel network to both one-stage and two-stage detection models, state of the art results are obtained on the most popular benchmark data sets, i.e. MS COCO and Pascal VOC.

This paper considers the multi-group multicast beamforming optimization problem, for which the optimal solution has been unknown due to the non-convex and NP-hard nature of the problem. By utilizing the successive convex approximation numerical method and Lagrangian duality, we obtain the optimal multicast beamforming solution structure for both the quality-of-service (QoS) problem and the max-min fair (MMF) problem. The optimal structure brings valuable insights into multicast beamforming: We show that the notion of uplink-downlink duality can be generalized to the multicast beamforming problem. The optimal multicast beamformer is a weighted MMSE filter based on a group-channel direction: a generalized version of the optimal downlink multi-user unicast beamformer. We also show that there is an inherent low-dimensional structure in the optimal multicast beamforming solution independent of the number of transmit antennas, leading to efficient numerical algorithm design, especially for systems with large antenna arrays. We propose efficient algorithms to compute the multicast beamformer based on the optimal beamforming structure. Through asymptotic analysis, we characterize the asymptotic behavior of the multicast beamformers as the number of antennas grows, and in turn, provide simple closed-form approximate multicast beamformers for both the QoS and MMF problems. This approximation offers practical multicast beamforming solutions with a near-optimal performance at very low computational complexity for large-scale antenna systems.

Reading comprehension (RC) has been studied in a variety of datasets with the boosted performance brought by deep neural networks. However, the generalization capability of these models across different domains remains unclear. To alleviate this issue, we are going to investigate unsupervised domain adaptation on RC, wherein a model is trained on labeled source domain and to be applied to the target domain with only unlabeled samples. We first show that even with the powerful BERT contextual representation, the performance is still unsatisfactory when the model trained on one dataset is directly applied to another target dataset. To solve this, we provide a novel conditional adversarial self-training method (CASe). Specifically, our approach leverages a BERT model fine-tuned on the source dataset along with the confidence filtering to generate reliable pseudo-labeled samples in the target domain for self-training. On the other hand, it further reduces domain distribution discrepancy through conditional adversarial learning across domains. Extensive experiments show our approach achieves comparable accuracy to supervised models on multiple large-scale benchmark datasets.

Interaction languages such as MSC are often associated with formal semantics by means of translations into distinct behavioral formalisms such as automatas or Petri nets. In contrast to translational approaches we propose an operational approach. Its principle is to identify which elementary communication actions can be immediately executed, and then to compute, for every such action, a new interaction representing the possible continuations to its execution. We also define an algorithm for checking the validity of execution traces (i.e. whether or not they belong to an interaction's semantics). Algorithms for semantic computation and trace validity are analyzed by means of experiments.

Histopathology is a reflection of the molecular changes and provides prognostic phenotypes representing the disease progression. In this study, we introduced feature scores generated from hematoxylin and eosin histology images based on deep learning (DL) models developed for prostate pathology. We demonstrated that these feature scores were significantly prognostic for time to event endpoints (biochemical recurrence and cancer-specific survival) and had simultaneously molecular biologic associations to relevant genomic alterations and molecular subtypes using already trained DL models that were not previously exposed to the datasets of the current study. Further, we discussed the potential of such feature scores to improve the current tumor grading system and the challenges that are associated with tumor heterogeneity and the development of prognostic models from histology images. Our findings uncover the potential of feature scores from histology images as digital biomarkers in precision medicine and as an expanding utility for digital pathology.

Robots are required to operate autonomously in response to changing situations. Previously, imitation learning using 4ch-bilateral control was demonstrated to be suitable for imitation of object manipulation. However, cooperative work between humans and robots has not yet been verified in these studies. In this study, the task was expanded by cooperative work between a human and a robot. 4ch-bilateral control was used to collect training data for training robot motion. We focused on serving salad as a task in the home. The task was executed with a spoon and a fork fixed to robots. Adjustment of force was indispensable in manipulating indefinitely shaped objects such as salad. Results confirmed the effectiveness of the proposed method as demonstrated by the success of the task.

Recent beyond worst-case optimal join algorithms Minesweeper and its generalization Tetris have brought the theory of indexing and join processing together by developing a geometric framework for joins. These algorithms take as input an index $mathcal{B}$, referred to as a box cover, that stores output gaps that can be inferred from traditional indexes, such as B+ trees or tries, on the input relations. The performances of these algorithms highly depend on the certificate of $mathcal{B}$, which is the smallest subset of gaps in $mathcal{B}$ whose union covers all of the gaps in the output space of a query $Q$. We study how to generate box covers that contain small size certificates to guarantee efficient runtimes for these algorithms. First, given a query $Q$ over a set of relations of size $N$ and a fixed set of domain orderings for the attributes, we give a $ ilde{O}(N)$-time algorithm called GAMB which generates a box cover for $Q$ that is guaranteed to contain the smallest size certificate across any box cover for $Q$. Second, we show that finding a domain ordering to minimize the box cover size and certificate is NP-hard through a reduction from the 2 consecutive block minimization problem on boolean matrices. Our third contribution is a $ ilde{O}(N)$-time approximation algorithm called ADORA to compute domain orderings, under which one can compute a box cover of size $ ilde{O}(K^r)$, where $K$ is the minimum box cover for $Q$ under any domain ordering and $r$ is the maximum arity of any relation. This guarantees certificates of size $ ilde{O}(K^r)$. We combine ADORA and GAMB with Tetris to form a new algorithm we call TetrisReordered, which provides several new beyond worst-case bounds. On infinite families of queries, TetrisReordered's runtimes are unboundedly better than the bounds stated in prior work.

Traffic signs are essential map features globally in the era of autonomous driving and smart cities. To develop accurate and robust algorithms for traffic sign detection and classification, a large-scale and diverse benchmark dataset is required. In this paper, we introduce a traffic sign benchmark dataset of 100K street-level images around the world that encapsulates diverse scenes, wide coverage of geographical locations, and varying weather and lighting conditions and covers more than 300 manually annotated traffic sign classes. The dataset includes 52K images that are fully annotated and 48K images that are partially annotated. This is the largest and the most diverse traffic sign dataset consisting of images from all over world with fine-grained annotations of traffic sign classes. We have run extensive experiments to establish strong baselines for both the detection and the classification tasks. In addition, we have verified that the diversity of this dataset enables effective transfer learning for existing large-scale benchmark datasets on traffic sign detection and classification. The dataset is freely available for academic research: https://www.mapillary.com/dataset/trafficsign.

When time-dependent partial differential equations (PDEs) are solved numerically in a domain with curved boundary or on a curved surface, mesh error and geometric approximation error caused by the inaccurate location of vertices and other interior grid points, respectively, could be the main source of the inaccuracy and instability of the numerical solutions of PDEs. The role of these geometric errors in deteriorating the stability and particularly the conservation properties are largely unknown, which seems to necessitate very fine meshes especially to remove geometric approximation error. This paper aims to investigate the effect of geometric approximation error by using a high-order mesh with negligible geometric approximation error, even for high order polynomial of order p. To achieve this goal, the high-order mesh generator from CAD geometry called NekMesh is adapted for surface mesh generation in comparison to traditional meshes with non-negligible geometric approximation error. Two types of numerical tests are considered. Firstly, the accuracy of differential operators is compared for various p on a curved element of the sphere. Secondly, by applying the method of moving frames, four different time-dependent PDEs on the sphere are numerically solved to investigate the impact of geometric approximation error on the accuracy and conservation properties of high-order numerical schemes for PDEs on the sphere.

The central importance of large scale eigenvalue problems in scientific computation necessitates the development of massively parallel algorithms for their solution. Recent advances in dense numerical linear algebra have enabled the routine treatment of eigenvalue problems with dimensions on the order of hundreds of thousands on the world's largest supercomputers. In cases where dense treatments are not feasible, Krylov subspace methods offer an attractive alternative due to the fact that they do not require storage of the problem matrices. However, demonstration of scalability of either of these classes of eigenvalue algorithms on computing architectures capable of expressing massive parallelism is non-trivial due to communication requirements and serial bottlenecks, respectively. In this work, we introduce the SISLICE method: a parallel shift-invert algorithm for the solution of the symmetric self-consistent field (SCF) eigenvalue problem. The SISLICE method drastically reduces the communication requirement of current parallel shift-invert eigenvalue algorithms through various shift selection and migration techniques based on density of states estimation and k-means clustering, respectively. This work demonstrates the robustness and parallel performance of the SISLICE method on a representative set of SCF eigenvalue problems and outlines research directions which will be explored in future work.

Our starting point are register automata for data words, in the style of Kaminski and Francez. We study the effects of the single-use restriction, which says that a register is emptied immediately after being used. We show that under the single-use restriction, the theory of automata for data words becomes much more robust. The main results are: (a) five different machine models are equivalent as language acceptors, including one-way and two-way single-use register automata; (b) one can recover some of the algebraic theory of languages over finite alphabets, including a version of the Krohn-Rhodes Theorem; (c) there is also a robust theory of transducers, with four equivalent models, including two-way single use transducers and a variant of streaming string transducers for data words. These results are in contrast with automata for data words without the single-use restriction, where essentially all models are pairwise non-equivalent.

Noisy Intermediate-Scale Quantum (NISQ) computers are entering an era in which they can perform computational tasks beyond the capabilities of the most powerful classical computers, thereby achieving "Quantum Supremacy", a major milestone in quantum computing. NISQ Supremacy requires comparison with a state-of-the-art classical simulator. We report HPC simulations of hard random quantum circuits (RQC), which have been recently used as a benchmark for the first experimental demonstration of Quantum Supremacy, sustaining an average performance of 281 Pflop/s (true single precision) on Summit, currently the fastest supercomputer in the World. These simulations were carried out using qFlex, a tensor-network-based classical high-performance simulator of RQCs. Our results show an advantage of many orders of magnitude in energy consumption of NISQ devices over classical supercomputers. In addition, we propose a standard benchmark for NISQ computers based on qFlex.

The problem of sampling from the stationary distribution of a Markov chain finds widespread applications in a variety of fields. The time required for a Markov chain to converge to its stationary distribution is known as the classical mixing time. In this article, we deal with analog quantum algorithms for mixing. First, we provide an analog quantum algorithm that given a Markov chain, allows us to sample from its stationary distribution in a time that scales as the sum of the square root of the classical mixing time and the square root of the classical hitting time. Our algorithm makes use of the framework of interpolated quantum walks and relies on Hamiltonian evolution in conjunction with von Neumann measurements.

There also exists a different notion for quantum mixing: the problem of sampling from the limiting distribution of quantum walks, defined in a time-averaged sense. In this scenario, the quantum mixing time is defined as the time required to sample from a distribution that is close to this limiting distribution. Recently we provided an upper bound on the quantum mixing time for Erd"os-Renyi random graphs [Phys. Rev. Lett. 124, 050501 (2020)]. Here, we also extend and expand upon our findings therein. Namely, we provide an intuitive understanding of the state-of-the-art random matrix theory tools used to derive our results. In particular, for our analysis we require information about macroscopic, mesoscopic and microscopic statistics of eigenvalues of random matrices which we highlight here. Furthermore, we provide numerical simulations that corroborate our analytical findings and extend this notion of mixing from simple graphs to any ergodic, reversible, Markov chain.

The Hierarchical Equal Area isoLatitude Pixelation (HEALPix) scheme is used extensively in astrophysics for data collection and analysis on the sphere. The scheme was originally designed for studying the Cosmic Microwave Background (CMB) radiation, which represents the first light to travel during the early stages of the universe's development and gives the strongest evidence for the Big Bang theory to date. Refined analysis of the CMB angular power spectrum can lead to revolutionary developments in understanding the nature of dark matter and dark energy. In this paper, we present a new method for performing spherical harmonic analysis for HEALPix data, which is a central component to computing and analyzing the angular power spectrum of the massive CMB data sets. The method uses a novel combination of a non-uniform fast Fourier transform, the double Fourier sphere method, and Slevinsky's fast spherical harmonic transform (Slevinsky, 2019). For a HEALPix grid with $N$ pixels (points), the computational complexity of the method is $mathcal{O}(Nlog^2 N)$, with an initial set-up cost of $mathcal{O}(N^{3/2}log N)$. This compares favorably with $mathcal{O}(N^{3/2})$ runtime complexity of the current methods available in the HEALPix software when multiple maps need to be analyzed at the same time. Using numerical experiments, we demonstrate that the new method also appears to provide better accuracy over the entire angular power spectrum of synthetic data when compared to the current methods, with a convergence rate at least two times higher.

Finding a good regularization parameter for Tikhonov regularization problems is a though yet often asked question. One approach is to use leave-one-out cross-validation scores to indicate the goodness of fit. This utilizes only the noisy function values but, on the downside, comes with a high computational cost. In this paper we present a general approach to shift the main computations from the function in question to the node distribution and, making use of FFT and FFT-like algorithms, even reduce this cost tremendously to the cost of the Tikhonov regularization problem itself. We apply this technique in different settings on the torus, the unit interval, and the two-dimensional sphere. Given that the sampling points satisfy a quadrature rule our algorithm computes the cross-validations scores in floating-point precision. In the cases of arbitrarily scattered nodes we propose an approximating algorithm with the same complexity. Numerical experiments indicate the applicability of our algorithms.

Funding the production of quality online content is a pressing problem for content producers. The most common funding method, online advertising, is rife with well-known performance and privacy harms, and an intractable subject-agent conflict: many users do not want to see advertisements, depriving the site of needed funding.

Because of these negative aspects of advertisement-based funding, paywalls are an increasingly popular alternative for websites. This shift to a "pay-for-access" web is one that has potentially huge implications for the web and society. Instead of a system where information (nominally) flows freely, paywalls create a web where high quality information is available to fewer and fewer people, leaving the rest of the web users with less information, that might be also less accurate and of lower quality. Despite the potential significance of a move from an "advertising-but-open" web to a "paywalled" web, we find this issue understudied.

This work addresses this gap in our understanding by measuring how widely paywalls have been adopted, what kinds of sites use paywalls, and the distribution of policies enforced by paywalls. A partial list of our findings include that (i) paywall use is accelerating (2x more paywalls every 6 months), (ii) paywall adoption differs by country (e.g. 18.75% in US, 12.69% in Australia), (iii) paywalls change how users interact with sites (e.g. higher bounce rates, less incoming links), (iv) the median cost of an annual paywall access is $108 per site, and (v) paywalls are in general trivial to circumvent.

Finally, we present the design of a novel, automated system for detecting whether a site uses a paywall, through the combination of runtime browser instrumentation and repeated programmatic interactions with the site. We intend this classifier to augment future, longitudinal measurements of paywall use and behavior.

Asymptotic expansions are derived for eigenvalues produced by both the Crouzeix-Raviart element and the enriched Crouzeix--Raviart element. The expansions are optimal in the sense that extrapolation eigenvalues based on them admit a fourth order convergence provided that exact eigenfunctions are smooth enough. The major challenge in establishing the expansions comes from the fact that the canonical interpolation of both nonconforming elements lacks a crucial superclose property, and the nonconformity of both elements. The main idea is to employ the relation between the lowest-order mixed Raviart--Thomas element and the two nonconforming elements, and consequently make use of the superclose property of the canonical interpolation of the lowest-order mixed Raviart--Thomas element. To overcome the difficulty caused by the nonconformity, the commuting property of the canonical interpolation operators of both nonconforming elements is further used, which turns the consistency error problem into an interpolation error problem. Then, a series of new results are obtained to show the final expansions.

We develop a supervised machine learning algorithm that is able to learn topological phases for finite condensed matter systems from bulk data in real lattice space. The algorithm employs diagonalization in real space together with any supervised learning algorithm to learn topological phases through an eigenvector ensembling procedure. We combine our algorithm with decision trees and random forests to successfully recover topological phase diagrams of Su-Schrieffer-Heeger (SSH) models from bulk lattice data in real space and show how the Shannon information entropy of ensembles of lattice eigenvectors can be used to retrieve a signal detailing how topological information is distributed in the bulk. The discovery of Shannon information entropy signals associated with topological phase transitions from the analysis of data from several thousand SSH systems illustrates how model explainability in machine learning can advance the research of exotic quantum materials with properties that may power future technological applications such as qubit engineering for quantum computing.

Within the field of multilinear algebra, inverses and generalized inverses of tensors based on the Einstein product have been investigated over the past few years. In this paper, we explore the singular value decomposition and full-rank decomposition of arbitrary-order tensors using {it reshape} operation. Applying range and null space of tensors along with the reshape operation; we further study the Moore-Penrose inverse of tensors and their cancellation properties via the Einstein product. Then we discuss weighted Moore-Penrose inverses of arbitrary-order tensors using such product. Following a specific algebraic approach, a few characterizations and representations of these inverses are explored. In addition to this, we obtain a few necessary and sufficient conditions for the reverse-order law to hold for weighted Moore-Penrose inverses of arbitrary-order tensors.

A classical problem in appointment scheduling, with applications in health care, concerns the determination of the patients' arrival times that minimize a cost function that is a weighted sum of mean waiting times and mean idle times. One aspect of this problem is the sequencing problem, which focuses on ordering the patients. We assess the performance of the smallest-variance-first (SVF) rule, which sequences patients in order of increasing variance of their service durations. While it was known that SVF is not always optimal, it has been widely observed that it performs well in practice and simulation. We provide a theoretical justification for this observation by proving, in various settings, quantitative worst-case bounds on the ratio between the cost incurred by the SVF rule and the minimum attainable cost. We also show that, in great generality, SVF is asymptotically optimal, i.e., the ratio approaches 1 as the number of patients grows large. While evaluating policies by considering an approximation ratio is a standard approach in many algorithmic settings, our results appear to be the first of this type in the appointment scheduling literature.

We investigate sparse matrix bipartitioning -- a problem where we minimize the communication volume in parallel sparse matrix-vector multiplication. We prove, by reduction from graph bisection, that this problem is $mathcal{NP}$-complete in the case where each side of the bipartitioning must contain a linear fraction of the nonzeros.

We present an improved exact branch-and-bound algorithm which finds the minimum communication volume for a given matrix and maximum allowed imbalance. The algorithm is based on a maximum-flow bound and a packing bound, which extend previous matching and packing bounds.

We implemented the algorithm in a new program called MP (Matrix Partitioner), which solved 839 matrices from the SuiteSparse collection to optimality, each within 24 hours of CPU-time. Furthermore, MP solved the difficult problem of the matrix cage6 in about 3 days. The new program is on average more than ten times faster than the previous program MondriaanOpt.

Benchmark results using the set of 839 optimally solved matrices show that combining the medium-grain/iterative refinement methods of the Mondriaan package with the hypergraph bipartitioner of the PaToH package produces sparse matrix bipartitionings on average within 10% of the optimal solution.

Autonomous robot manipulation involves estimating the translation and orientation of the object to be manipulated as a 6-degree-of-freedom (6D) pose. Methods using RGB-D data have shown great success in solving this problem. However, there are situations where cost constraints or the working environment may limit the use of RGB-D sensors. When limited to monocular camera data only, the problem of object pose estimation is very challenging. In this work, we introduce a novel method called SilhoNet that predicts 6D object pose from monocular images. We use a Convolutional Neural Network (CNN) pipeline that takes in Region of Interest (ROI) proposals to simultaneously predict an intermediate silhouette representation for objects with an associated occlusion mask and a 3D translation vector. The 3D orientation is then regressed from the predicted silhouettes. We show that our method achieves better overall performance on the YCB-Video dataset than two state-of-the art networks for 6D pose estimation from monocular image input.

We design and implement the first private and anonymous decentralized crowdsourcing system ZebraLancer, and overcome two fundamental challenges of decentralizing crowdsourcing, i.e., data leakage and identity breach.

First, our outsource-then-prove methodology resolves the tension between the blockchain transparency and the data confidentiality to guarantee the basic utilities/fairness requirements of data crowdsourcing, thus ensuring: (i) a requester will not pay more than what data deserve, according to a policy announced when her task is published via the blockchain; (ii) each worker indeed gets a payment based on the policy, if he submits data to the blockchain; (iii) the above properties are realized not only without a central arbiter, but also without leaking the data to the open blockchain. Second, the transparency of blockchain allows one to infer private information about workers and requesters through their participation history. Simply enabling anonymity is seemingly attempting but will allow malicious workers to submit multiple times to reap rewards. ZebraLancer also overcomes this problem by allowing anonymous requests/submissions without sacrificing accountability. The idea behind is a subtle linkability: if a worker submits twice to a task, anyone can link the submissions, or else he stays anonymous and unlinkable across tasks. To realize this delicate linkability, we put forward a novel cryptographic concept, i.e., the common-prefix-linkable anonymous authentication. We remark the new anonymous authentication scheme might be of independent interest. Finally, we implement our protocol for a common image annotation task and deploy it in a test net of Ethereum. The experiment results show the applicability of our protocol atop the existing real-world blockchain.

A chordless cycle, or equivalently a hole, in a graph $G$ is an induced subgraph of $G$ which is a cycle of length at least $4$. We prove that the ErdH{o}s-P'osa property holds for chordless cycles, which resolves the major open question concerning the ErdH{o}s-P'osa property. Our proof for chordless cycles is constructive: in polynomial time, one can find either $k+1$ vertex-disjoint chordless cycles, or $c_1k^2 log k+c_2$ vertices hitting every chordless cycle for some constants $c_1$ and $c_2$. It immediately implies an approximation algorithm of factor $mathcal{O}(sf{opt}log {sf opt})$ for Chordal Vertex Deletion. We complement our main result by showing that chordless cycles of length at least $ell$ for any fixed $ellge 5$ do not have the ErdH{o}s-P'osa property.

This work deals with the efficient numerical solution of the time-fractional heat equation discretized on non-uniform temporal meshes. Non-uniform grids are essential to capture the singularities of "typical" solutions of time-fractional problems. We propose an efficient space-time multigrid method based on the waveform relaxation technique, which accounts for the nonlocal character of the fractional differential operator. To maintain an optimal complexity, which can be obtained for the case of uniform grids, we approximate the coefficient matrix corresponding to the temporal discretization by its hierarchical matrix (${cal H}$-matrix) representation. In particular, the proposed method has a computational cost of ${cal O}(k N M log(M))$, where $M$ is the number of time steps, $N$ is the number of spatial grid points, and $k$ is a parameter which controls the accuracy of the ${cal H}$-matrix approximation. The efficiency and the good convergence of the algorithm, which can be theoretically justified by a semi-algebraic mode analysis, are demonstrated through numerical experiments in both one- and two-dimensional spaces.

Dense kernel matrices $Theta in mathbb{R}^{N imes N}$ obtained from point evaluations of a covariance function $G$ at locations ${ x_{i} }_{1 leq i leq N} subset mathbb{R}^{d}$ arise in statistics, machine learning, and numerical analysis. For covariance functions that are Green's functions of elliptic boundary value problems and homogeneously-distributed sampling points, we show how to identify a subset $S subset { 1 , dots , N }^2$, with $# S = O ( N log (N) log^{d} ( N /epsilon ) )$, such that the zero fill-in incomplete Cholesky factorisation of the sparse matrix $Theta_{ij} 1_{( i, j ) in S}$ is an $epsilon$-approximation of $Theta$. This factorisation can provably be obtained in complexity $O ( N log( N ) log^{d}( N /epsilon) )$ in space and $O ( N log^{2}( N ) log^{2d}( N /epsilon) )$ in time, improving upon the state of the art for general elliptic operators; we further present numerical evidence that $d$ can be taken to be the intrinsic dimension of the data set rather than that of the ambient space. The algorithm only needs to know the spatial configuration of the $x_{i}$ and does not require an analytic representation of $G$. Furthermore, this factorization straightforwardly provides an approximate sparse PCA with optimal rate of convergence in the operator norm. Hence, by using only subsampling and the incomplete Cholesky factorization, we obtain, at nearly linear complexity, the compression, inversion and approximate PCA of a large class of covariance matrices. By inverting the order of the Cholesky factorization we also obtain a solver for elliptic PDE with complexity $O ( N log^{d}( N /epsilon) )$ in space and $O ( N log^{2d}( N /epsilon) )$ in time, improving upon the state of the art for general elliptic operators.

In the era of Internet of Things (IoT), Malware has been proliferating exponentially over the past decade. Traditional anti-virus software are ineffective against modern complex Malware. In order to address this challenge, researchers have proposed Hardware-assisted Malware Detection (HMD) using Hardware Performance Counters (HPCs). The HPCs are used to train a set of Machine learning (ML) classifiers, which in turn, are used to distinguish benign programs from Malware. Recently, adversarial attacks have been designed by introducing perturbations in the HPC traces using an adversarial sample predictor to misclassify a program for specific HPCs. These attacks are designed with the basic assumption that the attacker is aware of the HPCs being used to detect Malware. Since modern processors consist of hundreds of HPCs, restricting to only a few of them for Malware detection aids the attacker. In this paper, we propose a Moving target defense (MTD) for this adversarial attack by designing multiple ML classifiers trained on different sets of HPCs. The MTD randomly selects a classifier; thus, confusing the attacker about the HPCs or the number of classifiers applied. We have developed an analytical model which proves that the probability of an attacker to guess the perfect HPC-classifier combination for MTD is extremely low (in the range of $10^{-1864}$ for a system with 20 HPCs). Our experimental results prove that the proposed defense is able to improve the classification accuracy of HPC traces that have been modified through an adversarial sample generator by up to 31.5%, for a near perfect (99.4%) restoration of the original accuracy.

The standard training algorithm in neural machine translation (NMT) suffers from exposure bias, and alternative algorithms have been proposed to mitigate this. However, the practical impact of exposure bias is under debate. In this paper, we link exposure bias to another well-known problem in NMT, namely the tendency to generate hallucinations under domain shift. In experiments on three datasets with multiple test domains, we show that exposure bias is partially to blame for hallucinations, and that training with Minimum Risk Training, which avoids exposure bias, can mitigate this. Our analysis explains why exposure bias is more problematic under domain shift, and also links exposure bias to the beam search problem, i.e. performance deterioration with increasing beam size. Our results provide a new justification for methods that reduce exposure bias: even if they do not increase performance on in-domain test sets, they can increase model robustness to domain shift.

In this paper, we focus on the task of small-footprint keyword spotting under the far-field scenario. Far-field environments are commonly encountered in real-life speech applications, and it causes serve degradation of performance due to room reverberation and various kinds of noises. Our baseline system is built on the convolutional neural network trained with pooled data of both far-field and close-talking speech. To cope with the distortions, we adopt the multi-task learning scheme with alignment loss to reduce the mismatch between the embedding features learned from different domains of data. Experimental results show that our proposed method maintains the performance on close-talking speech and achieves significant improvement on the far-field test set.

There exist two notions of equivalence of behavior between states of a Labelled Markov Process (LMP): state bisimilarity and event bisimilarity. The first one can be considered as an appropriate generalization to continuous spaces of Larsen and Skou's probabilistic bisimilarity, while the second one is characterized by a natural logic. C. Zhou expressed state bisimilarity as the greatest fixed point of an operator $mathcal{O}$, and thus introduced an ordinal measure of the discrepancy between it and event bisimilarity. We call this ordinal the "Zhou ordinal" of $mathbb{S}$, $mathfrak{Z}(mathbb{S})$. When $mathfrak{Z}(mathbb{S})=0$, $mathbb{S}$ satisfies the Hennessy-Milner property. The second author proved the existence of an LMP $mathbb{S}$ with $mathfrak{Z}(mathbb{S}) geq 1$ and Zhou showed that there are LMPs having an infinite Zhou ordinal. In this paper we show that there are LMPs $mathbb{S}$ over separable metrizable spaces having arbitrary large countable $mathfrak{Z}(mathbb{S})$ and that it is consistent with the axioms of $mathit{ZFC}$ that there is such a process with an uncountable Zhou ordinal.

We perform an effectivization of classical results concerning universal coding and prediction for stationary ergodic processes over an arbitrary finite alphabet. That is, we lift the well-known almost sure statements to statements about Martin-L"of random sequences. Most of this work is quite mechanical but, by the way, we complete a result of Ryabko from 2008 by showing that each universal probability measure in the sense of universal coding induces a universal predictor in the prequential sense. Surprisingly, the effectivization of this implication holds true provided the universal measure does not ascribe too low conditional probabilities to individual symbols. As an example, we show that the Prediction by Partial Matching (PPM) measure satisfies this requirement. In the almost sure setting, the requirement is superfluous.

The accurate assembly of the system matrix is an important step in any code that solves partial differential equations on a mesh. We either explicitly set up a matrix, or we work in a matrix-free environment where we have to be able to quickly return matrix entries upon demand. Either way, the construction can become costly due to non-trivial material parameters entering the equations, multigrid codes requiring cascades of matrices that depend upon each other, or dynamic adaptive mesh refinement that necessitates the recomputation of matrix entries or the whole equation system throughout the solve. We propose that these constructions can be performed concurrently with the multigrid cycles. Initial geometric matrices and low accuracy integrations kickstart the multigrid, while improved assembly data is fed to the solver as and when it becomes available. The time to solution is improved as we eliminate an expensive preparation phase traditionally delaying the actual computation. We eliminate algorithmic latency. Furthermore, we desynchronise the assembly from the solution process. This anarchic increase of the concurrency level improves the scalability. Assembly routines are notoriously memory- and bandwidth-demanding. As we work with iteratively improving operator accuracies, we finally propose the use of a hierarchical, lossy compression scheme such that the memory footprint is brought down aggressively where the system matrix entries carry little information or are not yet available with high accuracy.

We introduce $Psimathrm{ec}$, a local spectral exterior calculus for the two-sphere $S^2$. $Psimathrm{ec}$ provides a discretization of Cartan's exterior calculus on $S^2$ formed by spherical differential $r$-form wavelets. These are well localized in space and frequency and provide (Stevenson) frames for the homogeneous Sobolev spaces $dot{H}^{-r+1}( Omega_{ u}^{r} , S^2 )$ of differential $r$-forms. At the same time, they satisfy important properties of the exterior calculus, such as the de Rahm complex and the Hodge-Helmholtz decomposition. Through this, $Psimathrm{ec}$ is tailored towards structure preserving discretizations that can adapt to solutions with varying regularity. The construction of $Psimathrm{ec}$ is based on a novel spherical wavelet frame for $L_2(S^2)$ that we obtain by introducing scalable reproducing kernel frames. These extend scalable frames to weighted sampling expansions and provide an alternative to quadrature rules for the discretization of needlet-like scale-discrete wavelets. We verify the practicality of $Psimathrm{ec}$ for numerical computations using the rotating shallow water equations. Our numerical results demonstrate that a $Psimathrm{ec}$-based discretization of the equations attains accuracy comparable to those of spectral methods while using a representation that is well localized in space and frequency.

Generative feature matching network (GFMN) is an approach for training implicit generative models for images by performing moment matching on features from pre-trained neural networks. In this paper, we present new GFMN formulations that are effective for sequential data. Our experimental results show the effectiveness of the proposed method, SeqGFMN, for three distinct generation tasks in English: unconditional text generation, class-conditional text generation, and unsupervised text style transfer. SeqGFMN is stable to train and outperforms various adversarial approaches for text generation and text style transfer.

We present a novel numerical scheme to approximate the solution map $smapsto u(s) := mathcal{L}^{-s}f$ to partial differential equations involving fractional elliptic operators. Reinterpreting $mathcal{L}^{-s}$ as interpolation operator allows us to derive an integral representation of $u(s)$ which includes solutions to parametrized reaction-diffusion problems. We propose a reduced basis strategy on top of a finite element method to approximate its integrand. Unlike prior works, we deduce the choice of snapshots for the reduced basis procedure analytically. Avoiding further discretization, the integral is interpreted in a spectral setting to evaluate the surrogate directly. Its computation boils down to a matrix approximation $L$ of the operator whose inverse is projected to a low-dimensional space, where explicit diagonalization is feasible. The universal character of the underlying $s$-independent reduced space allows the approximation of $(u(s))_{sin(0,1)}$ in its entirety. We prove exponential convergence rates and confirm the analysis with a variety of numerical examples.

Further improvements are proposed in the second part of this investigation to avoid inversion of $L$. Instead, we directly project the matrix to the reduced space, where its negative fractional power is evaluated. A numerical comparison with the predecessor highlights its competitive performance.

Image dehazing is an ill-posed problem that has been extensively studied in the recent years. The objective performance evaluation of the dehazing methods is one of the major obstacles due to the lacking of a reference dataset. While the synthetic datasets have shown important limitations, the few realistic datasets introduced recently assume homogeneous haze over the entire scene. Since in many real cases haze is not uniformly distributed we introduce NH-HAZE, a non-homogeneous realistic dataset with pairs of real hazy and corresponding haze-free images. This is the first non-homogeneous image dehazing dataset and contains 55 outdoor scenes. The non-homogeneous haze has been introduced in the scene using a professional haze generator that imitates the real conditions of hazy scenes. Additionally, this work presents an objective assessment of several state-of-the-art single image dehazing methods that were evaluated using NH-HAZE dataset.

This paper is the first to consider online algorithms to schedule a proportionate flexible flow shop of batching machines (PFFB). The scheduling model is motivated by manufacturing processes of individualized medicaments, which are used in modern medicine to treat some serious illnesses. We provide two different online algorithms, proving also lower bounds for the offline problem to compute their competitive ratios. The first algorithm is an easy-to-implement, general local scheduling heuristic. It is 2-competitive for PFFBs with an arbitrary number of stages and for several natural scheduling objectives. We also show that for total/average flow time, no deterministic algorithm with better competitive ratio exists. For the special case with two stages and the makespan or total completion time objective, we describe an improved algorithm that achieves the best possible competitive ratio $varphi=frac{1+sqrt{5}}{2}$, the golden ratio. All our results also hold for proportionate (non-flexible) flow shops of batching machines (PFB) for which this is also the first paper to study online algorithms.

This paper proposes a faceted information exploration model that supports coarse-grained and fine-grained focusing of geographic maps by offering a graphical representation of data attributes within interactive widgets. The proposed approach enables (i) a multi-category projection of long-lasting geographic maps, based on the proposal of efficient facets for data exploration in sparse and noisy datasets, and (ii) an interactive representation of the search context based on widgets that support data visualization, faceted exploration, category-based information hiding and transparency of results at the same time. The integration of our model with a semantic representation of geographical knowledge supports the exploration of information retrieved from heterogeneous data sources, such as Public Open Data and OpenStreetMap. We evaluated our model with users in the OnToMap collaborative Web GIS. The experimental results show that, when working on geographic maps populated with multiple data categories, it outperforms simple category-based map projection and traditional faceted search tools, such as checkboxes, in both user performance and experience.

Predicate constraints of general-purpose knowledge bases (KBs) like Wikidata, DBpedia and Freebase are often limited to subproperty, domain and range constraints. In this demo we showcase CounQER, a system that illustrates the alignment of counting predicates, like staffSize, and enumerating predicates, like workInstitution^{-1} . In the demonstration session, attendees can inspect these alignments, and will learn about the importance of these alignments for KB question answering and curation. CounQER is available at https://counqer.mpi-inf.mpg.de/spo.

Sentence level quality estimation (QE) for machine translation (MT) attempts to predict the translation edit rate (TER) cost of post-editing work required to correct MT output. We describe our view on sentence-level QE as dictated by several practical setups encountered in the industry. We find consumers of MT output---whether human or algorithmic ones---to be primarily interested in a binary quality metric: is the translated sentence adequate as-is or does it need post-editing? Motivated by this we propose a quality classification (QC) view on sentence-level QE whereby we focus on maximizing recall at precision above a given threshold. We demonstrate that, while classical QE regression models fare poorly on this task, they can be re-purposed by replacing the output regression layer with a binary classification one, achieving 50-60\% recall at 90\% precision. For a high-quality MT system producing 75-80\% correct translations, this promises a significant reduction in post-editing work indeed.

For any finite Galois field extension $mathsf{K}/mathsf{F}$, with Galois group $G = mathrm{Gal}(mathsf{K}/mathsf{F})$, there exists an element $alpha in mathsf{K}$ whose orbit $Gcdotalpha$ forms an $mathsf{F}$-basis of $mathsf{K}$. Such an $alpha$ is called a normal element and $Gcdotalpha$ is a normal basis. We introduce a probabilistic algorithm for testing whether a given $alpha in mathsf{K}$ is normal, when $G$ is either a finite abelian or a metacyclic group. The algorithm is based on the fact that deciding whether $alpha$ is normal can be reduced to deciding whether $sum_{g in G} g(alpha)g in mathsf{K}[G]$ is invertible; it requires a slightly subquadratic number of operations. Once we know that $alpha$ is normal, we show how to perform conversions between the working basis of $mathsf{K}/mathsf{F}$ and the normal basis with the same asymptotic cost.

In this paper, we propose a new nonlinear detector with improved interference suppression in Multi-User Multiple Input, Multiple Output (MU-MIMO) system. The proposed detector is a combination of the following parts: QR decomposition (QRD), low complexity users sorting before QRD, sorting-reduced (SR) K-best method and minimum mean square error (MMSE) pre-processing. Our method outperforms a linear interference rejection combining (IRC, i.e. MMSE naturally) method significantly in both strong interference and additive white noise scenarios with both ideal and real channel estimations. This result has wide application importance for scenarios with strong interference, i.e. when co-located users utilize the internet in stadium, highway, shopping center, etc. Simulation results are presented for the non-line of sight 3D-UMa model of 5G QuaDRiGa 2.0 channel for 16 highly correlated single-antenna users with QAM16 modulation in 64 antennas of Massive MIMO system. The performance was compared with MMSE and other detection approaches.

Convolution is an equivariant operation, and image position does not affect its result. A recent study shows that the zero-padding employed in convolutional layers of CNNs provides position information to the CNNs. The study further claims that the position information enables accurate inference for several tasks, such as object recognition, segmentation, etc. However, there is a technical issue with the design of the experiments of the study, and thus the correctness of the claim is yet to be verified. Moreover, the absolute image position may not be essential for the segmentation of natural images, in which target objects will appear at any image position. In this study, we investigate how positional information is and can be utilized for segmentation tasks. Toward this end, we consider {em positional encoding} (PE) that adds channels embedding image position to the input images and compare PE with several padding methods. Considering the above nature of natural images, we choose medical image segmentation tasks, in which the absolute position appears to be relatively important, as the same organs (of different patients) are captured in similar sizes and positions. We draw a mixed conclusion from the experimental results; the positional encoding certainly works in some cases, but the absolute image position may not be so important for segmentation tasks as we think.

Real-world application scenarios like modern Internet services consist of diversity of AI and non-AI modules with very long and complex execution paths. Using component or micro AI benchmarks alone can lead to error-prone conclusions. This paper proposes a scenario-distilling AI benchmarking methodology. Instead of using real-world applications, we propose the permutations of essential AI and non-AI tasks as a scenario-distilling benchmark. We consider scenario-distilling benchmarks, component and micro benchmarks as three indispensable parts of a benchmark suite. Together with seventeen industry partners, we identify nine important real-world application scenarios. We design and implement a highly extensible, configurable, and flexible benchmark framework. On the basis of the framework, we propose the guideline for building scenario-distilling benchmarks, and present two Internet service AI ones. The preliminary evaluation shows the advantage of scenario-distilling AI benchmarking against using component or micro AI benchmarks alone. The specifications, source code, testbed, and results are publicly available from the web site url{this http URL}.

In this paper, we analyze and implement the Dirichlet spectral-Galerkin method for approximating simply supported vibrating plate eigenvalues with variable coefficients. This is a Galerkin approximation that uses the approximation space that is the span of finitely many Dirichlet eigenfunctions for the Laplacian. Convergence and error analysis for this method is presented for two and three dimensions. Here we will assume that the domain has either a smooth or Lipschitz boundary with no reentrant corners. An important component of the error analysis is Weyl's law for the Dirichlet eigenvalues. Numerical examples for computing the simply supported vibrating plate eigenvalues for the unit disk and square are presented. In order to test the accuracy of the approximation, we compare the spectral-Galerkin method to the separation of variables for the unit disk. Whereas for the unit square we will numerically test the convergence rate for a variable coefficient problem.

We present the Perceptimatic English Benchmark, an open experimental benchmark for evaluating quantitative models of speech perception in English. The benchmark consists of ABX stimuli along with the responses of 91 American English-speaking listeners. The stimuli test discrimination of a large number of English and French phonemic contrasts. They are extracted directly from corpora of read speech, making them appropriate for evaluating statistical acoustic models (such as those used in automatic speech recognition) trained on typical speech data sets. We show that phone discrimination is correlated with several types of models, and give recommendations for researchers seeking easily calculated norms of acoustic distance on experimental stimuli. We show that DeepSpeech, a standard English speech recognizer, is more specialized on English phoneme discrimination than English listeners, and is poorly correlated with their behaviour, even though it yields a low error on the decision task given to humans.

Neural style transfer is a well-known branch of deep learning research, with many interesting works and two major drawbacks. Most of the works in the field are hard to use by non-expert users and substantial hardware resources are required. In this work, we present a solution to both of these problems. We have applied neural style transfer to real-time video (over 25 frames per second), which is capable of running on mobile devices. We also investigate the works on achieving temporal coherence and present the idea of fine-tuning, already trained models, to achieve stable video. What is more, we also analyze the impact of the common deep neural network architecture on the performance of mobile devices with regard to number of layers and filters present. In the experiment section we present the results of our work with respect to the iOS devices and discuss the problems present in current Android devices as well as future possibilities. At the end we present the qualitative results of stylization and quantitative results of performance tested on the iPhone 11 Pro and iPhone 6s. The presented work is incorporated in Kunster - AR Art Video Maker application available in the Apple's App Store.

This paper reviews the second challenge on spectral reconstruction from RGB images, i.e., the recovery of whole-scene hyperspectral (HS) information from a 3-channel RGB image. As in the previous challenge, two tracks were provided: (i) a "Clean" track where HS images are estimated from noise-free RGBs, the RGB images are themselves calculated numerically using the ground-truth HS images and supplied spectral sensitivity functions (ii) a "Real World" track, simulating capture by an uncalibrated and unknown camera, where the HS images are recovered from noisy JPEG-compressed RGB images. A new, larger-than-ever, natural hyperspectral image data set is presented, containing a total of 510 HS images. The Clean and Real World tracks had 103 and 78 registered participants respectively, with 14 teams competing in the final testing phase. A description of the proposed methods, alongside their challenge scores and an extensive evaluation of top performing methods is also provided. They gauge the state-of-the-art in spectral reconstruction from an RGB image.

Rescue vessels are the main actors in maritime safety and rescue operations. At the same time, aerial drones bring a significant advantage into this scenario. This paper presents the research directions of the AutoSOS project, where we work in the development of an autonomous multi-robot search and rescue assistance platform capable of sensor fusion and object detection in embedded devices using novel lightweight AI models. The platform is meant to perform reconnaissance missions for initial assessment of the environment using novel adaptive deep learning algorithms that efficiently use the available sensors and computational resources on drones and rescue vessel. When drones find potential objects, they will send their sensor data to the vessel to verity the findings with increased accuracy. The actual rescue and treatment operation are left as the responsibility of the rescue personnel. The drones will autonomously reconfigure their spatial distribution to enable multi-hop communication, when a direct connection between a drone transmitting information and the vessel is unavailable.

With the rapidly worldwide spread of Coronavirus disease (COVID-19), it is of great importance to conduct early diagnosis of COVID-19 and predict the time that patients might convert to the severe stage, for designing effective treatment plan and reducing the clinicians' workloads. In this study, we propose a joint classification and regression method to determine whether the patient would develop severe symptoms in the later time, and if yes, predict the possible conversion time that the patient would spend to convert to the severe stage. To do this, the proposed method takes into account 1) the weight for each sample to reduce the outliers' influence and explore the problem of imbalance classification, and 2) the weight for each feature via a sparsity regularization term to remove the redundant features of high-dimensional data and learn the shared information across the classification task and the regression task. To our knowledge, this study is the first work to predict the disease progression and the conversion time, which could help clinicians to deal with the potential severe cases in time or even save the patients' lives. Experimental analysis was conducted on a real data set from two hospitals with 422 chest computed tomography (CT) scans, where 52 cases were converted to severe on average 5.64 days and 34 cases were severe at admission. Results show that our method achieves the best classification (e.g., 85.91% of accuracy) and regression (e.g., 0.462 of the correlation coefficient) performance, compared to all comparison methods. Moreover, our proposed method yields 76.97% of accuracy for predicting the severe cases, 0.524 of the correlation coefficient, and 0.55 days difference for the converted time.