arXiv subjects

Yu Wang

Publications and source records attributed to Yu Wang.

At least 1,081 records · Page 60Linked to original sources

arXiv2022

Heuristic Adaptability to Input Dynamics for SpMM on GPUs

Sparse Matrix-Matrix Multiplication (SpMM) has served as fundamental components in various domains. Many previous studies exploit GPUs for SpMM acceleration because GPUs provide high bandwidth and parallelism. We point out that a static design does not always improve the performance of SpMM on different input data (e.g., >85\% performance loss with a single algorithm). In this paper, we consider the challenge of input dynamics from a novel auto-tuning perspective, while following issues remain to be solved: (1) Orthogonal design principles considering sparsity. Orthogonal design principles for such a sparse problem should be extracted to form different algorithms, and further used for performance tuning. (2) Nontrivial implementations in the algorithm space. Combining orthogonal design principles to create new algorithms needs to tackle with new challenges like thread race handling. (3) Heuristic adaptability to input dynamics. The heuristic adaptability is required to dynamically optimize code for input dynamics. To tackle these challenges, we first propose a novel three-loop model to extract orthogonal design principles for SpMM on GPUs. The model not only covers previous SpMM designs, but also comes up with new designs absent from previous studies. We propose techniques like conditional reduction to implement algorithms missing in previous studies. We further propose DA-SpMM, a Data-Aware heuristic GPU kernel for SpMM. DA-SpMM adaptively optimizes code considering input dynamics. Extensive experimental results show that, DA-SpMM achieves 1.26x~1.37x speedup compared with the best NVIDIA cuSPARSE algorithm on average, and brings up to 5.59x end-to-end speedup to applications like Graph Neural Networks.

cs.AR↗

arXiv2022

Large-scale Personalized Video Game Recommendation via Social-aware Contextualized Graph Neural Network

Because of the large number of online games available nowadays, online game recommender systems are necessary for users and online game platforms. The former can discover more potential online games of their interests, and the latter can attract users to dwell longer in the platform. This paper investigates the characteristics of user behaviors with respect to the online games on the Steam platform. Based on the observations, we argue that a satisfying recommender system for online games is able to characterize: personalization, game contextualization and social connection. However, simultaneously solving all is rather challenging for game recommendation. Firstly, personalization for game recommendation requires the incorporation of the dwelling time of engaged games, which are ignored in existing methods. Secondly, game contextualization should reflect the complex and high-order properties of those relations. Last but not least, it is problematic to use social connections directly for game recommendations due to the massive noise within social connections. To this end, we propose a Social-aware Contextualized Graph Neural Recommender System (SCGRec), which harnesses three perspectives to improve game recommendation. We conduct a comprehensive analysis of users' online game behaviors, which motivates the necessity of handling those three characteristics in the online game recommendation.

cs.IR↗

arXiv2022

Opencpop: A High-Quality Open Source Chinese Popular Song Corpus for Singing Voice Synthesis

This paper introduces Opencpop, a publicly available high-quality Mandarin singing corpus designed for singing voice synthesis (SVS). The corpus consists of 100 popular Mandarin songs performed by a female professional singer. Audio files are recorded with studio quality at a sampling rate of 44,100 Hz and the corresponding lyrics and musical scores are provided. All singing recordings have been phonetically annotated with phoneme boundaries and syllable (note) boundaries. To demonstrate the reliability of the released data and to provide a baseline for future research, we built baseline deep neural network-based SVS models and evaluated them with both objective metrics and subjective mean opinion score (MOS) measure. Experimental results show that the best SVS model trained on our database achieves 3.70 MOS, indicating the reliability of the provided corpus. Opencpop is released to the open-source community WeNet, and the corpus, as well as synthesized demos, can be found on the project homepage.

cs.SD↗

arXiv2022

Null controllability for stochastic fourth order parabolic equations

We establish the null controllability for linear stochastic fourth order parabolic equations. Utilizing the duality argument, the null controllability is reduced to the observability for backward fourth order stochastic parabolic equations, and the desired observability estimate is obtained by a new global Carleman estimate. Our Carleman estimate is based on a new fundamental identity for a stochastic fourth order parabolic operator.

math.OC↗

arXiv2022

A Brief Review of Binary Driven Hypernova

Binary driven hypernova (BdHN) models long gamma-ray burst (GRBs) as occurring in the binary systems involving a carbon-oxygen core (CO$_{\rm core}$) and a companion neutron star (NS) or a black hole (BH). This model, first proposed in 2012, succeeds and improves upon the fireshell model and the induced gravitational collapse (IGC) paradigm. After nearly a decade of development, the BdHN model has reached a nearly complete structure, explaining all the observables of long bursts into its theoretical framework, and has given a refined classification of long GRBs according to the original properties of the progenitors. In this article, we present a summary of the BdHN model and the physical processes at work in each of the envisaged Episodes during its occurrence and lifetime, duly contextualized in the framework of GRB observations.

astro-ph.HE↗

arXiv2021

A High Spatial Resolution Muon Tomography Prototype System based on Micromegas Detector

Cosmic ray muon has strong penetrating power and no ionizing radiation hazards, which makes it an ideal probe for detecting special nuclear materials. In this paper, a high spatial resolution muon tomography system based on Micromegas detectors is proposed to optimize the imaging time and quality. The proposed system includes eight Micromegas detectors based on the thermal bonding technique and a scalable readout system. In addition, a multiplexing method based on position encoding is developed to reduce the number of electronics channels by order of magnitude. The spatial resolution of the proposed system with encoding readout can reach a value of hundred micrometers. Finally, a tomography test is performed, and test results show that this proposed system can image 2-cm objects and distinguish different materials.

physics.ins-det↗

arXiv2021

Stability of scar states in 2D PXP model against random disorders

Recently a class of quantum systems exhibiting weak ergodicity breaking has attracted much attention. These systems feature a special band of eigenstates called quantum many-body scar states in the energy spectrum. In this work we study the fate of quantum many-body scar states in a two-dimensional lattice against random disorders. We show that in both the square lattice and the honeycomb lattice the scar states can persist up to a finite disorder strength, before eventually being erased by the disorder. We further study the localization properties of the system in the presence of even stronger disorders and show that whether a full localization transition occurs depends on the type of disorder we introduce. Our study thus reveals the fascinating interplay between disorder and quantum many-body scarring in a two-dimensional system.

cond-mat.dis-nn↗

arXiv2021

Variational Automatic Curriculum Learning for Sparse-Reward Cooperative Multi-Agent Problems

We introduce a curriculum learning algorithm, Variational Automatic Curriculum Learning (VACL), for solving challenging goal-conditioned cooperative multi-agent reinforcement learning problems. We motivate our paradigm through a variational perspective, where the learning objective can be decomposed into two terms: task learning on the current task distribution, and curriculum update to a new task distribution. Local optimization over the second term suggests that the curriculum should gradually expand the training tasks from easy to hard. Our VACL algorithm implements this variational paradigm with two practical components, task expansion and entity progression, which produces training curricula over both the task configurations as well as the number of entities in the task. Experiment results show that VACL solves a collection of sparse-reward problems with a large number of agents. Particularly, using a single desktop machine, VACL achieves 98% coverage rate with 100 agents in the simple-spread benchmark and reproduces the ramp-use behavior originally shown in OpenAI's hide-and-seek project. Our project website is at https://sites.google.com/view/vacl-neurips-2021.

cs.LG↗

arXiv2021

Enhancing fiber atom interferometer by in-fiber laser cooling

We demonstrate an inertia sensitive atom interferometer optically guided inside a 22-cm-long negative curvature hollow-core photonic crystal fiber with an interferometer time of 20 ms. The result prolongs the previous fiber guided atom interferometer time by three orders of magnitude. The improvement arises from the realization of in-fiber Λ-enhanced gray molasses and delta-kick cooling to cool atoms from 32 μK to below 1 μK in 4 ms. The in-fiber cooling overcomes the inevitable heating during the atom loading process and allows a shallow guiding optical potential to minimize decoherence. Our results permit bringing atoms close to source fields for sensing and could lead to compact inertial quantum sensors with a sub-millimeter resolution.

physics.atom-ph↗

arXiv2021

Improving Self-supervised Learning with Automated Unsupervised Outlier Arbitration

Our work reveals a structured shortcoming of the existing mainstream self-supervised learning methods. Whereas self-supervised learning frameworks usually take the prevailing perfect instance level invariance hypothesis for granted, we carefully investigate the pitfalls behind. Particularly, we argue that the existing augmentation pipeline for generating multiple positive views naturally introduces out-of-distribution (OOD) samples that undermine the learning of the downstream tasks. Generating diverse positive augmentations on the input does not always pay off in benefiting downstream tasks. To overcome this inherent deficiency, we introduce a lightweight latent variable model UOTA, targeting the view sampling issue for self-supervised learning. UOTA adaptively searches for the most important sampling region to produce views, and provides viable choice for outlier-robust self-supervised learning approaches. Our method directly generalizes to many mainstream self-supervised learning approaches, regardless of the loss's nature contrastive or not. We empirically show UOTA's advantage over the state-of-the-art self-supervised paradigms with evident margin, which well justifies the existence of the OOD sample issue embedded in the existing approaches. Especially, we theoretically prove that the merits of the proposal boil down to guaranteed estimator variance and bias reduction. Code is available: at https://github.com/ssl-codelab/uota.

cs.LG↗

arXiv2021

Transferrable Contrastive Learning for Visual Domain Adaptation

Self-supervised learning (SSL) has recently become the favorite among feature learning methodologies. It is therefore appealing for domain adaptation approaches to consider incorporating SSL. The intuition is to enforce instance-level feature consistency such that the predictor becomes somehow invariant across domains. However, most existing SSL methods in the regime of domain adaptation usually are treated as standalone auxiliary components, leaving the signatures of domain adaptation unattended. Actually, the optimal region where the domain gap vanishes and the instance level constraint that SSL peruses may not coincide at all. From this point, we present a particular paradigm of self-supervised learning tailored for domain adaptation, i.e., Transferrable Contrastive Learning (TCL), which links the SSL and the desired cross-domain transferability congruently. We find contrastive learning intrinsically a suitable candidate for domain adaptation, as its instance invariance assumption can be conveniently promoted to cross-domain class-level invariance favored by domain adaptation tasks. Based on particular memory bank constructions and pseudo label strategies, TCL then penalizes cross-domain intra-class domain discrepancy between source and target through a clean and novel contrastive loss. The free lunch is, thanks to the incorporation of contrastive learning, TCL relies on a moving-averaged key encoder that naturally achieves a temporally ensembled version of pseudo labels for target data, which avoids pseudo label error propagation at no extra cost. TCL therefore efficiently reduces cross-domain gaps. Through extensive experiments on benchmarks (Office-Home, VisDA-2017, Digits-five, PACS and DomainNet) for both single-source and multi-source domain adaptation tasks, TCL has demonstrated state-of-the-art performances.

cs.CV↗

arXiv2021

A Style and Semantic Memory Mechanism for Domain Generalization

Mainstream state-of-the-art domain generalization algorithms tend to prioritize the assumption on semantic invariance across domains. Meanwhile, the inherent intra-domain style invariance is usually underappreciated and put on the shelf. In this paper, we reveal that leveraging intra-domain style invariance is also of pivotal importance in improving the efficiency of domain generalization. We verify that it is critical for the network to be informative on what domain features are invariant and shared among instances, so that the network sharpens its understanding and improves its semantic discriminative ability. Correspondingly, we also propose a novel "jury" mechanism, which is particularly effective in learning useful semantic feature commonalities among domains. Our complete model called STEAM can be interpreted as a novel probabilistic graphical model, for which the implementation requires convenient constructions of two kinds of memory banks: semantic feature bank and style feature bank. Empirical results show that our proposed framework surpasses the state-of-the-art methods by clear margins.

cs.CV↗

arXiv2021

Multi-Agent Vulnerability Discovery for Autonomous Driving with Hazard Arbitration Reward

Discovering hazardous scenarios is crucial in testing and further improving driving policies. However, conducting efficient driving policy testing faces two key challenges. On the one hand, the probability of naturally encountering hazardous scenarios is low when testing a well-trained autonomous driving strategy. Thus, discovering these scenarios by purely real-world road testing is extremely costly. On the other hand, a proper determination of accident responsibility is necessary for this task. Collecting scenarios with wrong-attributed responsibilities will lead to an overly conservative autonomous driving strategy. To be more specific, we aim to discover hazardous scenarios that are autonomous-vehicle responsible (AV-responsible), i.e., the vulnerabilities of the under-test driving policy. To this end, this work proposes a Safety Test framework by finding Av-Responsible Scenarios (STARS) based on multi-agent reinforcement learning. STARS guides other traffic participants to produce Av-Responsible Scenarios and make the under-test driving policy misbehave via introducing Hazard Arbitration Reward (HAR). HAR enables our framework to discover diverse, complex, and AV-responsible hazardous scenarios. Experimental results against four different driving policies in three environments demonstrate that STARS can effectively discover AV-responsible hazardous scenarios. These scenarios indeed correspond to the vulnerabilities of the under-test driving policies, thus are meaningful for their further improvements.

cs.RO↗

arXiv2021

Multiway Ensemble Kalman Filter

In this work, we study the emergence of sparsity and multiway structures in second-order statistical characterizations of dynamical processes governed by partial differential equations (PDEs). We consider several state-of-the-art multiway covariance and inverse covariance (precision) matrix estimators and examine their pros and cons in terms of accuracy and interpretability in the context of physics-driven forecasting when incorporated into the ensemble Kalman filter (EnKF). In particular, we show that multiway data generated from the Poisson and the convection-diffusion types of PDEs can be accurately tracked via EnKF when integrated with appropriate covariance and precision matrix estimators.

stat.ML↗

arXiv2021

Two-orbital model for possible superconductivity pairing mechanism in nickelates

The newly synthesized strontium doped RNiO$_2$ (R=Nd, La) superconductors have stimulated extensive interests in understanding their pairing mechanism and pairing nature. Here we study the pairing mechanism in this family from a two-orbital model comprising the Ni- $3d_{x^2-y^2}$- and $3d_{xy}$- orbitals, equipped with extended Hubbard interactions and induced low-energy effective superexchange interactions. We then study the pairing symmetry in this system by using large scale variational Monte Carlo approach. Our results yield the intraorbital $d_{x^2-y^2}$-wave singlet pairing as the leading pairing symmetry in the nickelates, which is analogous to the cuprates. However, there exist two important differences between the physical properties of the two families due to the fact that at the low Sr-doping regime, while the Ni-$3d_{x^2-y^2}$ orbitals remain half-filled and singly-occupied to form a Mott-insulating background, the Ni-$3d_{xy}$ orbitals accommodate nearly all the extra doped holes, which move freely on this background. The first difference lies in the single-particle aspect: while the $3d_{x^2-y^2}$ degree of freedom remains Mott insulating with spectra weight pinned down at zero at low dopings, the $3d_{xy}$ one behaves as Fermi liquid with spectra weight near 1. The second difference lies in the pairing aspect: while the huge intra-$3d_{x^2-y^2}$-orbital pairing gap is actually a pseudo gap which has nothing to do with the SC, the small intra-$3d_{xy}$-orbital pairing gap serves as the true superconducting pairing gap, which is related to the $T_c$ via the BCS relation. Both differences can be verified by the angle-resolved photo-emission spectrum.

cond-mat.str-el↗

arXiv2021

Distribution-Balanced Loss for Multi-Label Classification in Long-Tailed Datasets

We present a new loss function called Distribution-Balanced Loss for the multi-label recognition problems that exhibit long-tailed class distributions. Compared to conventional single-label classification problem, multi-label recognition problems are often more challenging due to two significant issues, namely the co-occurrence of labels and the dominance of negative labels (when treated as multiple binary classification problems). The Distribution-Balanced Loss tackles these issues through two key modifications to the standard binary cross-entropy loss: 1) a new way to re-balance the weights that takes into account the impact caused by label co-occurrence, and 2) a negative tolerant regularization to mitigate the over-suppression of negative labels. Experiments on both Pascal VOC and COCO show that the models trained with this new loss function achieve significant performance gains over existing methods. Code and models are available at: https://github.com/wutong16/DistributionBalancedLoss .

cs.CV↗

arXiv2021

Generating Synthetic Systems of Interdependent Critical Infrastructure Networks

The lack of data on critical infrastructure systems has hindered the research progress in modeling and optimizing the system performance. This work develops a method for generating Synthetic Interdependent Critical Infrastructure Networks (SICIN) using simulation and non-linear optimization techniques. SICIN consists of three components: (i) determining the location of facilities in individual networks via a modified simulated annealing algorithm, (ii) generating interdependent links based on a novel pseudo-tripartite graph algorithm, and (iii) simulating network flow using nonlinear optimization considering the operations of individual networks and their interdependencies. Two existing systems of interdependent infrastructure networks are used to validate the proposed method. The results demonstrate that SICIN outperforms state-of-the-art simulation methods according to multiple topological and flow measures of similarity between the simulated and real networks.

physics.soc-ph↗

arXiv2021

Learning Dynamic Compact Memory Embedding for Deformable Visual Object Tracking

Recently, template-based trackers have become the leading tracking algorithms with promising performance in terms of efficiency and accuracy. However, the correlation operation between query feature and the given template only exploits accurate target localization, leading to state estimation error especially when the target suffers from severe deformable variations. To address this issue, segmentation-based trackers have been proposed that employ per-pixel matching to improve the tracking performance of deformable objects effectively. However, most of existing trackers only refer to the target features in the initial frame, thereby lacking the discriminative capacity to handle challenging factors, e.g., similar distractors, background clutter, appearance change, etc. To this end, we propose a dynamic compact memory embedding to enhance the discrimination of the segmentation-based deformable visual tracking method. Specifically, we initialize a memory embedding with the target features in the first frame. During the tracking process, the current target features that have high correlation with existing memory are updated to the memory embedding online. To further improve the segmentation accuracy for deformable objects, we employ a point-to-global matching strategy to measure the correlation between the pixel-wise query features and the whole template, so as to capture more detailed deformation information. Extensive evaluations on six challenging tracking benchmarks including VOT2016, VOT2018, VOT2019, GOT-10K, TrackingNet, and LaSOT demonstrate the superiority of our method over recent remarkable trackers. Besides, our method outperforms the excellent segmentation-based trackers, i.e., D3S and SiamMask on DAVIS2017 benchmark.

cs.CV↗