检索结果-内蒙古大学图书馆

Joint 30th International Conference on computational Linguistics and 14th International Conference on Language Resources and Evaluation, LREC-COLING 2024

作者： Yang, Yuting Huang, Pei Ma, Feifei Cao, Juan Li, Jintao China University of Chinese Academy of Sciences China Department of Computer Science Stanford University United States State Key Laboratory of Computer Science Institute of Software CAS China Laboratory of Parallel Software and Computational Science Institute of Software CAS China

ISBN: (纸本)9782493814104

Deep neural networks can be vulnerable to adversarial attacks, even for the mainstream Transformer-based models. Although several robustness enhancement approaches have been proposed, they usually focus on some certain type of perturbation. As the types of attack can be various and unpredictable in practical scenarios, a general and strong defense method is urgently in require. We notice that most well-trained models can be weakly robust in the perturbation space, i.e., only a small ratio of adversarial examples exist. Inspired by the weak robust property, this paper presents a novel ensemble method for enhancing robustness. We propose a lightweight framework PAD to save computational resources in realizing an ensemble. Instead of training multiple models, a plugin module is designed to perturb the parameters of a base model which can achieve the effect of multiple models. Then, to diversify adversarial example distributions among different models, we promote each model to have different attention patterns via optimizing a diversity measure we defined. Experiments on various widely-used datasets and target models show that PAD can consistently improve the defense ability against many types of adversarial attacks while maintaining accuracy on clean data. Besides, PAD also presents good interpretability via visualizing diverse attention patterns. © 2024 ELRA Language Resource Association: CC BY-NC 4.0.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

A Distributed Data Collection System for Traffic Images 3rd

A Distributed Data Collection System for Traffic Images

引用

3rd International Conference on Cyberspace Data and Intelligence, Cyber DI 2019, and the International Conference on Cyber-Living, Cyber-Syndrome, and Cyber-Health, CyberLife 2019

作者： Bo, Wen Yang, Hongju Xiao, Jiaojiao Li, Liangyan Zhang, Changyou School of Computer and Information Technology Shanxi University TaiyuanShanxi China Laboratory of Parallel Software and Computational Science Institute of Software Chinese Academy of Sciences Beijing China

ISBN: (纸本)9789811519215

With the development of intelligent traffic system, high-definition cameras are spread along the urban roads. These devices transmit real-time captured images to data center for multi-purpose usability, but these bring higher requirements on network and storage capacity of the traffic images collection system. To address these problems, we proposed a compressed representation method for traffic images and collection system architecture. Firstly, the method proposed in this paper designed a distributed data collection system for traffic images based on edge computing mode. Secondly, we studied on the image feature representation methods for vehicle type/version retrieval, and formed a compressed representation method based on structural relationships selections. In this method, the retrieval precision reaches to 97.78% with the recall ratio of 90%, which proved the usability in this image collection system. Finally, we set up an analysis model based on Petri-net to observe the system requirements on storage, computing and transmission with different setting parameters. This model is powerful on finding bottlenecks of system in early stage and keeping balance in multi-aspects. The simulation experiments show that the data volume needs to be transported and preserved was compressed to 1/2250 comparing to the method of original images and the system transport delay was reduced more than 1/9 of original method. The experimental result showed that compared with the original collection method, the amount of data to be transmitted and stored was compressed by 1/2250, and the system transmission delay of the system was reduced to 1/9.15. This distributed data collection method and system proposed in this paper provided a novel referable revolution for traffic images processing system in intelligent traffics. © 2019, Springer Nature Singapore Pte Ltd.

关键词： Search engines

来源：评论

学校读者我要写书评

暂无评论

Learning the Satisfiability of Pseudo-Boolean Problem with Graph Neural Networks 26th

Learning the Satisfiability of Pseudo-Boolean Problem with G...

引用

26th International Conference on Principles and Practice of Constraint Programming, CP 2020

作者： Liu, Minghao Zhang, Fan Huang, Pei Niu, Shuzi Ma, Feifei Zhang, Jian State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Beijing China Laboratory of Parallel Software and Computational Science Institute of Software Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China

ISBN: (纸本)9783030584740

Graph Neural Network (GNN) has shown great power on many practical tasks in the past few years. It is also considered to be a potential technique in bridging the gap between machine learning and symbolic reasoning. Experimental investigations have also shown that some-Hard constraint satisfaction problems can be well learned by the GNN models. In this paper, a GNN-based classification model to learn the satisfiability of pseudo-Boolean (PB) problem is proposed. After constructing the bipartite graph representation, a two-phase message passing process is executed. Experiments on 0–1 knapsack and weighted independent set problems show that the model can effectively learn the features related to the problem distribution and satisfiability. As a result, competitive prediction accuracy has been achieved with some generalization to larger-scale problems. The studies indicate that GNN has great potential in solving constraint satisfaction problems with numerical coefficients. © 2020, Springer Nature Switzerland AG.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Can Graph Neural Networks Learn to Solve the MaxSAT Problem? 37

Can Graph Neural Networks Learn to Solve the MaxSAT Problem?

引用

37th AAAI Conference on Artificial Intelligence, AAAI 2023

作者： Liu, Minghao Huang, Pei Jia, Fuqi Zhang, Fan Sun, Yuchen Cai, Shaowei Ma, Feifei Zhang, Jian State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences China Laboratory of Parallel Software and Computational Science Institute of Software Chinese Academy of Sciences China University of Chinese Academy of Sciences China Inspir.ai

ISBN: (纸本)9781577358800

The paper presents an attempt to bridge the gap between machine learning and symbolic reasoning. We build graph neural networks (GNNs) to predict the solution of the Maximum Satisfiability (MaxSAT) problem, an optimization variant of SAT. Two closely related graph representations are adopted, and we prove their theoretical equivalence. We also show that GNNs can achieve attractive performance to solve hard MaxSAT problems in certain distributions even compared with state-of-the-art solvers through experimental evaluation. Copyright © 2023, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Study on parallel Computing

引用

Journal of Computer science & Technology 2006年第5期21卷 665-673页

作者：陈国良孙广中张云泉莫则尧 Anhui Province-MOST Key Co-Lab of High Performance Computing and Its Applications Department of Computer Science and Technology University of Science and Technology of China Hefei 230027 P.R. China Laboratory of Parallel Computing Institute of Software Chinese Academy of Sciences Beijing 100080 P.R. China Laboratory of Computational Physics Institute of Applied Physics and Computational Mathematics Beijing 100088 P.R. China

In this paper, we present a general survey on parallel computing. The main contents include parallel computer system which is the hardware platform of parallel computing, parallel algorithm which is the theoretical base of parallel computing, parallel programming which is the software support of parallel computing. After that, we also introduce some parallel applications and enabling technologies. We argue that parallel computing research should form an integrated methodology of ＂architecture algorithm programming application＂. Only in this way, parallel computing research becomes continuous development and more realistic.

关键词： parallel computing parallel architecture parallel programming parallel algorithm parallel application

来源：评论

学校读者我要写书评

暂无评论

Resource Scheduling Strategies for Cloud Editors for Low-Code Development 5

Resource Scheduling Strategies for Cloud Editors for Low-Cod...

引用

5th International Conference on Electronics and Communication, Network and Computer Technology, ECNCT 2023

作者： Gao, Yuhao Duan, Shufeng Bo, Wen Zhang, Changyou Shijiazhuang Tiedao University School of Information Science and Technology Shijiazhuang China University of Chinese Academy of Sciences Laboratory of Parallel Software and Computational Science Beijing China

ISBN: (纸本)9798350328660

software ecosystems are accelerating the vertical division of labour and horizontal integration in the software industry, attracting widespread attention from industry and academia. The low-code development cloud editor oriented in this paper is a generalisation of building software ecosystems, which reduces the technical threshold of developers in the ecosystem and improves the development efficiency by supplementing the interfaces and services and customising the cloud editor platform to support core technologies such as multi-language, sharing, service publishing, and low-code. In this context, this paper focuses on the editor resource scheduling strategy for cloud editor: 1) for low-code programming, a collaborative service planning method is proposed to start containers running synchronously on demand in units of service components, avoiding the start of temporary unnecessary service startups and improving the resource utilisation rate;2) designing a threetier platform service framework that supports on-demand running of component code, and formulating a related service strategy that Improve the efficiency of component execution;3) For three different scenarios, a resource scheduling method based on greedy algorithm is proposed to reasonably utilise resources;4) Experiments verify the effect of the algorithmic strategy in this paper in terms of resource scheduling efficiency, quality and stability. The work in this paper is used in project development practice to support low-code development software ecosystem. © 2023 IEEE.

关键词： Efficiency

来源：评论

学校读者我要写书评

暂无评论

Investigation on Discontinuous Galerkin Solvers for Different Simulations

引用

Procedia Engineering 2013年 61卷 158-165页

作者： Xiaohe Zhufu Ruifeng Zhao Jin Xu Laboratory of Parallel Software and Computational Science Institute of Software Chinese Academy of Sciences Beijing 100190 CHINA

Discontinuous Galerkin(DG) method has been successfully used in time dependent simulations. This paper will investigate DG solvers for di_erent simulations. They are compressible flow and electromagnetic wave simulations. The compressible flow simulation needs to solve the Euler equation, while electromagnetic wave simulation needs to solve the Maxwell's equations. Two solvers have been studied in deep separately. All solvers apply domain decomposition method for the parallelization, and they both show the advantages of DG method: easy parallelization and good scalability. Verifications will be shown on simple geometries in detail for both cases, and at last a wakefield simulation ha been shown with electromagnetic time domain solver.

关键词： Compressible flow Euler equation Maxwell equation Time domain Discontinuous Galerkin method Nodal basis

来源：评论

学校读者我要写书评

暂无评论

Solving multi-objective constrained minimum weighted bipartite assignment problem: a case study on energy-aware radio broadcast scheduling

引用

science China(Information sciences) 2022年第8期65卷 23-38页

作者： Yupeng ZHOU Mingjie FAN Feifei MA Minghao YIN College of Information Science and Technology Northeast Normal University Laboratory of Parallel Software and Computational Science Institute of Software Chinese Academy of Sciences State Key laboratory of Computer Science Institute of Software Chinese Academy of Sciences

This paper proposes a multi-objective constrained minimum weighted bipartite assignment problem(MCMWBAP), which is considered an extension of the classical bipartite matching problem(BMP). We first provide the formulation of the MCMWBAP and prove that it is an NP-hard combinatorial optimization problem. Based on this formulation, multi-objective energy-aware shortwave radio broadcast resource allocation problem(MSRBRAP) application is studied. The goal of this problem is to allocate radio programs to transmission devices to broadcast all radio programs felicitously with a maximized objective of total qualified monitoring sites and a minimized objective of energy consumption. Then, a novel multi-objective hybrid evolutionary algorithm(MOHEA), which is integrated with push and pull initialization, the dynamic resource allocation strategy, and the aggregate local search procedure, is developed to solve the *** proposed method is evaluated using two categories of benchmarks for MCMWBAP together with a real scenario case study for MSRBRAP. Furthermore, the key components of MOHEA are analyzed, and the experimental results demonstrate that MOHEA outperforms two classical multi-objective evolutionary algorithms(NSGA-II and MOEA/D), improving working efficiency.

关键词： constrained minimum weighted bipartite assignment multi-objective energy-aware radio broadcast scheduling evolutionary algorithms dynamic resource allocation aggregate local search

来源：评论

学校读者我要写书评

暂无评论

EFFICIENT SPECTRAL METHODS FOR EIGENVALUE PROBLEMS OF THE INTEGRAL FRACTIONAL LAPLACIAN ONABALLOFANYDIMENSION

引用

Journal of computational Mathematics 2024年第4期42卷 1032-1062页

作者： Suna Ma Huiyuan Li Zhimin Zhang Hu Chen Lizhen Chen College of Science Nanjing University of Posts and TelecommunicationsNanjingChina State Key Laboratory of Computer Science/Laboratory of Parallel Computing Institute of SoftwareChinese Academy of SciencesBeijingChina Department of Mathematics Wayne State UniversityDetroitUSA School of Mathematical Sciences Ocean University of ChinaQingdaoChina Beijing Computational Science Research Center BeijingChina

An efficient spectral-Galerkin method for eigenvalue problems of the integral fractional Laplacian on a unit ball of any dimension is proposed in this *** symmetric positive definite linear system is retained explicitly which plays an important role in the numerical *** a sharp estimate on the algebraic system's condition number is established which behaves as N4s with respect to the polynomial degree N,where 2s is the fractional derivative *** regularity estimate of solutions to source problems of the fractional Laplacian in arbitrary dimensions is firstly investigated in weighted Sobolev *** the regularity of eigenfunctions of the fractional Laplacian eigenvalue problem is readily ***,rigorous error estimates of the eigenvalues and eigenvectors are *** experiments are presented to demonstrate the accuracy and efficiency and to validate the theoretical results.

关键词： Integral fractional Laplacian Spectral method Eigenvalue problem Regularity analysis Error estimate

来源：评论

学校读者我要写书评

暂无评论

C1-Conforming Quadrilateral Spectral Element Method for Fourth-Order Equations

引用

Communications on Applied Mathematics and Computation 2019年第3期1卷 403-434页

作者： Huiyuan Li Weikun Shan Zhimin Zhang State Key Laboratory of Computer Science/Laboratory of Parallel Computing Institute of SoftwareChinese Academy of SciencesBeijing 100190China School of Science Henan University of TechnologyZhengzhou 450001HenanChina Beijing Computational Science Research Center Beijing 100193China Department of Mathematics Wayne State UniversityDetroitMI 48202USA

This paper is devoted to Professor Benyu Guo's open question on the C1-conforming quadrilateral spectral element method for fourth-order equations which has been endeavored for years. Starting with generalized Jacobi polynomials on the reference square, we construct the C1-conforming basis functions using the bilinear mapping from the reference square onto each quadrilateral element which fall into three categories-interior modes, edge modes, and vertex modes. In contrast to the triangular element, compulsively compensatory requirements on the global C1-continuity should be imposed for edge and vertex mode basis functions such that their normal derivatives on each common edge are reduced from rational functions to polynomials, which depend on only parameters of the common edge. It is amazing that the C1-conforming basis functions on each quadrilateral element contain polynomials in primitive variables, the completeness is then guaranteed and further confirmed by the numerical results on the Petrov-Galerkin spectral method for the non-homogeneous boundary value problem of fourth-order equations on an arbitrary quadrilateral. Finally, a C1-conforming quadrilateral spectral element method is proposed for the biharmonic eigenvalue problem, and numerical experiments demonstrate the effectiveness and efficiency of our spectral element method.

关键词： Quadrilateral spectral element method Fourth-order equations Mapped polynomials C1-conforming basis Polynomial inclusion Completeness

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：