检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Camps, Daan Mach, Thomas Vandebril, Raf Watkins, David S. National Energy Research Scientific Computing Center Lawrence Berkeley National Laboratory United States University of Potsdam Institute of Mathematics Karl-Liebknecht-Str. 24–25 Potsdam14476 Germany Department of Computer Science KU Leuven Belgium Department of Mathematics Washington State University United States

Pole-swapping algorithms, generalizations of bulge-chasing algorithms, have been shown to be a viable alternative to the bulge-chasing QZ algorithm for solving the generalized eigenvalue problem for a matrix pencil A − λB. It is natural to try to devise a pole-swapping algorithm that solves the standard eigenvalue problem for a single matrix A. This paper introduces such an algorithm and shows that it is competitive with Francis’s bulge-chasing QR *** Codes 65F15, 15A18 © 2024, CC BY.

关键词： Consensus algorithm

来源：评论

学校读者我要写书评

暂无评论

Adam-family Methods for Nonsmooth Optimization with Convergence Guarantees

arXiv

引用

arXiv 2023年

作者： Xiao, Nachuan Hu, Xiaoyin Liu, Xin Toh, Kim-Chuan Institute of Operations Research and Analytics National University of Singapore 3 Research Link Singapore117602 Singapore School of Computer and Computing Science Hangzhou City University Hangzhou310015 China State Key Laboratory of Scientific and Engineering Computing Academy of Mathematics and Systems Science Chinese Academy of Sciences Beijing100190 China Department of Mathematics Institute of Operations Research and Analytics National University of Singapore 10 Lower Kent Ridge Road Singapore119076 Singapore

In this paper, we present a comprehensive study on the convergence properties of Adam-family methods for nonsmooth optimization, especially in the training of nonsmooth neural networks. We introduce a novel two-timescale framework that adopts a two-timescale updating scheme, and prove its convergence properties under mild assumptions. Our proposed framework encompasses various popular Adam-family methods, providing convergence guarantees for these methods in training nonsmooth neural networks. Furthermore, we develop stochastic subgradient methods that incorporate gradient clipping techniques for training nonsmooth neural networks with heavy-tailed noise. Through our framework, we show that our proposed methods converge even when the evaluation noises are only assumed to be integrable. Extensive numerical experiments demonstrate the high efficiency and robustness of our proposed methods. Copyright © 2023, The Authors. All rights reserved.

关键词： Stochastic systems

来源：评论

学校读者我要写书评

暂无评论

Make 1 regularization effective in training sparse CNN

arXiv

引用

arXiv 2018年

作者： He, Juncai Jia, Xiaodong Xu, Jinchao Zhang, Lian Zhao, Liang Department of Mathematics Pennsylvania State University University ParkPA16802 United States Department of Computer Science and Engineering Pennsylvania State University University ParkPA16802 United States State Key Laboratory of Scientific and Engineering Computing Academy of Mathematics and Systems Science Chinese Academy of Sciences University of Chinese Academy of Sciences Beijing100190 China

Compressed Sensing using 1 regularization is among the most powerful and popular sparsification technique in many applications, but why has it not been used to obtain sparse deep learning model such as convolutional neural network (CNN)? This paper is aimed to provide an answer to this question and to show how to make it work. Following Xiao (2010), We first demonstrate that the commonly used stochastic gradient decent (SGD) and variants training algorithm is not an appropriate match with 1 regularization and then replace it with a different training algorithm based on a regularized dual averaging (RDA) method. The RDA method of Xiao (2010) was originally designed specifically for convex problem, but with new theoretical insight and algorithmic modifications (using proper initialization and adaptivity), we have made it an effective match with 1 regularization to achieve a state-of-the-art sparsity for the highly non-convex CNN compared to other weight pruning methods without compromising accuracy (achieving 95% sparsity for ResNet-18 on CIFAR-10, for example). Copyright © 2018, The Authors. All rights reserved.

关键词： Stochastic systems

来源：评论

学校读者我要写书评

暂无评论

DIVIDE AND CONQUER: LEARNING CHAOTIC DYNAMICAL SYSTEMS WITH MULTI-STEP PENALTY NEURAL ORDINARY DIFFERENTIAL EQUATIONS

arXiv

引用

arXiv 2024年

作者： Chakraborty, Dibyajyoti Chung, Seung Whan Arcomano, Troy Maulik, Romit Information Sciences and Technology Pennsylvania State University University ParkPA United States Center for Applied Scientific Computing Lawrence Livermore National Laboratory LivermoreCA United States Environmental Science Division Argonne National Laboratory LemontIL United States Mathematics and Computer Science Division Argonne National Laboratory LemontIL United States

Forecasting high-dimensional dynamical systems is a fundamental challenge in various fields, such as geosciences and engineering. Neural Ordinary Differential Equations (NODEs), which combine the power of neural networks and numerical solvers, have emerged as a promising algorithm for forecasting complex nonlinear dynamical systems. However, classical techniques used for NODE training are ineffective for learning chaotic dynamical systems. In this work, we propose a novel NODE-training approach that allows for robust learning of chaotic dynamical systems. Our method addresses the challenges of non-convexity and exploding gradients associated with underlying chaotic dynamics. Training data trajectories from such systems are split into multiple, non-overlapping time windows. In addition to the deviation from the training data, the optimization loss term further penalizes the discontinuities of the predicted trajectory between the time windows. The window size is selected based on the fastest Lyapunov time scale of the system. Multi-step penalty(MP) method is first demonstrated on Lorenz equation, to illustrate how it improves the loss landscape and thereby accelerates the optimization convergence. MP method can optimize chaotic systems in a manner similar to least-squares shadowing with significantly lower computational costs. Our proposed algorithm, denoted the Multistep Penalty NODE, is applied to chaotic systems such as the Kuramoto–Sivashinsky equation, the two-dimensional Kolmogorov flow, and ERA5 reanalysis data for the atmosphere. It is observed that MP-NODE provide viable performance for such chaotic systems, not only for short-term trajectory predictions but also for invariant statistics that are hallmarks of the chaotic nature of these dynamics. © 2024, CC BY-NC-SA.

关键词： Trajectories

来源：评论

学校读者我要写书评

暂无评论

Evaluating the Impact of Energy Efficient Networks on HPC Workloads

Evaluating the Impact of Energy Efficient Networks on HPC Wo...

引用

International Conference on High Performance computing

作者： Giorgis Georgakoudis Nikhil Jain Takatsugu Ono Koji Inoue Shinobu Miwa Abhinav Bhatele Center for Applied Scientific Computing Lawrence Livermore National Labor0atory NVIDIA Inc. Department of Advanced Information Technology Kyushu University Department of Informatics Kyushu University Department of Information System Fundamentals The University of Electro-Communications Department of Computer Science University of Maryland College Park

ISBN: (纸本)9781728145365

Interconnection networks grow larger as supercomputers include more nodes and require higher bandwidth for performance. This scaling significantly increases the fraction of power consumed by the network, by increasing the number of network components (links and switches). Typically, network links consume power continuously once they are turned on. However, recent proposals for energy efficient interconnects have introduced low-power operation modes for periods when network links are idle. Low-power operation can increase messaging time when switching a link from low-power to active operation. We extend the TraceR-CODES network simulator for power modeling to evaluate the impact of energy efficient networking on power and performance. Our evaluation presents the first study on both single-job and multi-job execution to realistically simulate power consumption and performance under congestion for a large-scale HPC network. Results on several workloads consisting of HPC proxy applications show that single-job and multi-job execution favor different modes of low power operation to have significant power savings at the cost of minimal performance degradation.

关键词： Delays Power demand Transmitters Receivers Ethernet Standards Proposals

来源：评论

学校读者我要写书评

暂无评论

computing and data grids for science and engineering 00

Computing and data grids for science and engineering

引用

Proceedings of the 2000 ACM/IEEE conference on Supercomputing

作者： William E. Johnston Dennis Gannon Bill Nitzberg Leigh Ann Tanner Bill Thigpen Alex Woo National Energy Research Scientific Computing Division Lawrence Berkeley National Laboratory and Numerical Aerospace Simulation Division NASA Ames Research Center Computer Science Department University of Indiana and NAS NASA Ames Veridian Systems PBS Products Numerical Aerospace Simulation Division NASA Ames Research Center

ISBN: (纸本)9780780398023

We use the term "Grid" to refer to a software system that provides uniform and location independent access to geographically and organizationally dispersed, heterogeneous resources which are persistent and supported. While, in general, Grids will provide the infrastructure to support a wide range of services in the scientific environment (e.g. collaboration and remote instrument control) in this paper we focus on services for high performance computing and data handling. We describe the services and architecture of NASA's Information Power Grid ("IPG")-an early example of a large-scale Grid-and some of the issues that have come up in its implementation.

关键词： Grids

来源：评论

学校读者我要写书评

暂无评论

Heuristics for estimating contact area of supports in layered manufacturing

引用

ACM Journal of Experimental Algorithmics 2006年 11卷 1.6–es页

作者： Ilinkin, Ivayio Janardan, Ravi Smid, Michiel Johnson, Eric Castillo, Paul Schwerdt, Jörg Department of Mathematics and Computer Science Rhodes College Memphis 38112 TN United States Department of Computer Science and Engineering University of Minnesota Minneapolis 55455 MN United States School of Computer Science Carleton University Ottawa K1S 5B6 ON Canada Center for Applied Scientific Computing Lawrence Livermore National Laboratory Livermore 94551 CA United States Softwarebüro Bubel GmbH Kirkel 66459 Germany

Layered manufacturing is a technology that allows physical prototypes of three-dimensional(3D) models to be built directly from their digital representation, as a stack of two-dimensional(2D) layers. A key design problem here is the choice of a suitable direction in which the digital model should be oriented and built so as to minimize the area of contact between the prototype and temporary support structures that are generated during the build. Devising an efficient algorithm for computing such a direction has remained a difficult problem for quite some time. In this paper, a suite of efficient and practical heuristics is presented for estimating the minimum contact area. Also given is a technique for evaluating the quality of the estimate provided by any heuristic, which does not require knowledge of the (unknown and hard-to-compute) optimal solution;instead, it provides an indirect upper bound on the quality of the estimate via two relatively easy-to-compute quantities. The algorithms are based on various techniques from computational geometry, such as ray-shooting, convex hulls, boolean operations on polygons, and spherical arrangements, and have been implemented and tested. Experimental results on a wide range of real-world models show that the heuristics perform quite well in practice. © 2006 ACM.

关键词： Algorithm implementation and testing Algorithms Computational geometry E.1 [Data structures] Experimentation F.2.2 [Analysis of algorithms and problem complexity]: Nonnumerical algorithms and problems I.6 [computer-aided engineering] Measurement Performance Theory

来源：评论

学校读者我要写书评

暂无评论

Euro-Par 2010 Parallel Processing Workshops 1

引用

丛书名： Lecture Notes in computer science

1000年

作者： Mario R. Guarracino Frédéric Vivien Jesper Larsson Träff Mario Cannatoro Marco Danelutto Anders Hast Francesca Perla Andreas Knüpfer Beniamino Martino Michael Alexander

ISBN: (数字)9783642218781

ISBN: (纸本)9783642218774

This book constitutes thoroughly refereed post-conference proceedings of the workshops of the 16th International Conference on Parallel computing, Euro-Par 2010, held in Ischia, Italy, in August/September 2010. The papers of these 9 workshops HeteroPar, HPCC, HiBB, CoreGrid, UCHPC, HPCF, PROPER, CCPI, and VHPC focus on promotion and advancement of all aspects of parallel and distributed computing.

关键词： System Performance and Evaluation Software Engineering computer Communication Networks Algorithm Analysis and Problem Complexity Programming Techniques Special Purpose and Application-Based Systems

来源：评论

学校读者我要写书评

暂无评论

IoT Security Frameworks: A Comparative Review with a Focus on Privacy 2

IoT Security Frameworks: A Comparative Review with a Focus o...

引用

2nd International Conference on Cyber Resilience, ICCR 2024

作者： Al-Hasnawi, Abduljaleel Niu, Yitong Tawfeq, Jamal F. Ranjan Pradhan, Manas Salahat, Mohammed Ghazal, Taher M. Ministry Of Higher Education And Scientific Research Baghdad Iraq College Of Information Technology Imam Ja'afar Al-Sadiq University Department Of Communication Technology Engineering Baghdad Iraq School Of Aeronautical Engineering AnYang University Anyang China Medical Technical College Al-Farahidi University Department Of Medical Instrumentation Technical Engineering Baghdad10070 Iraq School Of Computing Skyline University College University City Sharjah Sharjah United Arab Emirates College Of Engineering And Technology University Of Science And Technology Of Fujairah Fujairah United Arab Emirates Khalifa University Centre For Cyber Physical Systems Computer Science Department United Arab Emirates Center For Cyber Security Faculty Of Information Science And Technology Selangor Bangi43600 Malaysia

ISBN: (纸本)9798350394962

The privacy of sensitive data within the Internet of Things (IoT) is at risk due to potential privacy violations. These violations can occur through illegal access by external entities, as well as inadvertent disclosures or intentional attacks by those within the system who exceed their authorized access privileges. Furthermore, there needs to be more control experienced by data owners regarding their sensitive data as it transitions from their domain to other entities within the Internet of Things (IoT) ecosystem. Numerous methods have been presented in the academic literature over the past few years to effectively tackle the privacy concerns associated with the Internet of Things (IoT). This paper aims to classify existing literature by organizing it according to the specific technology or technique employed to safeguard privacy. The classification encompasses four primary areas of privacy solutions, specifically, solutions based on encryption, solutions based on identity management, solutions based on policy enforcement, and solutions focused on self-protection. The paper examines various state-of-the-art solutions within each category. It begins by providing a broad comparison between these solutions and our work. Subsequently, it focuses on a specific comparison of the privacy objectives of the offered PEFM solutions. © 2024 IEEE.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

A PCA-based fuzzy tensor evaluation model for multiple-criteria group decision making

引用

Applied Soft computing 2023年 132卷

作者： Singh, Meenu Pant, Millie Kong, Lingping Alijani, Zahra Snášel, Václav Department of Applied Mathematics and Scientific Computing Indian Institute of Technology Roorkee Uttarakhand247667 India Mehta Family School of Data Science and Artificial Intelligence Indian Institute of Technology Roorkee Uttarakhand247667 India Faculty of Electrical Engineering and Computer Science VŠB-TU Ostrava Czech Republic Institute for Research and Applications of Fuzzy Modeling University of Ostrava 30. dubna 22 701 03 Ostrava Czech Republic

Multiple-criteria group decision-making (MCGDM) problems mainly consist of multiple factors and multiple Decision Makers (DMs) or Users, for which dimension extension is necessary when considering all the entries of DMs together. Tensor, a generalized form of a matrix, displays a multi-way array item, which is the most suitable and practical way to represent high-dimensional data without losing any information. In this paper, we first reduce the dimension through Principal Component Analysis (PCA), which helps consider the most-informative criteria. Then, we reintroduced the tensor as a fuzzy-form tensor for the MCGDM problem because of user information uncertainty. We choose Interval-valued Neutrosophic Fuzzy numbers (IVNFNs) as the basis for the tensor form because of their ability to distinguish between truth, indeterminacy, and falsity in the data. Lastly, a Generalized Interval-valued Neutrosophic Fuzzy Weighted Geometric (GIVNFWG) operator is defined. Moreover, a generalized framework for fuzzy-form tensors for high-dimensional MCGDM problems is proposed. The feasibility and efficiency of this proposed process is illustrated for a real-world MCGDM problem of ranking the most efficient Third Party Reverse Logistics Partners (3PRLPs), i.e., recycled fiber-based paper mills for the packaging industry. The obtained results are according to the experts and are validated using sensitivity analysis. This analysis facilitates in assessing the impact on the overall ranking performance of 3PRLPs by considering various combinations of environment and technological sub-criteria. © 2022 Elsevier B.V.

关键词： Principal component analysis

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：