检索结果-内蒙古大学图书馆

IEEE JOURNAL ON SELECTED AREAS IN INFORMATION THEORY 2021年第3期2卷 830-844页

作者： Krishnan, M. Nikhil Hosseini, Erfan Khisti, Ashish Univ Toronto Dept Elect & Comp Engn Toronto ON M5S 3G4 Canada Int Inst Informat Technol Bangalore Bengaluru 560100 India

In this work, we consider a sequence of J matrix multiplication jobs which needs to be distributed by a master across multiple worker nodes. For i is an element of{1, 2,..., J}, job-i begins in round-i and has to be completed by round-(i + T). In order to provide resiliency against slow workers (stragglers), previous works focus on coding across workers, which is the special case of T = 0. We propose here two schemes with T > 0, which allow for coding across workers as well as the dimension of time. Our first scheme is a modification of the polynomial coding scheme introduced by Yu et al. and places no assumptions on the straggler model. Exploitation of the temporal dimension helps the scheme handle a larger set of straggler patterns than the polynomial coding scheme, for a given computational load per worker per round. The second scheme assumes a particular straggler model to further improve performance (in terms of encoding/decoding complexity). We develop theoretical results establishing (i) optimality of our proposed schemes for certain classes of straggler patterns and (ii) improved performance for the case of i.i.d. stragglers. These are further validated by experiments, where we implement our schemes to train neural networks.

关键词： distributed computation matrix multiplication polynomial codes erasure coding

来源：评论

学校读者我要写书评

暂无评论

Bivariate Polynomial Coding for Efficient distributed Matrix Multiplication

IEEE JOURNAL ON SELECTED AREAS IN INFORMATION THEORY

引用

IEEE JOURNAL ON SELECTED AREAS IN INFORMATION THEORY 2021年第3期2卷 814-829页

作者： Hasircioglu, Burak Gomez-Vilardebo, Jesus Gunduz, Deniz Imperial Coll London Dept Elect & Elect Engn London SW7 2AZ England Ctr Tecnol Telecomun Catalunya CTTC CERCA Barcelona 08860 Spain

Coded computing is an effective technique to mitigate "stragglers" in large-scale and distributed matrix multiplication. In particular, univariate polynomial codes have been shown to be effective in straggler mitigation by making the computation time depend only on the fastest workers. However, these schemes completely ignore the work done by the straggling workers resulting in a waste of computational resources. To reduce the amount of work left unfinished at workers, one can further decompose the matrix multiplication task into smaller sub-tasks, and assign multiple sub-tasks to each worker, possibly heterogeneously, to better fit their particular storage and computation capacities. In this work, we propose a novel family of bivariate polynomial codes to efficiently exploit the work carried out by straggling workers. We show that bivariate polynomial codes bring significant advantages in terms of upload communication costs and storage efficiency, measured in terms of the number of sub-tasks that can be computed per worker. We propose two bivariate polynomial coding schemes. The first one exploits the fact that bivariate interpolation is always possible on a rectangular grid of evaluation points. We obtain such points at the cost of adding some redundant computations. For the second scheme, we relax the decoding constraints and require decodability for almost all choices of the evaluation points. We present interpolation sets satisfying such decodability conditions for certain storage configurations of workers. Our numerical results show that bivariate polynomial coding considerably reduces the average computation time of distributed matrix multiplication. We believe this work opens up a new class of previously unexplored coding schemes for efficient coded distributed computation.

关键词： Bivariate polynomial interpolation coded computation distributed computation distributed matrix multiplication polynomial codes

来源：评论

学校读者我要写书评

暂无评论

Recent Developments in distributed Particle Filtering: Towards Fast and Accurate Algorithms

引用

IFAC Proceedings Volumes 2009年第20期42卷 138-143页

作者： Andrea Simonetto Tamás Keviczky Delft Center for Systems and Control Delft University of Technology The Netherlands

Particle filters have been widely used for the solution of optimal estimation problems in nonlinear non-Gaussian environments. One of their drawbacks is that these methods are computationally expensive. In the past few years, new developments have been made in trying to distribute the particle filter algorithm among different computing agents in order to make the underlying computations tractable. This period also witnessed the rise of general purpose GPU devices, which are making massive code parallelization possible. These developments have the potential to make the particle filter a viable alternative for real-time implementations in the near future, even when the number of required particles is high. In this paper we review the state-of-the-art in distributed particle filtering and propose a method that is applicable to distributed computing architectures.

关键词： Particle filters distributed estimation algorithms distributed particle filters consensus filters distributed computation mobile robot tracking general purpose GPU

来源：评论

学校读者我要写书评

暂无评论

distributed virtual reality system based on hybrid rendering

引用

Progress in Natural Science:Materials International 2000年第3期 60-66+83页

作者：郑文庭鲍虎军彭群生孙汉秋 China Department of Computer Science and Engineering Hangzhou 310027 NT Shatin State Key Laboratory of CAD&CG Zhejiang University the Chinese University of Hong Kong

A distributed virtual reality system is presented based on a hybrid rendering scheme. This system first predicts the camera position of the next key frame in real-time according to the tendency of the current camera movement. The predicted key frame is then generated using a geometry-based algorithm. Finally, a real-time backward image warping algorithm is employed to create a smooth transition between the two adjacent key frames. Since the rendering procedure can be divided into two independent steps, this system can be efficiently implemented in distributed mode.

关键词： virtual reality texture mapping distributed computation parallel computation.

来源：评论

学校读者我要写书评

暂无评论

Probabilistic localization in sensor networks using distributed Kalman Filter

引用

IFAC Proceedings Volumes 2010年第16期43卷 366-371页

作者： Maurizio Di Rocco Federica Pascucci Dip. Informatica e Automazione Università degli Studi “Roma Tre” Via della Vasca Navale 79 00146 Roma Italy

In recent years sensor networks have interested fields such as environment monitoring, surveillance and other distributed applications for data elaboration. This interest has been based on the decentralized approach in treating the information. However it is still a challenge to manipulate such streams of data when the dimension of the net becomes large despite computational capabilities and consumption constraints. In most of applications, location awareness is fundamental to accomplish common tasks. In this paper a probabilistic approach to solve localization problem in wireless sensor networks is presented. The algorithm, based on the Kalman Filter, estimates the sensors' location by an adaptive behavior. The technique proposed allows a reduction of the computation burden respect to the traditional Kalman Filter showing, as explained in simulations and real world experiments, good performances.

关键词： localization extended kalman filter sensor networks distributed computation Bayesian filtering

来源：评论

学校读者我要写书评

暂无评论

Analysis of Known Linear distributed Average Consensus Algorithms on Cycles and Paths

引用

SENSORS 2018年第4期18卷 968-968页

作者： Gutierrez-Gutierrez, Jesus Zarraga-Rodriguez, Marta Insausti, Xabier Univ Navarra Dept Biomed Engn & Sci Tecnun Manuel Lardizabal 13 San Sebastian 20018 Spain

In this paper, we compare six known linear distributed average consensus algorithms on a sensor network in terms of convergence time (and therefore, in terms of the number of transmissions required). The selected network topologies for the analysis (comparison) are the cycle and the path. Specifically, in the present paper, we compute closed-form expressions for the convergence time of four known deterministic algorithms and closed-form bounds for the convergence time of two known randomized algorithms on cycles and paths. Moreover, we also compute a closed-form expression for the convergence time of the fastest deterministic algorithm considered on grids.

关键词： average consensus algorithms distributed computation sensor networks convergence time number of transmissions

来源：评论

学校读者我要写书评

暂无评论

Sequential LQG optimization of hierarchically structured systems

引用

IFAC Proceedings Volumes 1988年第19期21卷 115-121页

作者： S.S. Stanković D.D. Šiljak Faculty of Electrical Engineering University of Belgrade 1100 Belgrade Yugoslavia Santa Clara University Santa Clara CA 95053 USA

Sequential LQG algorithms are formulated for stochastic linear discrete-time systems having a Lower Block Triangular (LBT) decomposition. Three distinct information structures are discussed from the point of view of c... 详细信息

关键词： Large-scale systems Stochastic control Information structure constraints Sparsity distributed computation

来源：评论

学校读者我要写书评

暂无评论

distributed Relaxation-Type Algorithms for Noncooperative Games

引用

IFAC Proceedings Volumes 1989年第5期22卷 13-18页

作者： T. Başar Decision and Control Laboratory Coordinated Science Laboratory and the Department of Electrical and Computer Engineering University of Illionis 1101W Springfield Avenue Urbana IL 61801 USA

The main objective of this paper is to review some recent results and also to present some new algorithms for the on-line, distributed computation of equilibria in noncooperative deterministic and stochastic games. The algorithms to be studied are of the relaxation-type , which require memory on either the past policy choices or the past actions of the players, depending on whether the iteration (and learning) is in the policy or the action space. Our earlier work on this topic has shown that incorporating memory into Jacobi and Gauss-Seidel type iterations in two-person Nash games could lead to considerable improvements in both the region of convergence and the speed of convergence towards noncooperative equilibrium. under both asynchronous and totally asynchronous modes of update. The present paper shows that these appealing features carry over to partially asynchronous modes of update, when there is a bound on the memory available to one or more players. The paper also discusses robustness of the partially asynchronous relaxation-type algorithms to inaccuracies in each player's knowledge regarding the others' response functions.

关键词： Noncooperative games distributed computation robustness Nash equilibrium

来源：评论

学校读者我要写书评

暂无评论

A Survey on Noncooperative Games and distributed Nash Equilibrium Seeking over Multi-Agent Networks

引用

CAAI Artificial Intelligence Research 2022年第1期1卷 8-27页

作者： Peng Yi Jinlong Lei Xiuxian Li Shu Liang Min Meng Jie Chen Department of Control Science and Engineering Tongji UniversityShanghai 201804China Shanghai Research Institute for Intelligent Autonomous Systems Shanghai 201210China

The work gives a review on the distributed Nash equilibrium seeking of noncooperative games in multi-agent networks,which emerges as one of the frontier research topics in the area of systems and control ***,we give the basic formulation and analysis of noncooperative games with continuous action spaces,and provide the motivation and basic setting for distributed Nash equilibrium *** we introduce both the gradient-based algorithms and best-response based algorithms for various type of games,including zero-sum games,aggregative games,potential games,monotone games,and multi-cluster *** addition,we provide some applications of noncooperative games.

关键词： noncooperative games multi-agent systems optimization and decision making cyber-physical systems Nash equilibrium distributed computation

来源：评论

学校读者我要写书评

暂无评论

Integrated Modeling, Simulation, and Optimization of Multibody Systems

引用

IFAC Proceedings Volumes 1995年第23期28卷 83-88页

作者： P. Eberhard D. Bestle Institute B of Mechanics University of Stuttgart D-70550 Stuttgart Germany

Design of nonlinear dynamic systems is a challenging task. Computers may help to analyze and optimize virtual prototypes in an earl design phase, but there is a lack of design tools. In this paper, a design concept on the basis of a multibody system approach is described. Programs for modeling, simulation, sensitivity analysis, optimization, and animation are operating under a common administration which ensures model consistency. The designer is supported by a graphical user interface, distributed computation improves efficiency.

关键词： Multibody Systems Optimization Sensitivity Analysis Concurrent System Design distributed computation

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：