检索结果-内蒙古大学图书馆

6th IEEE International Conference on Data Science and Advanced Analytics (DSAA)

作者： Gardner, Steven Golovidov, Oleg Griffin, Joshua Koch, Patrick Thompson, Wayne Wujek, Brett Xu, Yan SAS Inst Inc Cary NC 27513 USA

ISBN: (纸本)9781728144931

Automated machine learning has gained a lot of attention recently. Building and selecting the right machine learning models is often a multi-objective optimization problem. General purpose machine learning software that simultaneously supports multiple objectives and constraints is scant, though the potential benefits are great. In this work, we present a framework called Autotune that effectively handles multiple objectives and constraints that arise in machine learning problems. Autotune is built on a suite of derivative-free optimization methods, and utilizes multi-level parallelism in a distributed computing environment for automatically training, scoring, and selecting good models. Incorporation of multiple objectives and constraints in the model exploration and selection process provides the flexibility needed to satisfy trade-offs necessary in practical machine learning applications. Experimental results from standard multi-objective optimization benchmark problems show that Autotune is very efficient in capturing Pareto fronts. These benchmark results also show how adding constraints can guide the search to more promising regions of the solution space, ultimately producing more desirable Pareto fronts. Results from two real-world case studies demonstrate the effectiveness of the constrained multi-objective optimization capability offered by Autotune.

关键词： Multi-objective Optimization Automated Machine Learning distributed computing system

来源：评论

学校读者我要写书评

暂无评论

RHiNET: A network for high performance parallel computing using locally distributed computers

RHiNET: A network for high performance parallel computing us...

引用

International Workshop on Innovative Architecture for Future Generation High-Performance Processors and systems

作者： Kudoh, T Nishimura, S Yamamoto, J Nishi, H Tatebe, O Amano, H Tsukuba Res Ctr RWCP Tsukuba Ibaraki 3050032 Japan

ISBN: (纸本)076950650X

A local area distributed computing system called RHiNET (RWCP High Performance Network), in which thousands of commodity personal computers and workstations can be connected by high speed optical interconnection is being developed. RHiNET consists of network interfaces, network switches and optical interconnections. RHiNET will provide a single system image to users and support high performance parallel computing. This paper presents an overview of the system.

关键词： PARALLEL PROCESSING (COMPUTERS) HIGH-PERFORMANCE distributed computer High Performance computing optical interconnections Workstations distributed computing system Network interfaces

来源：评论

学校读者我要写书评

暂无评论

KunPeng: Parameter Server based distributed Learning systems and Its Applications in Alibaba and Ant Financial 17

KunPeng: Parameter Server based Distributed Learning Systems...

引用

23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD)

作者： Zhou, Jun Li, Xiaolong Zhao, Peilin Chen, Chaochao Li, Longfei Yang, Xinxing Cui, Qing Yu, Jin Chen, Xu Ding, Yi Qi, Yuan Alan Ant Financial Serv Grp Hangzhou Zhejiang Peoples R China Alibaba Cloud Hangzhou Zhejiang Peoples R China

ISBN: (纸本)9781450348874

In recent years, due to the emergence of Big Data (terabytes or petabytes) and Big Model (tens of billions of parameters), there has been an ever-increasing need of parallelizing machine learning (ML) algorithms in both academia and industry. Although there are some existing distributed computing systems, such as Hadoop and Spark, for parallelizing ML algorithms, they only provide synchronous and coarse-grained operators (e.g., Map, Reduce, and Join, etc.), which may hinder developers from implementing more efficient algorithms. This motivated us to design a universal distributed platform termed KunPeng, that combines both distributed systems and parallel optimization algorithms to deal with the complexities that arise from large-scale ML. Specifically, KunPeng not only encapsulates the characteristics of data/model parallelism, load balancing, model sync-up, sparse representation, industrial fault-tolerance, etc., but also provides easy-to-use interface to empower users to focus on the core ML logics. Empirical results on terabytes of real datasets with billions of samples and features demonstrate that, such a design brings compelling performance improvements on ML programs ranging from Follow-the-Regularized-Leader Proximal algorithm to Sparse Logistic Regression and Multiple Additive Regression Trees. Furthermore, KunPeng's encouraging performance is also shown for several real-world applications including the Alibaba's Double 11 Online Shopping Festival and Ant Financial's transaction risk estimation.

关键词： Parameter Server distributed computing system distributed optimization algorithm

来源：评论

学校读者我要写书评

暂无评论

Strategy for dynamic scheduling on distributed computing systems

引用

Huazhong Ligong Daxue Xuebao/Journal Huazhong (Central China) University of Science and Technology 2001年第SUPPL. 1期29卷 87-89页

作者： Wang, D. Wei, J. Chen, Z. Xue, Z. College of Computer Sci. and Tech. HUST Wuhan 430074 China

The dynamic scheduling strategies based on homogeneous computing systems are analyzed. Global Dynamic distributed Scheduling (GDDS) strategy is discussed for heterogeneous distributed computing systems. GDDS uses processing coefficient to represent heterogeneity of computing systems and is executed distributively on every processor and utilizes the global load information of systems. GDDS possesses a low-overhead, highly effective and scalable scheduling strategy.

关键词： distributed computing system Dynamic scheduling Heterogeneous Processing coefficient

来源：评论

学校读者我要写书评

暂无评论

New Method to Use Idle Personal Computers for Solving Coherent Tasks

引用

AASRI Procedia 2014年 9卷 131-137页

作者： Anatoly Kalyaev Iakov Korovin Scientific Research Institute of Multiprocessor Computing Systems of Southern Federal University Chekov st. 2 Taganrog 347922 Russia

In this paper authors offer the new method for solving coherent tasks in distributed computing system, based on resources of personal computers. Parameters of such resources are dynamically varying and that makes it hard for their application in distributed computations. To achieve ability of effective usage of personal computers, proposed method uses multiagent approach: proactive agent controls every personal computer in distributed computing system, and process of task solving is dispatched decentralized by interactions of agents. To solve every incoming coherent task agents of the system unite into community and that makes it easier to dispatch and perform computations. The main benefit of proposed method is decreasing the price for creating and maintenance of distributed computing system.

关键词： distributed computing system multiagent system coherent tasks communities proactive agents decentralized system.

来源：评论

学校读者我要写书评

暂无评论

The distributed program reliability analysis on a star topology: Efficient algorithms and approximate solution

引用

IEICE TRANSACTIONS ON INFORMATION AND systemS 1999年第6期E82D卷 1020-1029页

作者： Chang, MS Chen, DJ Lin, MS Ku, KL Natl Chiao Tung Univ Inst Comp Sci & Informat Engn Hsinchu 30050 Taiwan Tamsui Oxford Univ Coll Dept Informat Management Taipei Taiwan Chung Shan Inst Sci & Technol Tao Yuan Taiwan

A distributed computing system consists of processing elements, communication links, memory units, data files, and programs. These resources are interconnected via a communication network and controlled by a distributed operating system. The distributed program reliability (DPR) in a distributed computing system is the probability that a program which runs on multiple processing elements and needs to retrieve data files from other processing elements will be executed successfully. This reliability varies according to 1) the topology of the distributed computing system, 2) the reliability of the communication edges, 3) the data files and programs distribution among processing elements, and 4) the data files required to execute a program. In this paper, we show that computing the distributed program reliability on a star distributed computing system is #P-complete. A polynomially solvable case is developed for computing the distributed program reliability when some additional file distribution is restricted on the star topology. We also propose a polynomial time algorithm for computing the distributed program reliability with approximate solutions when the star star topology has no the additional file distribution.

关键词： distributed program reliability distributed computing system algorithms

来源：评论

学校读者我要写书评

暂无评论

Language Structures and Management Method in a distributed Real-Time Environment

引用

IFAC Proceedings Volumes 1981年第1期14卷 103-113页

作者： M. Tsukamoto Electrotechnical Laboratory Tsukuba Ibaraki 305 Japan

Recent advances in hardware technology have made it economically feasible to construct micro computer controlled subsystems and to connect them to distributed real-time control systems. However it is difficult to integrate and develop the potential of subsystems without understanding or changing the existing systems. In this paper, we will introduce a new language concept and a management method to overcome the difficulties. Firstly, a concurrent programming concept, the guarding process (GP), is introduced. In the concept a program consists only of modules, each of which defines objects and servers. Using the concept, programs are described in a structured manner allowing the modules to detect incorrect sequences of interactions. Secondly, a management method that guarantees timing and response to modules, the deadline monitor (DM), is introduced. Servers may be specified with requirements of timing and response. DM fills these requirements and monitors servers' behavior. Finally, in order to deal with the exceptions raised by GP and DM, the exception handler is introduced. In the testing phase, GP reports to the handler why the algorithm is wrong nd how to use the module. The concepts are presented in a programming language and augmented by the implementation of a distributed robot system.

关键词： distributed computing system distributed computer control concurrent programming distributed programming modular programming path expression error detection exception handling critical real-time deadline mechanism performance monitor robot

来源：评论

学校读者我要写书评

暂无评论

A Different Approach for Allocating Tasks in a DCS Using A

A Different Approach for Allocating Tasks in a DCS Using A

引用

The Fourth International Conference on Parallel and distributed computing, Applications and Technologies

作者： Biplab Kumer Sarker Anil Kumar Tripathi Deo Prakash Vidyarthi Kuniaki Uehara Graduate School of Science and Technology Kobe University Department of Computer Engineering I.TBanaras Hindu University Department of Computer Science Banaras Hindu University

In a distributed computing systems(DCS) tasks may consist of multiple *** submitted to a DCS,are usually partitioned into different modules and the modules may be allocated to different processing nodes so as to achieve minimum turn around time of the tasks utilizing the maximum resources of the existing system such as CPU speed,memory capacities *** problem lies on how to obtain the optimal allocation of these multiple tasks by keeping in mind that no processing node is overloaded due to this *** paper proposes an algorithm A*RS using well-known A* which aims to reduce the search space and time for allocating the tasks by minimizing the turn around time of tasks in the way so that processing nodes do not become overloaded due to this *** experimental results justify the claims with necessary supports by comparing it with the earlier algorithm for multiple tasks allocation.

关键词： distributed computing system Task Allocation Inter Module Communication Multiple Task Allocation Load balancing

来源：评论

学校读者我要写书评

暂无评论

The Economics of Introducing the Peer-to-peer system of Storage and Processing of Protected Information at an Enterprise

The Economics of Introducing the Peer-to-peer System of Stor...

引用

International Conference on Complex systems: Control and Modeling Problems

作者： Alexey Nenashev Victor Khryashchev Samara State Technical University

ISBN: (纸本)9781728167015

The method of evaluating the effectiveness of the economic parameters for the implementation at industrial enterprises of "Peer-to-Peer distributed Storage and Processing of Protected Information in Corporate Networks," designed primarily to protect distributed computing and isolate private networks without limiting the possibilities of useful interaction in the framework of the technologies Industry 5.0 and Super Smart Society is proposed. Within the framework of these concepts, there is a gradual integration of production, logistics, sales, and information systems into a global information exchange system. The proposed system provides for effective counteraction to the threats using cryptography, protection against unauthorized access using biometrics, and an innovative protocol for information interaction and control of topology based on distributed ledger technologies, it optimizes the use of computing power and provides tools for distributed computing without changing the logic of existing software.

关键词： Industry 5.0 Super smart society distributed ledger Cybersecurity Performance evaluation Computer simulation Peer-to-peer networks Information security distributed computing systems distributed computing system Network security peer to peer file sharing computer information security Performance Evaluation Computer Simulation distributed computing economics Peer-to-peer network Confidential Information peer to peer Private networks Industrial enterprises unauthorized access

来源：评论

学校读者我要写书评

暂无评论

A Different Approach for Allocating Tasks in a DCS Using A

A Different Approach for Allocating Tasks in a DCS Using A

引用

Proceedings of The Fourth International Conference on Parallel and Distribyted computing,Applications and Technologies(第四届并行与分布式计算应用与技术国际会议)

作者： Kuniaki Uehara Biplab Kumer Sarker Anil Kumar Tripathi Deo Prakash Vidyarthi Graduate School of Science and Technology Kobe University Kobe 657-8501 Japan Department of Computer Engineering I.T Banaras Hindu University India Department of Computer Science Banaras Hindu University India

ISBN: (纸本)0780378407

In a distributed computing systems (DCS) tasks may consist of multiple modules. Tasks submitted to a DCS, are usually partitioned into different modules and the modules may be allocated to different processing nodes so as to achieve minimum turn around time of the tasks utilizing the maximum resources of the existing system such as CPU speed, memory capacities etc. The problem lies on how to obtain the optimal allocation of these multiple tasks by keeping in mind that no processing node is overloaded due to this allocation. This paper proposes an algorithm A*RS using well-known A* which aims to reduce the search space and time for allocating the tasks by minimizing the turn around time of tasks in the way so that processing nodes do not become overloaded due to this allocation. Our experimental results justify the claims with necessary supports by comparing it with the earlier algorithm for multiple tasks allocation.

关键词： distributed computing system Task Allocation Inter Module Communication Multiple Task Allocation Load balancing

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：