检索结果-内蒙古大学图书馆

Fourth International Conference on Advances in Recent Technologies in Communication and Computing (ARTCom2012)

作者： G. Vijay Kumar V. Valli Kumari School of Computing K L University Guntur India Department of CS&SE AU College of Engineering Visakhapatnam India

ISBN: (纸本)9781634396226

A good number of parallel and distributed frequent pattern mining algorithms have been proposed so far for the large and/or distributed databases. Not only occurrence frequency of a pattern but also occurrence behavior (regularity) of a pattern may be treated as an emerging area in data mining research. So far some efforts have been made to mine regular patterns but there is no suitable algorithm exists to mine frequent-regular patterns in parallel and distributed environment. Therefore, in this paper we introduced a new method called PFRP-method (parallel Frequent Regular Pattern-method) to discover frequent-regular patterns in large databases using vertical data format which requires only one database scan. Our method works in parallel at each local site in order to reduce I/O cost and inter-process communication, generates all frequent-regular patterns in the final phase. Our experiment results show that our PFRP-method is highly efficient in large databases.

关键词： Frequent patterns Regular patterns Large databases Vertical data format parallel and distributed algorithms parallel Lines Database large database LARGE gene indium MINE data formats

来源：评论

学校读者我要写书评

暂无评论

GPU Strategies for Distance-Based Outlier Detection

引用

IEEE TRANSACTIONS ON parallel AND distributed SYSTEMS 2016年第11期27卷 3256-3268页

作者： Angiulli, Fabrizio Basta, Stefano Lodi, Stefano Sartori, Claudio Univ Calabria DIMES Dept Via P Bucci41C I-87036 Arcavacata Di Rende CS Italy Italian Natl Res Council Inst High Performance Comp & Networking Via P Bucci41C I-87036 Arcavacata Di Rende CS Italy Univ Bologna Dept Comp Sci & Engn Viale Risorgimento 2 I-40136 Bologna Italy

The process of discovering interesting patterns in large, possibly huge, data sets is referred to as data mining, and can be performed in several flavours, known as "data mining functions." Among these functions, outlier detection discovers observations which deviate substantially from the rest of the data, and has many important practical applications. Outlier detection in very large data sets is however computationally very demanding and currently requires high-performance computing facilities. We propose a family of parallel and distributed algorithms for graphic processing units (GPU) derived from two distance-based outlier detection algorithms: BruteForce and SolvingSet. The algorithms differ in the way they exploit the architecture and memory hierarchy of the GPU and guarantee significant improvements with respect to the CPU versions, both in terms of scalability and exploitation of parallelism. We provide a detailed discussion of their computational properties and measure performances with an extensive experimentation, comparing the several implementations and showing significant speedups.

关键词： Distance-based outliers outlier detection parallel and distributed algorithms

来源：评论

学校读者我要写书评

暂无评论

parallel newton two-stage multisplitting iterative methods for nonlinear systems

引用

BIT NUMERICAL MATHEMATICS 2003年第5期43卷 849-861页

作者： Arnal, J Migallón, V Penadés, J Univ Alicante Dept Ciencia Computac & Inteligencia Artificial E-03071 Alicante Spain

parallel Newton two-stage iterative methods to solve nonlinear systems are studied. These algorithms are based on both the multisplitting technique and the two-stage iterative methods. Convergence properties of these methods are studied when the Jacobian matrix is either monotone or an H-matrix. Furthermore, in order to illustrate the performance of the algorithms studied, computational results about these methods on a distributed memory multiprocessor are discussed.

关键词： parallel and distributed algorithms nonlinear systems numerical methods multisplitting methods two-stage methods distributed memory

来源：评论

学校读者我要写书评

暂无评论

Special Issue: Bio-Inspired Optimization Techniques for High Performance Computing Preface

引用

NEW GENERATION COMPUTING 2011年第2期29卷 125-128页

作者： Folino, Gianluigi Mastroianni, Carlo ICAR CNR I-87036 Arcavacata Di Rende CS Italy

来源：评论

学校读者我要写书评

暂无评论

distributed nearest neighbor-based condensation of very large data sets

引用

IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 2007年第12期19卷 1593-1606页

作者： Angiulli, Fabrizio Folino, Gianluigi Univ Calabria Dipartimento Elettr Informat & Sistemist I-87036 Arcavacata Di Rende CS Italy Italian Natl Res Council Inst High Performance Comp & Networking I-87036 Arcavacata Di Rende CS Italy

In this work, the parallel Fast Condensed Nearest Neighbor (PFCNN) rule, a distributed method for computing a consistent subset of a very large data set for the nearest neighbor classification rule is presented. In order to cope with the communication overhead typical of distributed environments and to reduce memory requirements, different variants of the basic PFCNN method are introduced. An analysis of spatial cost, CPU cost, and communication overhead is accomplished for all the algorithms. Experimental results, performed on both synthetic and real very large data sets, revealed that these methods can be profitably applied to enormous collections of data. Indeed, they scale up well and are efficient in memory consumption, confirming the theoretical analysis, and achieve noticeable data reduction and good classification accuracy. To the best of our knowledge, this is the first distributed algorithm for computing a training set consistent subset for the nearest neighbor rule.

关键词： classification parallel and distributed algorithms nearest neighbor rule data condensation

来源：评论

学校读者我要写书评

暂无评论

Iterated Vector Fields and Conservatism, with Applications to Federated Learning 33

Iterated Vector Fields and Conservatism, with Applications t...

引用

33rd International Conference on Algorithmic Learning Theory (ALT)

作者： Charles, Zachary Rush, Keith Google Res Mountain View CA 94043 USA

We study whether iterated vector fields (vector fields composed with themselves) are conservative. We give explicit examples of vector fields for which this self-composition preserves conservatism. Notably, this includes gradient vector fields of loss functions associated to some generalized linear models. In the context of federated learning, we show that when clients have loss functions whose gradient satisfies this condition, federated averaging is equivalent to gradient descent on a surrogate loss function. We leverage this to derive novel convergence results for federated learning. By contrast, we demonstrate that when the client losses violate this property, federated averaging can yield behavior which is fundamentally distinct from centralized optimization. Finally, we discuss theoretical and practical questions our analytical framework raises for federated learning.

关键词： Optimization federated learning dynamical systems parallel and distributed algorithms

来源：评论

学校读者我要写书评

暂无评论

An upper diagonal work distribution for LU decomposition on the HP SuperDome

An upper diagonal work distribution for LU decomposition on ...

引用

3rd International Conference on Communications in Computing

作者： Steck, T Meyer, GGL Johns Hopkins Univ Baltimore MD 21218 USA

ISBN: (纸本)1932415068

A new family of parameterized LU decomposition algorithms, JHU LU, is presented. For moderate size matrices, JHU LU is twice as fast as vendor-tuned ScaLAPACK pdgetrf In contrast to algorithms inspired by geometrically regular data distribution, JHU L U is based on an upper diagonal work distribution. Comparatively coarse-grained parallelism is used to reduce the total number of communications between processors while exploiting high per-processor performance available by using BLAS subroutines. Another key departure from ScaLAPACK and other libraries is a requirement that the algorithms finish with the resulting factors L and U residing in a single processor memory - in particular, the processor memory that instially stores the matrix A to be decomposed. During processing, JHU L U iteratively updates the data distribution with processors dropping off successively, much like a pipeline discharge. This allows computation to continue while gathering results. Algorithm parameters control the number of processors used, problem size, and block size. A new set of parameters controls the quantity of work done by each processor. JHU L U is implemented on an HP SuperDome system with 550 MHz processors and is compared to ScaLAPACK pdgetrf Using 16 processors, the time to factor a 1000 x 1000 matrix is 195 ms for JHU and 390 ms for ScaLAPACK. In addition, for matrices between 500 x 500 and 3000 x 3000, JHU L U achieves faster execution.

关键词： LU Gaussian elimination parallel and distributed algorithms high performance computing block cyclic decomposition

来源：评论

学校读者我要写书评

暂无评论

A Framework for parallelizing Hierarchical Clustering Methods

A Framework for Parallelizing Hierarchical Clustering Method...

引用

European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD)

作者： Lattanzi, Silvio Lavastida, Thomas Lu, Kefu Moseley, Benjamin Google Zurich Zurich Switzerland Carnegie Mellon Univ Tepper Sch Business Pittsburgh PA 15213 USA Washington & Lee Univ Comp Sci Dept Lexington VA 24450 USA

ISBN: (纸本)9783030461508;9783030461492

Hierarchical clustering is a fundamental tool in data mining, machine learning and statistics. Popular hierarchical clustering algorithms include top-down divisive approaches such as bisecting k-means, k-median, and k-center and bottom-up agglomerative approaches such as single-linkage, average-linkage, and centroid-linkage. Unfortunately, only a few scalable hierarchical clustering algorithms are known, mostly based on the single-linkage algorithm. So, as datasets increase in size every day, there is a pressing need to scale other popular methods. We introduce efficient distributed algorithms for bisecting k-means, k-median, and k-center as well as centroid-linkage. In particular, we first formalize a notion of closeness for a hierarchical clustering algorithm, and then we use this notion to design new scalable distributed methods with strong worst case bounds on the running time and the quality of the solutions. Finally, we show experimentally that the introduced algorithms are efficient and close to their sequential variants in practice.

关键词： Hierarchical clustering parallel and distributed algorithms Clustering Unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

Techniques for Graph Analytics on Big Data

Techniques for Graph Analytics on Big Data

引用

IEEE International Congress on Big Data

作者： Nisar, M. Usman Fard, Arash Miller, John A. Univ Georgia Dept Comp Sci Athens GA 30602 USA

ISBN: (纸本)9780769550060

Graphs enjoy profound importance because of their versatility and expressivity. They can be effectively used to represent social networks, web search engines and genome sequencing. The field of graph pattern matching has been of significant importance and has wide-spread applications. Conceptually, we want to find subgraphs that match a pattern in a given graph. Much work has been done in this field with solutions like Subgraph Isomorphism and Regular Expression matching. With Big Data, scientists are frequently running into massive graphs that have amplified the challenge that this area poses. We study the speedup and communication behavior of three distributed algorithms for inexact graph pattern matching. We also study the impact of different graph partitionings on runtime and network I/O. Our extensive results show that the algorithms exhibit excellent scalable behavior and min-cut partitioning can lead to improved performance under some circumstances, and can drastically reduce the network traffic as well.

关键词： graph analytics big data graph simulation parallel and distributed algorithms

来源：评论

学校读者我要写书评

暂无评论

Multi-Stream LDPC Decoder on GPU of Mobile Devices 9

Multi-Stream LDPC Decoder on GPU of Mobile Devices

引用

9th IEEE Annual Computing and Communication Workshop and Conference (CCWC)

作者： Amiri, Roohollah Mehrpouyan, Hani Boise State Univ Elect & Comp Engn Boise ID 83725 USA

ISBN: (纸本)9781728105543

Low-density parity check (LDPC) codes have been extensively applied in mobile communication systems due to their excellent error correcting capabilities. However, their broad adoption has been hindered by the high complexity of the LDPC decoder. Although to date, dedicated hardware has been used to implement low latency LDPC decoders, recent advancements in the architecture of mobile processors have made it possible to develop software solutions. In this paper, we propose a multi-stream LDPC decoder designed for a mobile device. The proposed decoder uses graphics processing unit (GPU) of a mobile device to achieve efficient real-time decoding. The proposed solution is implemented on an NVIDIA Tegra board as a system on a chip (SoC), where our results indicate that we can control the load on the central processing units through the multi-stream structure.

关键词： parallel and distributed algorithms Multiprocessor Architectures LDPC Decoder GPU Processing

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：