建國科技大學圖書館 |

語系: 繁體中文

說明(常見問題)

圖書館個人資料蒐集告知聲明

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Data management in machine learning ...

Boehm, Matthias.

Data management in machine learning systems

紀錄類型:	書目-語言資料,印刷品 : 單行本
作者:	BoehmMatthias.,
其他作者:	KumarArun.,
其他作者:	YangJun.,
出版地:	[San Rafael, CA]
出版者:	Morgan & Claypool Publishers;
出版年:	c2019.
面頁冊數:	xv, 157 p.ill. : 24 cm.;
集叢名:	Synthesis lectures on data management57
標題:	Machine learning. -
標題:	Database management. -
摘要註:	Large-scale data analytics using machine learning (ML) underpins many modern data-driven applications. ML systems provide means of specifying and executing these ML workloads in an efficient and scalable manner. Data management is at the heart of many ML systems due to data-driven application characteristics, data-centric workload characteristics, and system architectures inspired by classical data management techniques. In this book, we follow this data-centric view of ML systems and aim to provide a comprehensive overview of data management in ML systems for the end-to-end data science or ML lifecycle. We review multiple interconnected lines of work: (1) ML support in database (DB) systems, (2) DB-inspired ML systems, and (3) ML lifecycle systems. Covered topics include: in-database analytics via query generation and user-defined functions, factorized and statistical-relational learning; optimizing compilers for ML workloads; execution strategies and hardware accelerators; data access methods such as compression, partitioning and indexing; resource elasticity and cloud markets; as well as systems for data preparation for ML, model selection, model management, model debugging, and model serving. Given the rapidly evolving field, we strive for a balance between an up-to-date survey of ML systems, an overview of the underlying concepts and techniques, as well as pointers to open research questions. Hence, this book might serve as a starting point for both systems researchers and developers. -- Provided by publisher.
ISBN:	9781681734965
內容註:	1. Introduction: Overview of ML lifecycle and ML users Motivation Outline and scope 2. ML through database queries and UDFs: Linear algebra Iterative algorithms Sampling-based methods Discussion Summary 3. Multi-table ML and deep systems integration Learning over joins Statistical relational learning and non-IID models Deeper integration and specialized DBMSs Summary 4. Rewrites and optimization: Optimization scope Logical rewrites and planning Physical rewrites and operators Automatic operator fusion Runtime adaptation Summary 5. Execution strategies: Data-parallel execution Task-parallel execution Parameter servers (model-parallel execution) Hybrid execution strategies Accelerators (GPUs, FPGAs, ASICs) Summary 6. Data access methods: Caching and buffer pool management Compression NUMA-aware partitioning and replication Index structures Summary 7. Resource heterogeneity and elasticity: Provisioning, configuration, and scheduling Handling failures Working with markets of transient resources Summary 8. Systems for ML lifecycle tasks: Data sourcing and cleaning for ML Feature engineering and deep learning Model selection and model management Interaction, visualization, debugging, and inspection Model deployment and serving Benchmarking ML systems Summary 9. Conclusions: Bibliography Authors' biographies.

Data management in machine learning systems
Boehm, Matthias.

Data management in machine learning systems / Matthias Boehm, Arun Kumar, Jun Yang. - [San Rafael, CA] : Morgan & Claypool Publishers, c2019.. - xv, 157 p. ; ill. ; 24 cm.. - (Synthesis lectures on data management ; 57).
1. Introduction: Overview of ML lifecycle and ML users.
Includes bibliographical references (p. 127-156)..
ISBN 9781681734965ISBN 1681734982
Machine learning.Database management.

Kumar, Arun.

Data management in machine learning systems
LDR:03707cam a2200241 450 001 390391
010 1 $a 9781681734965 $b pbk. $d NT1912
010 1 $a 1681734982 $b pbk.
100 $a 20190309d2019 k y0engy50 b
101 0 $a eng
102 $a us $b ca
105 $a a a 000yy
200 1 $a Data management in machine learning systems $f Matthias Boehm, Arun Kumar, Jun Yang.
210 $a [San Rafael, CA] $c Morgan & Claypool Publishers $d c2019.
215 1 $a xv, 157 p. $c ill. $d 24 cm.
225 2 $a Synthesis lectures on data management $v 57
320 $a Includes bibliographical references (p. 127-156).
327 1 $a 1. Introduction: Overview of ML lifecycle and ML users $a Motivation $a Outline and scope $a 2. ML through database queries and UDFs: Linear algebra $a Iterative algorithms $a Sampling-based methods $a Discussion $a Summary $a 3. Multi-table ML and deep systems integration $a Learning over joins $a Statistical relational learning and non-IID models $a Deeper integration and specialized DBMSs $a Summary $a 4. Rewrites and optimization: Optimization scope $a Logical rewrites and planning $a Physical rewrites and operators $a Automatic operator fusion $a Runtime adaptation $a Summary $a 5. Execution strategies: Data-parallel execution $a Task-parallel execution $a Parameter servers (model-parallel execution) $a Hybrid execution strategies $a Accelerators (GPUs, FPGAs, ASICs) $a Summary $a 6. Data access methods: Caching and buffer pool management $a Compression $a NUMA-aware partitioning and replication $a Index structures $a Summary $a 7. Resource heterogeneity and elasticity: Provisioning, configuration, and scheduling $a Handling failures $a Working with markets of transient resources $a Summary $a 8. Systems for ML lifecycle tasks: Data sourcing and cleaning for ML $a Feature engineering and deep learning $a Model selection and model management $a Interaction, visualization, debugging, and inspection $a Model deployment and serving $a Benchmarking ML systems $a Summary $a 9. Conclusions: Bibliography $a Authors' biographies.
330 $a Large-scale data analytics using machine learning (ML) underpins many modern data-driven applications. ML systems provide means of specifying and executing these ML workloads in an efficient and scalable manner. Data management is at the heart of many ML systems due to data-driven application characteristics, data-centric workload characteristics, and system architectures inspired by classical data management techniques. In this book, we follow this data-centric view of ML systems and aim to provide a comprehensive overview of data management in ML systems for the end-to-end data science or ML lifecycle. We review multiple interconnected lines of work: (1) ML support in database (DB) systems, (2) DB-inspired ML systems, and (3) ML lifecycle systems. Covered topics include: in-database analytics via query generation and user-defined functions, factorized and statistical-relational learning; optimizing compilers for ML workloads; execution strategies and hardware accelerators; data access methods such as compression, partitioning and indexing; resource elasticity and cloud markets; as well as systems for data preparation for ML, model selection, model management, model debugging, and model serving. Given the rapidly evolving field, we strive for a balance between an up-to-date survey of ML systems, an overview of the underlying concepts and techniques, as well as pointers to open research questions. Hence, this book might serve as a starting point for both systems researchers and developers. -- Provided by publisher.
410 0 $1 2001 $a Synthesis lectures on data management $v 57.
410 0 $1 2001 $a Synthesis lectures on data management $v 57
606 $a Machine learning. $2 lc $3 32680
606 $a Database management. $2 lc $3 27993
676 $a 006.31 $v 23
680 $a Q325.5 $b .B643 2019
700 1 $a Boehm $b Matthias. $3 381741
702 1 $a Kumar $b Arun. $3 381742
702 1 $a Yang $b Jun. $3 381743
801 0 $a cw $b CTU $c 20200710 $g AACR2