杭州电子科技大学图书馆书目检索系统

Online Public Access Catalogue

题名/责任者:

Data management in machine learning systems / Matthias Boehm, Arun Kumar, Jun Yang.

出版发行项:

[San Rafael, California] : Morgan and Claypool, 2019.

ISBN:

9781681734989

ISBN:

9781681734965

载体形态项:

xv, 157 pages : illustrations ; 24 cm.

附加统一题名:

Synthesis digital library of engineering and computer science.

丛编说明:

Synthesis lectures on data management, 2153-5426 ; # 57

个人责任者:

Boehm, Matthias, author.

附加个人名称:

Kumar, Arun, author.

附加个人名称:

Yang, Jun, 1975 September 9- author.

论题主题:

Database management.

论题主题:

Machine learning.

中图法分类号:

TP181

一般附注:

Part of: Synthesis digital library of engineering and computer science.

书目附注:

Includes bibliographical references (pages 127-156)

内容附注:

1. Introduction -- 1.1 Overview of ML lifecycle and ML users -- 1.2 Motivation -- 1.3 Outline and scope --

内容附注:

2. ML through database queries and UDFs -- 2.1 Linear algebra -- 2.2 Iterative algorithms -- 2.3 Sampling-based methods -- 2.4 Discussion -- 2.5 Summary --

内容附注:

3. Multi-table ML and deep systems integration -- 3.1 Learning over joins -- 3.2 Statistical relational learning and non-IID models -- 3.3 Deeper integration and specialized DBMSs -- 3.4 Summary --

内容附注:

4. Rewrites and optimization -- 4.1 Optimization scope -- 4.2 Logical rewrites and planning -- 4.3 Physical rewrites and operators -- 4.4 Automatic operator fusion -- 4.5 Runtime adaptation -- 4.6 Summary --

内容附注:

5. Execution strategies -- 5.1 Data-parallel execution -- 5.2 Task-parallel execution -- 5.3 Parameter servers (model-parallel execution) -- 5.4 Hybrid execution strategies -- 5.5 Accelerators (GPUs, FPGAs, ASICs) -- 5.6 Summary --

内容附注:

6. Data access methods -- 6.1 Caching and buffer pool management -- 6.2 Compression -- 6.3 NUMA-aware partitioning and replication -- 6.4 Index structures -- 6.5 Summary --

内容附注:

7. Resource heterogeneity and elasticity -- 7.1 Provisioning, configuration, and scheduling -- 7.2 Handling failures -- 7.3 Working with markets of transient resources -- 7.4 Summary --

内容附注:

8. Systems for ML lifecycle tasks -- 8.1 Data sourcing and cleaning for ML -- 8.2 Feature engineering and deep learning -- 8.3 Model selection and model management -- 8.4 Interaction, visualization, debugging, and inspection -- 8.5 Model deployment and serving -- 8.6 Benchmarking ML systems -- 8.7 Summary --

内容附注:

9. Conclusions -- Bibliography -- Authors' biographies.

摘要附注:

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.

豆瓣简介：