Machine Learning Systems: Designs That Scale

✍ Scribed by Jeff Smith

• Publisher: Manning Publications
• Year: 2017
• Tongue: English
• Leaves: 224
• Category: Library

✦ Synopsis

Summary

Machine Learning Systems: Designs that scale is an example-rich guide that teaches you how to implement reactive design solutions in your machine learning systems to make them as reliable as a well-built web app.

Foreword by Sean Owen, Director of Data Science, Cloudera

Purchase of the print book includes a free eBook in PDF, Kindle, and ePub formats from Manning Publications.

About the Technology

If you're building machine learning models to be used on a small scale, you don't need this book. But if you're a developer building a production-grade ML application that needs quick response times, reliability, and good user experience, this is the book for you. It collects principles and practices of machine learning systems that are dramatically easier to run and maintain, and that are reliably better for users.

About the Book

Machine Learning Systems: Designs that scale teaches you to design and implement production-ready ML systems. You'll learn the principles of reactive design as you build pipelines with Spark, create highly scalable services with Akka, and use powerful machine learning libraries like MLib on massive datasets. The examples use the Scala language, but the same ideas and tools work in Java, as well.

What's Inside

• Working with Spark, MLlib, and Akka
• Reactive design patterns
• Monitoring and maintaining a large-scale system
• Futures, actors, and supervision

About the Reader

Readers need intermediate skills in Java or Scala. No prior machine learning experience is assumed.

About the Author

Jeff Smith builds powerful machine learning systems. For the past decade, he has been working on building data science applications, teams, and companies as part of various teams in New York, San Francisco, and Hong Kong. He blogs (https://medium.com/@jeffksmithjr), tweets (@jeffksmithjr), and speaks (www.jeffsmith.tech/speaking) about various aspects of building real-world machine learning systems.

Table of Contents

PART 1 - FUNDAMENTALS OF REACTIVE MACHINE LEARNING

1. Learning reactive machine learning
2. Using reactive tools

PART 2 - BUILDING A REACTIVE MACHINE LEARNING SYSTEM

1. Collecting data
2. Generating features
3. Learning models
4. Evaluating models
5. Publishing models
6. Responding

PART 3 - OPERATING A MACHINE LEARNING SYSTEM

1. Delivering
2. Evolving intelligence

✦ Table of Contents

Machine Learning Systems......Page 1
brief contents......Page 4
contents......Page 6
foreword......Page 10
preface......Page 13
acknowledgments......Page 16
about this book......Page 18
How this book is organized......Page 19
Book forum......Page 20
Other online resources......Page 21
about the author......Page 22
about the cover illustration......Page 23
Part 1: Fundamentals of reactive machine learning......Page 24
Chapter 1: Learning reactive machine learning......Page 26
1.1.1 Building a prototype system......Page 27
1.2 Reactive machine learning......Page 30
1.2.1 Machine learning......Page 31
1.2.2 Reactive systems......Page 34
1.2.3 Making machine learning systems reactive......Page 38
1.2.4 When not to use reactive machine learning......Page 43
Chapter 2: Using reactive tools......Page 46
2.1 Scala, a reactive language......Page 47
2.1.1 Reacting to uncertainty in Scala......Page 48
2.1.2 The uncertainty of time......Page 49
2.2.1 The actor model......Page 52
2.2.2 Ensuring resilience with Akka......Page 54
2.3 Spark, a reactive big data framework......Page 57
Part 2: Building a reactive machine learning system......Page 64
Chapter 3: Collecting data......Page 66
3.1 Sensing uncertain data......Page 67
3.2.1 Maintaining state in a distributed system......Page 71
3.2.2 Understanding data collection......Page 75
3.3 Persisting data......Page 76
3.3.1 Elastic and resilient databases......Page 77
3.3.2 Fact databases......Page 78
3.3.3 Querying persisted facts......Page 80
3.3.4 Understanding distributed-fact databases......Page 85
3.4 Applications......Page 89
3.5 Reactivities......Page 90
Chapter 4: Generating features......Page 92
4.2 Extracting features......Page 94
4.3 Transforming features......Page 97
4.3.1 Common feature transforms......Page 99
4.3.2 Transforming concepts......Page 102
4.4 Selecting features......Page 103
4.5.1 Feature generators......Page 105
4.5.2 Feature set composition......Page 109
4.7 Reactivities......Page 113
Chapter 5: Learning models......Page 116
5.1 Implementing learning algorithms......Page 117
5.1.1 Bayesian modeling......Page 119
5.1.2 Implementing Naive Bayes......Page 121
5.2.1 Building an ML pipeline......Page 125
5.2.2 Evolving modeling techniques......Page 130
5.3 Building facades......Page 132
5.3.1 Learning artistic style......Page 133
5.4 Reactivities......Page 138
Chapter 6: Evaluating models......Page 140
6.1 Detecting fraud......Page 141
6.2 Holding out data......Page 142
6.3 Model metrics......Page 145
6.4 Testing models......Page 150
6.5 Data leakage......Page 152
6.6 Recording provenance......Page 153
6.7 Reactivities......Page 155
Chapter 7: Publishing models......Page 158
7.2 Persisting models......Page 159
7.3.1 Microservices......Page 164
7.3.2 Akka HTTP......Page 165
7.4 Containerizing applications......Page 167
7.5 Reactivities......Page 170
Chapter 8: Responding......Page 172
8.2 Building services with tasks......Page 173
8.3 Predicting traffic......Page 176
8.4 Handling failure......Page 180
8.5 Architecting response systems......Page 183
8.6 Reactivities......Page 185
Part 3: Operating a machine learning system......Page 188
Chapter 9: Delivering......Page 190
9.1 Shipping fruit......Page 191
9.2 Building and packaging......Page 192
9.3 Build pipelines......Page 193
9.4 Evaluating models......Page 194
9.5 Deploying......Page 195
9.6 Reactivities......Page 198
10.1 Chatting......Page 200
10.3 Reflex agents......Page 201
10.4 Intelligent agents......Page 203
10.5 Learning agents......Page 204
10.6.1 Reactive principles......Page 208
10.7 Reactivities......Page 209
10.7.1 Libraries......Page 210
10.7.2 System data......Page 211
10.8.1 Users......Page 213
10.8.2 System dimensions......Page 214
10.8.3 Applying reactive principles......Page 215
sbt......Page 218
Docker......Page 219
D......Page 220
K......Page 221
Q......Page 222
Z......Page 223