Programming ML.NET (Developer Reference)

✍ Scribed by Dino Esposito, Francesco Esposito

• Publisher: Microsoft Press
• Year: 2022
• Tongue: English
• Leaves: 562
• Edition: 1
• Category: Library

✦ Synopsis

With .NET 5’s ML.NET and Programming ML.NET, any Microsoft .NET developer can solve serious machine learning problems, increasing their value and competitiveness in some of today’s fastest-growing areas of software development. World-renowned Microsoft development expert Dino Esposito covers everything you need to know about ML.NET, the machine learning pipeline, and real-world machine learning solutions development. 

Modeled on his popular Programming ASP.NET books, this guide takes the same scenario-based approach Microsoft’s team used to build the ML.NET framework itself. Esposito presents and illuminates ML.NET’s dedicated mini-frameworks (“ML Tasks”) for specific classes of problems, and draws on personal experience to help developers apply these in the real world, where a problem’s complexity can vary widely based on data availability or the specific results you need. In a full section on ML.NET neural networks, Esposito introduces key concepts and presents realistic examples you can reuse in your own applications. Along the way, Esposito also shows how to leverage powerful Python-based machine learning tools in the .NET environment. 

Programming ML.NET will help you add machine learning and artificial intelligence to your tool belt, whether you have a background in these high-demand technologies or not. 

✦ Table of Contents

Title Page
Copyright Page
Dedication Page
Contents
Contents at a Glance
Acknowledgments
Introduction
Who Should Read This Book?
Who Should Not Read This Book?
Organization of This Book
System Requirements
Code Samples
Errata, updates, & book support
Stay in Touch
Chapter 1. Artificially Intelligent Software
How We Ended Up with Software
The Role of Software Today
AI Is Just Software
Chapter 2. An Architectural Perspective of ML.NET
Life Beyond Python
Introducing ML.NET
Consuming a Trained Model
Summary
Chapter 3. The Foundation of ML.NET
On the Way to Data Engineering
The Data to Start From
The Training Step
Consuming the Model from a Client Application
Summary
Chapter 4. Prediction Tasks
The Pipeline and the Chain of Estimators
The Regression ML Task
Using the Regression Task
The ML Devil’s Advocate
Summary
Chapter 5. Classification Tasks
The Binary Classification ML Task
Binary Classification for Sentiment Analysis
The Multiclass Classification ML Task
Using the Multiclass Classification Task
The ML Devil’s Advocate
Summary
Chapter 6. Clustering Tasks
The Clustering ML Task
The ML Devil’s Advocate
Summary
Chapter 7. Anomaly Detection Tasks
What Is an Anomaly?
General Approaches to Detect Anomalies
The Anomaly Detection ML Task
The ML Devil’s Advocate
Summary
Chapter 8. Forecasting Tasks
Predicting the Future
The Forecast ML Task
The ML Devil’s Advocate
Summary
Chapter 9. Recommendation Tasks
Inside Information Retrieval Systems
The ML Recommendation Task
ML Devil’s Advocate
Summary
Chapter 10. Image Classification Tasks
Transfer Learning
Transfer Learning via Composition
The ML Image Classification Task
The ML Devil’s Advocate
Summary
Chapter 11. Overview of Neural Networks
Feed-forward Neural Networks
More Sophisticated Neural Networks
Summary
Chapter 12. A Neural Network to Recognize Passports
Using Azure Cognitive Services
Crafting Your Own Neural Network
The ML Devil’s Advocate
Summary
Appendix A. Model Explainability
Software Intelligence
The Super Theory of Artificial Intelligence
Machine Learning Black Boxes
Interpretability and Explainability
Explainability Techniques
Conclusion
Index
Code Snippets