Explainable AI for Practitioners (Early Release, Ch1&2/8)

✍ Scribed by Michael Munn; David Pitman

• Publisher: O'Reilly Media, Inc.
• Year: 2022
• Tongue: English
• Leaves: 74
• Category: Library

✦ Synopsis

Most intermediate-level machine learning books usually focus on how to optimize models by increasing accuracy or decreasing prediction error. But this approach often overlooks the importance and the need to be able to explain why and how your ML model makes the predictions that it does.
This practical guide brings together the best-in-class techniques for model interpretability and explains model predictions in a hands-on approach. Experienced ML practitioners will be able to more easily apply these tools in their daily workflow.

✦ Table of Contents

1. An Overview of Explainability
   What Are Explanations?
   Explainability Consumers
   Practitioners: Data Scientists and ML Engineers
   Observers: Business Stakeholders & Regulators
   End-Users: Domain Experts & Affected Users
   Types of Explanations
   Pre-modeling Explainability
   Intrinsic vs. Post-Hoc Explainability
   Local, Cohort, and Global Explanations
   Attributions, Counterfactual, and Example-based
   Themes Throughout Explainability
   Feature Attributions
   Surrogate Models
   Activation
   Putting It All Together
   Summary
2. Explainability for Image Data
   Integrated Gradients
   Choosing a Baseline
   Accumulating Gradients
   Improvements on Integrated Gradients
   XRAI
   How XRAI works
   Implementing XRAI
   Grad-CAM
   How Grad-CAM works
   Implementing Grad-CAM
   Improving Grad-CAM
   LIME
   How LIME Works
   Implementing LIME
   Guided Backpropagation and Guided Grad-CAM
   Guided Backprop and DeConvNets
   Guided Grad-CAM
   Summary

+6 chapters in full edition
About the Authors