Interpretable Machine Learning (IML)

This website offers an open and free introductory course on Interpretable Machine Learning. The course is constructed as self-contained as possible, and enables self-study through lecture videos, PDF slides, exercises (with solutions), and notebooks.

The course starts with a brief outline of some basic concepts, such as different dimensions of interpretability as well as correlation, dependence and interaction of features. Furthermore, common sources of error in the application and evaluation of ML models and IML methods are given. The content becomes more advanced as the chapters progress, so prior knowledge of machine learning and statistics is recommended. An introduction to ML can for instance be found here.

Chapter 1: IML intro
Chapter 2: Interpretable Models
Chapter 3: Feature Effects
Chapter 4: Shapley
Chapter 5: Feature Importance
Chapter 6: Local Interpretable Model-agnostic Explanations (LIME)
Chapter 7: Counterfactuals and Adversarial Examples