Explainable AI

The courses is for professionals and part of the programme MAISTR (hh.se/maistr) where participants can study the entire programme or individual courses. The course is part of the course track machine learning and is held online in English.

The course covers the following topics:

• Introduction to the multidisciplinary topics of explainable AI, what is XAI, why is it important, plus related terminologies
• Broad taxonomy of XAI methods including Intrinsic vs post hoc, model-specific vs model-agnostic, and local vs global
• Trade-off between accuracy and explainability, human-friendly explanations,
• Intrinsically explainable models including Linear Regression, Logistic Regression, Generalized Linear Model (GLM), Generalized Additive Model (GAM), and Decision Tree.
• XAI methods including, Partial Dependence Plot (PDP), Conformal Prediction, Individual Conditional Expectation (ICE), Feature Importance, Saliency Maps, Local Interpretable Model-Agnostic Explanations (LIME), SHAP, Integrated Gradient (IG)
• Evaluation of explainability