Explainable AI

About this course

Learn how to choose, interpret and assess model explanations so that they support meaningful decisions and communicate limitations.

This advanced course is part of the AI & Generative AI school. Work through the curriculum below in a instructor-led / cohort format, with an estimated study time of 28 h.

Course curriculum

• Module 1 — Why and What to Explain
• Module 2 — Interpretable by Design
• Module 3 — Global Model-Agnostic Explanations
• Module 4 — Shapley Values and SHAP
• Module 5 — Local Explanations, Reasons and Recourse
• Module 6 — Explaining Deep Models: Images and Text
• Module 7 — Evaluating and Attacking Explanations
• Module 8 — Explanations in Operation

Applied learning

• Case study — Marlow Mutual: getting a credit model through approval
• Lab 1 — A global audit of two models
• Lab 2 — Reasons and recourse for declined applicants
• Lab 3 — Can you trust this explanation?
• Capstone — An Explainability Case

Who this course is for

Technology Professional; Data Professional; Engineer / Architect

Prerequisites

NE-AI-201

Course details

• Course code: NE-AI-305
• Level: Advanced
• Estimated duration: 28 h
• Delivery: Instructor-Led / Cohort