Course

Explainability –What and Why, When not to use explainability; Types of explainability and Taxonomy, Explainability in the model development process.

Visual Explanations – Transparent models and RuleFit, Partial Dependency Plots, Partial Dependency plot (PDP), Individual conditional expectations (ICE), Applications of these plots, Global Explanations – Surrogate Models, Feature Importance; Local Explanations – LIME, Shapley values.

Explaining Structured Data, Explaining Images and Unstructured Data and Text Deep Explainers, Time Series Explainers, LLM, Foundation Models.

Preamble

This course provides an Introduction to Explainable AI (XAI) through practical applications and real-world examples. Students will gain a basic proficiency in interpreting and explaining the decisions of ML and AI systems, in a transparent and understandable manner to humans. The course will cover various XAI techniques and algorithms, including rule-based models, feature importance analysis, model-agnostic approaches, and post-hoc explanations.

Course Objectives

• To understand how to explain machine learning models with various techniques.
• To explain various factors affecting the efficiency of machine learning models.
• To apply the explainability while applying ML models in practical applications.

Course Outcomes

Prerequisites

• Machine Learning, Python Programming

Prerequisites

• Machine Learning, Python Programming

CO-PO Mapping

Evaluation Pattern – 70:30

• Midterm Exam – 20%
• Quizzes – 20%
• Lab Assignments & Case Study – 30%
• End Semester Exam – 30%

Text Book / References

1. Denis Rothman “Hands-On Explainable AI (XAI with Python”, Packt Publishing, 2020
2. Explainable AI for Practitioners by Michael Munn, David Pitman, O’Reilly Media 2022, ISBN: 9781098119133
3. Explainable AI: Interpreting, Explaining and Visualizing Deep Learning, Lecture Notes in Artificial Intelligence, Springer Nature, 2019