Explainability in Machine Learning / July-Dec 2023
Updates
- Nov 20: New Lecture is up: (xml-15) Mechanistic Interpretability [slides]
- Nov 16: New Lecture is up: (xml-14) Causal Antehoc Explanations [slides]
- Nov 09: New Lecture is up: (xml-13) Generating Robust Counterfactuals (Algorithmic Recourse) [slides]
- Nov 08: New Lecture is up: (xml-12) Attention and Explanations [slides]
- Nov 06: New Lecture is up: (xml-11) Causal Inference and Explainability [slides]
- Sep 21: New Lecture is up: (xml-10) Interpreting Neural Networks (Concept-based Explanations) [slides]
- Sep 14: New Lecture is up: (xml-9) Interpreting Neural Networks (Pixel Attribution) [slides]
Course Description
Algorithmic decision-making is getting more prominent. Machine learning models are regularly helping us with some of the essential aspects of our lives (e.g., healthcare, banking, navigation, etc.). It is of high significance that the end-users correctly understand these models, thereby establishing trust in both the functionality and inference of these models. This new course at IIT Hyderabad introduces the basic notions of explainability and the recent advances in the emerging field of interpretability and explainability in machine learning..
(AI51020) Explainability in Machine Learning Course Contents
This course aims to understand the notion of interpretability and explainability of a machine learning model. The first part details different classes of interpretable models, such as prototype-based approaches, sparse linear models, rule-based techniques, generalized additive models, etc. Later, it discusses post-hoc explanations such as black-box explanations, counterfactual explanations, saliency maps, etc. Further, the course explores the connections between interpretability, causality, debugging, bias, and fairness. Throughout the course, there will be an emphasis on multiple applications, such as criminal justice, computer vision, healthcare, etc., that can immensely benefit from model interpretability. This course involves discussing seminal and recent work that is relevant in the form of student presentations and a semester-long project.
Logistics
Class Room: C-106
Timings: Slot-B (Monday-10:00-10:55, Wednesday-09:00-09:55, Thursday-11:00-11:55)
Visit this page regularly for updates and information regarding the course.
Instructors
Teaching Assistants
Pranav K Nayak (ES20BTECH11035@iith.ac.in)
Tanmay Garg (CS20BTECH11063@iith.ac.in)