Notes on Lecture Slides
You can access the lecture slides for this course on Canvas by navigating to the "Files" tab. These slides are currently only available to enrolled students, but we plan to release them to the public once the course has concluded.
Tentative schedule & reading
Week 1
4/1: Introduction
Reading: No reading
4/3: Background
Week 2
4/8: Removal-based explanations 1
4/10: Removal-based explanations 2
Week 3
4/15: Shapley values 1
4/17: Shapley values 2
Week 4
4/22: Propagation and gradient-based explanations 1
4/24: Propagation and gradient-based explanations 2 + Representation explainability
Week 5
4/29: Amortized optimization + Evaluating explanation methods 1
5/1: Evaluating explanation methods 2
Week 6: Inherently interpretable models
5/6: Inherently interpretable models 1
5/8: Inherently interpretable models 2
Week 7
5/13: Concept-based explanations
Reading: Concept Bottleneck Models, (Optional) Feature Visualization
5/15: Neuron interpretation
Week 8
5/20: Instance explanations 1
5/22: Instance explanations 2
Week 9:
5/27: Counterfactual explanations
5/29: LLM explainability
Week 10
6/3: XAI in practice I –
6/5: XAI in practice II –Model improvement, applications to healthcare