Assignment #2: Model Explainability
Due Date: 23:59
Download [problems] [attachment]
Late Policy
- You have free 8 late days.
- You can use late days for assignments. A late day extends the deadline 24 hours.
- Once you have used all 8 late days, the penalty is 10% for each additional late day.
This assignment focuses on making black-box models more interpretable and explainable. You will:
- Model Training: Train a complex model (e.g., neural network, random forest) on a provided dataset
- Global Interpretability: Analyze feature importance and model behavior across the dataset
- Local Explanations: Generate explanations for individual predictions using LIME and SHAP
- Comparison: Compare different explanation methods and their insights
- User Study: Design a simple user study to evaluate explanation quality
Learning Objectives:
- Implement various explainability techniques
- Compare and contrast different explanation methods
- Evaluate the quality and usefulness of explanations
- Consider human factors in explainability
Deliverables:
- Implementation of explanation methods
- Comparative analysis report
- User study design and pilot results
Resources: