Lab 10: Explainability with Alibi

In this lab, you'll learn about the Alibi Explain library and implement global and local explanations of tabular and image classification models.

Alibi Explain is an open source Python library aimed at machine learning model inspection and interpretation. The focus of the library is to provide high-quality implementations of black-box, white-box, local and global explanation methods for classification and regression models.

Complete all the Deliverables mentioned below and show it to a TA for credit.

Deliverables

• Finish all the TODOs in Section 1
• Generate PD Plots for chosen features. Discuss your findings about model performance with the TA.
• Play around with parameters for Anchor. Show final results to the TA.

Getting Started

Clone this repository and follow instructions in the notebook.

Install Dependencies

For this assignment, make sure you have the required packages installed.

pip install -r requirements.txt

(If there are any major unsolvable issues prefer running this notebook on Google Colaboratory)

Possible Issues with installing Alibi

TypeError: issubclass() arg 1 must be a class
Solution: https://stackoverflow.com/questions/76313592/import-langchain-error-typeerror-issubclass-arg-1-must-be-a-class

If there's any more issues, please contact a TA to update this list (with a solution if its solved)

References

1. https://docs.seldon.io/projects/alibi/en/stable/examples/pdp_regression_bike.html
2. https://docs.seldon.io/projects/alibi/en/stable/methods/PartialDependence.html
3. https://docs.seldon.io/projects/alibi/en/stable/examples/anchor_image_imagenet.html