How to explain the decision process of a Machine Learning Algorithm.
-
Very fast practical intro to the topic using python (eli, shap, dependence plots)
https://www.kaggle.com/learn/machine-learning-explainability -
Captum: A library implementing the main ML explainability algorithms for PyTorch models
https://github.com/pytorch/captum, https://captum.ai/docs/introduction -
Shap: Similar to Captum but for all kind of ML models (trees, model agnostic) and for both PyTorch and Keras NNs.
https://github.com/slundberg/shap -
All papers in articles