Explaining Behavior of Machine Learning Models with ELI5 Library

Explore techniques for interpreting and debugging Machine Learning models in this 30-minute EuroPython Conference talk. Gain insights into explaining the behavior of linear models, decision trees, tree ensembles, and arbitrary classifiers using the LIME algorithm. Discover how to leverage the open-source eli5 library to interpret and debug estimators from popular Python ML libraries like scikit-learn and xgboost. Learn about the benefits of model interpretability, including easier debugging and increased human trust in model decisions. Acquire both practical and theoretical understanding of explanation methods to improve ML pipeline quality and effectively communicate model behavior to clients and stakeholders.

Mikhail Korobov - Explaining behavior of Machine Learning models with eli5 library