ELI5

• Values: { explainability }
  • Explanation type: { white box gradient-based local surrogate perturbation }
• Categories: { model-specific model-agnostic }
• Stage: { in-processing post-processing }
• Repository: https://github.com/TeamHG-Memex/eli5
• Tasks: { classification NLP }
• Input data: { tabular text image }
• Licence: MIT
• Languages: { Python }
• Frameworks: { scikit-learn lightning XGBoost LightGBM CatBoost }

ELI5 (“Explain Like I’m 5”) provides model-specific support for models from scikit-learn, lightning, decision tree ensembles using the xgboost, LightGBM, CatBoost libraries. ELI5 mainly provides convenient wrappers to couple the feature importance coefficients that these libraries already provide with feature names, as well as convenient ways to visualize importances, e.g. by highlighting words in a text. For Keras image classifiers an implementation of the gradient-based Grad-CAM visualizations is offered, but the TensorFlow V2 backend is not supported.

Additionally, two model agnostic approaches are provided that work for black box models. The TextExplainer class offers an implementation of LIME , specifically tailored towards text classifiers (which is slightly counter-intuitive, since LIME is a model-agnostic approach). The PermutationImportance computes attribute based permutation importance as the “Mean Decrease Accuracy (MDA)”.