Consistent Sufficient Explanations and Minimal Local Rules for explaining regression and classification models

Abstract

To explain the decision of any model, we extend the notion of probabilistic Sufficient Explanations (P-SE). For each instance, this approach selects the minimal subset of features that is sufficient to yield the same prediction with high probability, while removing other features. The crux of P-SE is to compute the conditional probability of maintaining the same prediction. Therefore, we introduce an accurate and fast estimator of this probability via random Forests for any data $(\mathbf{X},Y)$ and show its efficiency through a theoretical analysis of its consistency. As a consequence, we extend the P-SE to regression problems. In addition, we deal with non-discrete features, without learning the distribution of $\mathbf{X}$ nor having the model for making predictions. Finally, we introduce local rule-based explanations for regression/classification based on the P-SE and compare our approaches w.r.t other explainable AI methods. These methods are available as a Python package at \url{www.github.com/salimamoukou/acv00}.

arXiv commentAccepted at the 36th Conference on Neural Information Processing Systems (NeurIPS 2022)

References in corpus (6)

• Consistency of random forests
• On The Reasons Behind Decisions
• SHAFF: Fast and consistent SHApley eFfect estimates via random Forests
• Have We Learned to Explain?: How Interpretability Methods Can Learn to Encode Predictions in their Interpretations
• Estimation and Inference of Heterogeneous Treatment Effects using Random Forests
• On Locality of Local Explanation Models