BayLIME: Bayesian Local Interpretable Model-Agnostic Explanations

TL;DR

BayLIME introduces a Bayesian extension to LIME, enhancing explanation consistency, robustness, and fidelity in XAI by leveraging prior knowledge and Bayesian reasoning, with demonstrated theoretical and experimental benefits.

Contribution

It presents BayLIME, a novel Bayesian framework that improves local explanations in XAI by integrating prior knowledge and Bayesian inference, outperforming existing methods.

Findings

BayLIME offers more consistent explanations across repeated runs.

BayLIME demonstrates increased robustness to kernel parameter variations.

BayLIME achieves higher explanation fidelity than LIME, SHAP, and GradCAM.

Abstract

Given the pressing need for assuring algorithmic transparency, Explainable AI (XAI) has emerged as one of the key areas of AI research. In this paper, we develop a novel Bayesian extension to the LIME framework, one of the most widely used approaches in XAI -- which we call BayLIME. Compared to LIME, BayLIME exploits prior knowledge and Bayesian reasoning to improve both the consistency in repeated explanations of a single prediction and the robustness to kernel settings. BayLIME also exhibits better explanation fidelity than the state-of-the-art (LIME, SHAP and GradCAM) by its ability to integrate prior knowledge from, e.g., a variety of other XAI techniques, as well as verification and validation (V&V) methods. We demonstrate the desirable properties of BayLIME through both theoretical analysis and extensive experiments.