Baylime: Bayesian local interpretable model-agnostic explanations

Baylime: Bayesian local interpretable model-agnostic explanations

Zhao, Xingyu (Heriot-Watt University)
Huang, Wei (University of Liverpool)
Huang, Xiaowei (University of Liverpool)
Robu, Valentin (TU Delft Algorithmics)
Flynn, David (Heriot-Watt University)

de Campos, Cassio (editor)
Maathuis, Marloes H. (editor)

2021

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 Grad- CAM) 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.

http://resolver.tudelft.nl/uuid:94f2d335-b03d-4555-b71d-7f50f59ee0d4

Uncertainty in Artificial Intelligence, 27-30 July 2021, Online, 161

37th International Conference on Uncertainty in Artificial Intelligence, 2021-07-26 → 2021-07-30

Proceedings of Machine Learning Research, 2640-3498, 161

Institutional Repository

conference paper

© 2021 Xingyu Zhao, Wei Huang, Xiaowei Huang, Valentin Robu, David Flynn