On the failings of Shapley values for explainability

X Huang, J Marques-Silva - International Journal of Approximate …, 2024 - Elsevier
Abstract Explainable Artificial Intelligence (XAI) is widely considered to be critical for building
trust into the deployment of systems that integrate the use of machine learning (ML) models …

The inadequacy of shapley values for explainability

X Huang, J Marques-Silva - arxiv preprint arxiv:2302.08160, 2023 - arxiv.org
This paper develops a rigorous argument for why the use of Shapley values in explainable
AI (XAI) will necessarily yield provably misleading information about the relative importance …

A refutation of shapley values for explainability

X Huang, J Marques-Silva - arxiv preprint arxiv:2309.03041, 2023 - arxiv.org
Recent work demonstrated the existence of Boolean functions for which Shapley values
provide misleading information about the relative importance of features in rule-based …

Recent advances in formal explainability

X Huang - 2023 - theses.hal.science
In the past decade, monumental breakthroughs in Artificial Intelligence (AI), particularly in
Machine Learning (ML), have shaped various fields. The widespread integration of complex …

[PDF][PDF] Trustworthy AI at KDD Lab.

F Giannotti, R Guidotti, A Monreale, L Pappalardo… - Ital-IA, 2023 - ceur-ws.org
This document summarizes the activities regarding the development of Responsible AI
(Responsible Artificial Intelligence) conducted by the Knowledge Discovery and Data mining …