Salvato in:
Dettagli Bibliografici
Autori principali: Biradar, Gagan, Viswanathan, Vignesh, Zick, Yair
Natura: Preprint
Pubblicazione: 2021
Soggetti:
Accesso online:https://arxiv.org/abs/2109.03890
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913235946438656
author Biradar, Gagan
Viswanathan, Vignesh
Zick, Yair
author_facet Biradar, Gagan
Viswanathan, Vignesh
Zick, Yair
contents Explaining the decisions of black-box models is a central theme in the study of trustworthy ML. Numerous measures have been proposed in the literature; however, none of them take an axiomatic approach to causal explainability. In this work, we propose three explanation measures which aggregate the set of all but-for causes -- a necessary and sufficient explanation -- into feature importance weights. Our first measure is a natural adaptation of Chockler and Halpern's notion of causal responsibility, whereas the other two correspond to existing game-theoretic influence measures. We present an axiomatic treatment for our proposed indices, showing that they can be uniquely characterized by a set of desirable properties. We also extend our approach to derive a new method to compute the Shapley-Shubik and Banzhaf indices for black-box model explanations. Finally, we analyze and compare the necessity and sufficiency of all our proposed explanation measures in practice using the Adult-Income dataset. Thus, our work is the first to formally bridge the gap between model explanations, game-theoretic influence, and causal analysis.
format Preprint
id arxiv_https___arxiv_org_abs_2109_03890
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle Model Explanations via the Axiomatic Causal Lens
Biradar, Gagan
Viswanathan, Vignesh
Zick, Yair
Machine Learning
Explaining the decisions of black-box models is a central theme in the study of trustworthy ML. Numerous measures have been proposed in the literature; however, none of them take an axiomatic approach to causal explainability. In this work, we propose three explanation measures which aggregate the set of all but-for causes -- a necessary and sufficient explanation -- into feature importance weights. Our first measure is a natural adaptation of Chockler and Halpern's notion of causal responsibility, whereas the other two correspond to existing game-theoretic influence measures. We present an axiomatic treatment for our proposed indices, showing that they can be uniquely characterized by a set of desirable properties. We also extend our approach to derive a new method to compute the Shapley-Shubik and Banzhaf indices for black-box model explanations. Finally, we analyze and compare the necessity and sufficiency of all our proposed explanation measures in practice using the Adult-Income dataset. Thus, our work is the first to formally bridge the gap between model explanations, game-theoretic influence, and causal analysis.
title Model Explanations via the Axiomatic Causal Lens
topic Machine Learning
url https://arxiv.org/abs/2109.03890