Mitigating belief projection in explainable artificial intelligence via Bayesian teaching

Abstract State-of-the-art deep-learning systems use decision rules that are challenging for humans to model. Explainable AI (XAI) attempts to improve human understanding but rarely accounts for how people typically reason about unfamiliar agents. We propose explicitly modelling the human explainee v...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Scott Cheng-Hsin Yang, Wai Keen Vong, Ravi B. Sojitra, Tomas Folke, Patrick Shafto
Formato:	article
Lenguaje:	EN
Publicado:	Nature Portfolio 2021
Materias:	Medicine R Science Q
Acceso en línea:	https://doaj.org/article/feea80d1ede04d4f84ba50138beb7648
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:feea80d1ede04d4f84ba50138beb7648
record_format	dspace
spelling	oai:doaj.org-article:feea80d1ede04d4f84ba50138beb76482021-12-02T17:01:49ZMitigating belief projection in explainable artificial intelligence via Bayesian teaching10.1038/s41598-021-89267-42045-2322https://doaj.org/article/feea80d1ede04d4f84ba50138beb76482021-05-01T00:00:00Zhttps://doi.org/10.1038/s41598-021-89267-4https://doaj.org/toc/2045-2322Abstract State-of-the-art deep-learning systems use decision rules that are challenging for humans to model. Explainable AI (XAI) attempts to improve human understanding but rarely accounts for how people typically reason about unfamiliar agents. We propose explicitly modelling the human explainee via Bayesian teaching, which evaluates explanations by how much they shift explainees’ inferences toward a desired goal. We assess Bayesian teaching in a binary image classification task across a variety of contexts. Absent intervention, participants predict that the AI’s classifications will match their own, but explanations generated by Bayesian teaching improve their ability to predict the AI’s judgements by moving them away from this prior belief. Bayesian teaching further allows each case to be broken down into sub-examples (here saliency maps). These sub-examples complement whole examples by improving error detection for familiar categories, whereas whole examples help predict correct AI judgements of unfamiliar cases.Scott Cheng-Hsin YangWai Keen VongRavi B. SojitraTomas FolkePatrick ShaftoNature PortfolioarticleMedicineRScienceQENScientific Reports, Vol 11, Iss 1, Pp 1-17 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Medicine R Science Q
spellingShingle	Medicine R Science Q Scott Cheng-Hsin Yang Wai Keen Vong Ravi B. Sojitra Tomas Folke Patrick Shafto Mitigating belief projection in explainable artificial intelligence via Bayesian teaching
description	Abstract State-of-the-art deep-learning systems use decision rules that are challenging for humans to model. Explainable AI (XAI) attempts to improve human understanding but rarely accounts for how people typically reason about unfamiliar agents. We propose explicitly modelling the human explainee via Bayesian teaching, which evaluates explanations by how much they shift explainees’ inferences toward a desired goal. We assess Bayesian teaching in a binary image classification task across a variety of contexts. Absent intervention, participants predict that the AI’s classifications will match their own, but explanations generated by Bayesian teaching improve their ability to predict the AI’s judgements by moving them away from this prior belief. Bayesian teaching further allows each case to be broken down into sub-examples (here saliency maps). These sub-examples complement whole examples by improving error detection for familiar categories, whereas whole examples help predict correct AI judgements of unfamiliar cases.
format	article
author	Scott Cheng-Hsin Yang Wai Keen Vong Ravi B. Sojitra Tomas Folke Patrick Shafto
author_facet	Scott Cheng-Hsin Yang Wai Keen Vong Ravi B. Sojitra Tomas Folke Patrick Shafto
author_sort	Scott Cheng-Hsin Yang
title	Mitigating belief projection in explainable artificial intelligence via Bayesian teaching
title_short	Mitigating belief projection in explainable artificial intelligence via Bayesian teaching
title_full	Mitigating belief projection in explainable artificial intelligence via Bayesian teaching
title_fullStr	Mitigating belief projection in explainable artificial intelligence via Bayesian teaching
title_full_unstemmed	Mitigating belief projection in explainable artificial intelligence via Bayesian teaching
title_sort	mitigating belief projection in explainable artificial intelligence via bayesian teaching
publisher	Nature Portfolio
publishDate	2021
url	https://doaj.org/article/feea80d1ede04d4f84ba50138beb7648
work_keys_str_mv	AT scottchenghsinyang mitigatingbeliefprojectioninexplainableartificialintelligenceviabayesianteaching AT waikeenvong mitigatingbeliefprojectioninexplainableartificialintelligenceviabayesianteaching AT ravibsojitra mitigatingbeliefprojectioninexplainableartificialintelligenceviabayesianteaching AT tomasfolke mitigatingbeliefprojectioninexplainableartificialintelligenceviabayesianteaching AT patrickshafto mitigatingbeliefprojectioninexplainableartificialintelligenceviabayesianteaching
_version_	1718382060259770368

Mitigating belief projection in explainable artificial intelligence via Bayesian teaching

Ejemplares similares