BOVE Clara
Supervision : Marie-Jeanne LESOT
Co-supervision : Charles TIJUS (CHArt), Marcin DETYNIECKI (AXA)
Design and evaluation of explanation user interfaces for complex machine-learning systems
This thesis is set in the field of human-centered eXplicable AI (XAI), and more specifically on the intelligibility of explanations for non-expert users. The technical context is as follows: on the one hand, an opaque classifier or regressor provides a prediction, and a post-hoc XAI approach generates information that acts as explanations; on the other hand, the user receives both the prediction and these explanations. In this context, there are several issues that can limit the quality of explanations. Those we focus on are: the lack of contextual information in explanations, the lack of guidance for designing features to allow the user to explore the explanatory interface and the potential confusion that can be generated by the amount of information.
We develop an experimental procedure for designing explanation user interfaces (XUIs) and evaluating their intelligibility with non-expert users. We propose several generic XAI principles and their implementation in different XUIs for explanations expressed as feature importance scores and counterfactual examples. We evaluate the objective understanding and satisfaction of these interfaces through user studies. At a fundamental level, we consider the theoretical question of possible inconsistencies in these explanations. We study these inconsistencies and propose a typology to structure the most common inconsistencies in the literature.
Defence : 09/15/2023
Jury members :
VERBERT Katrien, Professeure des universités, Université KU Leuven [Rapporteur]
LABROCHE Nicolas, Maître de conférence HDR, Université de Tours [Rapporteur]
GANASCIA Jean-Gabriel, Professeur des universités, Sorbonne Université
EVGENIOU Theodoros, Professeur des universités, INSEAD
LESOT Marie-Jeanne, Professeure des universités, Sorbonne Université
TIJUS Charles, Professeur des universités, Université Paris 8
DETYNIECKI Marcin, Responsable de l’équipe Recherche, AXA
2021-2023 Publications
-
2023
- C. Bove : “Design and evaluation of explanation user interfaces for complex machine-learning systems”, thesis, phd defence 09/15/2023, supervision Lesot, Marie-Jeanne, co-supervision : Charles, TIJUS (CHArt), Marcin, DETYNIECKI (AXA) (2023)
- C. Bove, M.‑J. Lesot, Ch. Tijus, M. Detyniecki : “Investigating the Intelligibility of Plural Counterfactual Examples for Non-Expert Users: an Explanation User Interface Proposition and User Study”, IUI '23: Proceedings of the 28th International Conference on Intelligent User Interfaces (IUI ’23), Sydney, Australia, pp. 188-203, (ACM), (ISBN: 979-8-4007-0106-1) (2023)
-
2022
- C. Bove, J. Aigrain, M.‑J. Lesot, Ch. Tijus, M. Detyniecki : “Contextualization and Exploration of Local Feature Importance Explanations to Improve Understanding and Satisfaction of Non-Expert Users”, IUI '22: 27th International Conference on Intelligent User Interfaces, Helsinki, Finland, pp. 807-819, (ACM) (2022)
-
2021
- C. Bove, J. Aigrain, M.‑J. Lesot, Ch. Tijus, M. Detyniecki : “Contextualising local explanations for non-expert users: an XAI pricing interface for insurance”, Joint Proceedings of the ACM IUI 2021 Workshops co-located with 26th ACM Conference on Intelligent User Interfaces (ACM IUI 2021), vol. 2903, CEUR Workshop Proceedings, College Station, United States, (CEUR-WS.org) (2021)