Learning Reasoning Computing Axiomatization Decision Estimation
Armand Gaudillier Picture of Armand Gaudillier Doctorant

LinkedIn

Direction de thèse : Wassila Ouerdane (MICS), Khaled Belahcène (MICS), Sébastien Destercke (Heudiasyc UTC)

Mots-Clefs : Aide à la décision multi-critères, Élicitation, Explication, Incertain, Théorie des possibilités, Préférences

PhD student since November 2024, my thesis topic is about the role of uncertainty in elicitation and explanation of preferences. In a framework where error can come from both the user (change of mind during the process, inattention errors) and our modelling, how can we adapt incremental elicitation and explanation procedures? Several avenues are open to us, in particular by using possibility theory, which offers an interesting framework for modelling uncertainty, as well as a plethora of tools for working on an inconsistent basis. In terms of explanation, the presence of uncertainty provides us with new criteria on which to measure the quality of the explanation. Ultimately, we would like to combine these two processes into a single one. This thesis is based on the work of former doctoral students, in particular that of Loïc Adam on elicitation, and of Hénoïk Willow and Manuel Amoussou on explanation.