[Publication list below]
Publications
- My main research activities are positioned in the domain of decision modelling and preference learning, in particular when decision models involve multiple criteria. From a multiple criteria decision making perspective I have been studying both pairwise comparison (so-called outranking) methods, and value-based aggregation methods, but also on multi-criteria interactive methods.
- A large part of my research has been devoted to the study of algorithmic processes enabling to elicit a preference model from preference statement obtained from decision-makers. Such type of research amounts at preference learning, and aims at computing the set of preference parameters value that best match the input/learning set provided by the decision-maker. A typical example of such research has been published in [Mousseau,Slowinski 1998], and proposes an optimization based approach to learn the parameters of the Electre Tri method from assignment examples; this seminal work has been extended namely based on the axiomatic variant of Electre Tri (NCS, NonCompensatory Sorting). Concerning value-based aggregation methods, I have contributed to the robust ordinal regression so as to infer the partial ranking resulting from a set of pairwise comparison statements, see [Greco et al. 2008].
- This past decade, I also worked on explanability of recommendations produced by multi-criteria decision models. So as to increase trustworthiness of recommendations for users, the recommendations based preference models should come with interpretability and explanations. Convincing recommendations can not rely on technical elements of the model of preference, but on ``primitives'' used by the decision maker in their reasoning. An example of such contribution in the context of an additive value-based model can be found in [Amoussou et al. 2026].
- From a broader perspective, my research on preference elicitation and learning has contributed to bridge the gap between formerly distinct research communities : MCDM researchers that are operations research oriented, and researchers in preference learning, mostly coming from the IA and statistical learning community. Such link has been materialized by a serie of workshop that I initiated with Marc Pirlot in 2012: DA2PL, from Decision Analysis to Preference Learning. Since 2012, these workshops takes place every second year and gather ~50-70 participants. This perspective also explains the reason for which, during the last decade, I have been publishing my work in major international artificial Intelligence conferences (such as IJCAI, ECAI, AAMAS).
Publications