Research

My research work focuses on collective decision-making and multiagent planning. My main topics of interest are:

• Multiagent resource allocation: A first part of my research work focuses on resource allocation and lies at the intersection of computational social choice and distributed decision-making in multi-agent systems. Solving a multi-agent resource allocation problem consists in allocating one or more resources to agents in such a way as to maximize the efficiency or fairness of the allocation. My work studies distributed mechanisms to allocate a set of (indivisible) resources among a set of agents. The vast majority of my work focuses on the fairness of the allocation process and of the outcomes. I'm interested in relaxations of the envy-freeness notions and extensions to account for broader contexts such as group fairness or diversity constraints.

• Planning, coordination, and multi-agent negotiation: The second major focus of my research concerns planning, coordination, and negotiation in multi-agent systems. A part of this research work thus focuses on models and algorithms that solve multi-agent planning problems under uncertainty.Decentralized Markov models (DEC-POMDPs and DEC-MDPs) provide a powerful mathematical framework to formalize and solve cooperative multiagent sequential decision problems in stochastic environments. However, these models suffer from a high complexity and limited time and space representations. I'm interested in bridging the gap between DEC-POMDP models and real-world applications by providing more suited and interpretable models and by defining scalable algorithms. I'm also interested in multiagent reinforcement learning (MARL) to cope with partially known environments. Recently, I studied multiagent and multi-objective reinforcement learning to learn ethical behaviors. My research work also deals with issues related to knowledge representation, belief revision, negotiation protocols, distributed coordination with limited knowledge...

• Argumentation-based reasoning: The third area of my research work concerns argumentation-based reasoning. The objective of my work is to enable agents to exchange and debate in order to increase their knowledge and improve collective decision-making. I'm interested in developing strategic argumentative agents able to plan the sequence of moves to  put forward  in the debate. If a probabilistic model of the other agents exists, the decision-problem of the strategic agent can be viewed as a sequential decision problem under uncertainty and it can be modelled using Markovian models. My work also aims to enable better collective decision-making through structured debates. The objective is to use formal argumentation to ensure better organization of the debate by studying how the presentation of arguments can influence the perception of the debate. My goal is also to provide a more accurate summary of the debate and to be able to justify its outcome by explaining, for example, which contributors or arguments are the most influential.