Jean-Michel Dussoux

Université Paris Nanterre, Temple College

Papers

Total Citations

H-Index

About

Jean-Michel Dussoux’s research lies at the intersection of human-robot interaction, language grounding, and curiosity-driven learning. His work explores how autonomous agents can acquire language not through pre-programmed commands, but through natural, social interactions—much like a child learns. In his highly cited 2019 paper, “Usage-Based Learning in Human Interaction With an Adaptive Virtual Assistant” (10 citations), Dussoux demonstrated how virtual assistants could move beyond rigid, pre-specified command structures to learn from real-time user interactions, making them more adaptive and intuitive. Expanding on this, his 2019 study “Language Grounding through Social Interactions and Curiosity-Driven Multi-Goal Learning” (6 citations) introduced a framework where reinforcement learning agents discover their own goals and reward functions, using language exposure to organize their learning trajectories. This work challenges conventional AI paradigms by emphasizing intrinsic motivation and social scaffolding. Dussoux’s contributions are particularly notable for bridging developmental psychology and machine learning, offering a path toward more human-like, socially aware artificial intelligence. His research continues to influence the design of adaptive virtual assistants and autonomous robots capable of lifelong learning through interaction.