↓Skip to main content

Axes

The SyCoSMA team is organized along 2 axes.

  • Artificial Cognitive Systems

This axis addresses questions of intelligence, the intelligibility of knowledge, the explicability of processes and the ethics of decisions. Inspired by Piaget’s work on developmental learning, the SyCoSMA team is exploring how an agent, with no a priori knowledge, can learn a predictive sensorimotor model of the world. Developed incrementally and autonomously, this model enables the agent to anticipate its actions and adapt to changes in its environment.

  • Multi-Agent Systems

Multi-agent coordination models target collective problem-solving. To this end, the SyCoSMA team is developing techniques for argumentation, automatic negotiation and the generation of coalition structures to enable scaling up. The team also combines the techniques of multi-agent deep reinforcement learning and developmental learning to make the exploration process more robust.

The various projects the team leads, or is involved in, listed below, belong to one or both of these axes.

Projects

  1. Cognitive Systems ECIÉA Scientific Lead: CHAPUT Rémy
    Ethics Reinforcement Learning Argumentation
  2. Cognitive Systems; Multi-Agent Systems CADE-AI - Collaborative Argumentation and Debate Enhancement with AI Scientific Lead: YUN Bruno
    Argumentation Large Language Models Debates
  3. Multi-Agent Systems GAAMAS Scientific Lead: MORGE Maxime
    Evaluation of Generative Agents Cooperation and Coordination
  4. Cognitive Systems SSL4LP Scientific Lead: LEFORT Mathieu
    Self-Supervised Learning Graph Link Prediction
  5. Cognitive Systems ATOS Scientific Lead: LEFORT Mathieu
    Multi-Label Classification Medical Scanner Data
  6. Cognitive Systems SENS LEFORT Scientific Lead: LEFORT Mathieu
    Continual Learning Self-Supervised Learning
  7. Cognitive Systems MeSMRise Scientific Lead: LEFORT Mathieu
    Self-Supervised Learning Sensorimotor Theories Multimodal Fusion Active Learning Active Perception
  8. Cognitive Systems DATAWISE Scientific Lead: LEFORT Mathieu
    Self-Supervised Learning Transfer Learning Frugal AI
  9. Cognitive Systems; Multi-Agent Systems Assistance in designing complex technological tools Scientific Lead: ARMETTA
    Complex Technological Tools AI Assistance LLM
  10. Multi-Agent Systems RORES-CL - Rôles et réseaux sociaux dans les communautés en ligne. Scientific Lead: MORGE
    Agent-Based Simulation Agent Behaviour Social Network Analysis