Olivier Sigaud

About

Olivier Sigaud is a prominent French researcher whose work sits at the intersection of machine learning, developmental robotics, and autonomous learning systems. Over more than two decades, he has made substantial contributions to how robots acquire, refine, and generalize skills — spanning reinforcement learning, evolution strategies, and intrinsically motivated exploration. His highly cited 2011 survey on online regression algorithms for robot mechanical modeling (136 citations) established a foundational reference for the field, while his 2013 work on policy improvement methods (127 citations) helped clarify the theoretical landscape connecting reinforcement learning and evolutionary approaches to robot skill acquisition. Sigaud has been a persistent advocate for developmental and open-ended learning frameworks, exploring how robots can autonomously set goals, build curricula, and discover what is controllable in their environments — themes central to his influential 2018 contributions on goal space learning and the CURIOUS architecture. His research extends into human-robot interaction, multimodal perception, and active object exploration, reflecting a holistic vision of embodied machine intelligence. With a body of work accumulating hundreds of citations and spanning cognitive architectures, child-like robotic learning, and deep unsupervised representations, Sigaud remains a key voice shaping how autonomous agents learn in complex, open-ended worlds.

Research Focus

Key Achievements

Top Papers

1. 1
   On-line regression algorithms for learning mechanical models of robots: A survey

   136 citations · 2011
2. 2
   Robot Skill Learning: From Reinforcement Learning to Evolution Strategies

   127 citations · 2013
3. 3
   From Motor Learning to Interaction Learning in Robots

   73 citations · 2009
4. 4
   Anticipatory Behavior in Adaptive Learning Systems: Foundations, Theories, and Systems

   70 citations · 2003
5. 5
   Object Learning Through Active Exploration

   68 citations · 2014
6. 6
   Deep unsupervised network for multimodal perception, representation and classification

   63 citations · 2014
7. 7
   Open-Ended Learning: A Conceptual Framework Based on Representational Redescription

   61 citations · 2018
8. 8
   Unsupervised Learning of Goal Spaces for Intrinsically Motivated Goal Exploration

   49 citations · 2018
9. 9
   CURIOUS: Intrinsically Motivated Modular Multi-Goal Reinforcement Learning

   42 citations · 2018
10. 10
    Robot initiative in a team learning task increases the rhythm of interaction but not the perceived engagement

    41 citations · 2014

Key Collaborators

Contact & Links