Can a robot learn like a child? Can it learn new skills and new knowledge in an unknown and changing environment? How can it discover its body and its relationships with the physical and social environment? How can it discover language through natural interactions with humans? How can its cognitive capacities continuously develop without the intervention of an engineer once it is "out of the factory"?

These are the questions that we are investigating in the FLOWERS research team at INRIA Bordeaux Sud-Ouest and ENSTA-Paristech in Paris. Rather than trying to imitate the intelligence of adult humans like in the field of Artificial Intelligence, we follow an idea formulated sixty years ago by Alan Turing, but that really began to be explored at the beginning of the 21st century: we try to reconstruct the processes of development of the child's mind, rooted in the dynamical interactions between its brain, its body and its environment. This approach is called developmental robotics, or epigenetic robotics, and imports concepts and theories from developmental sciences, in particular developmental psychology, developmental and cognitive neuroscience, biology and linguistics. This approach aims both at exploring how to build new kinds of machines, capable to adapt and learn robustly and efficiently in the real world, and at exploring new theories and understanding of animal and human development.

Our team, headed by Pierre-Yves Oudeyer, focuses in particular on the study of developmental constraints that allow for efficient open-ended learning of novel skills. In particular, we study constraints that guide exploration in large sensorimotor spaces:

1) Mechanisms of intrinsically motivated exploration and learning, including artificial curiosity;

2) Mechanisms for social learning, e.g. learning by imitation or demonstration, which implies both issues related to machine learning and human-robot interaction;

3) Constraints related to embodiment, in particular through the concept of morphological computation, as well as the structure of motor primitives/muscle synegies that can leverage the properties of morphology and physics;

4) Maturational constraints which, coupled with the other constraints, can allow the progressive release of novel sensorimotor degrees of freedom to be explored;

We also study how these constraints on exploration can allow a robot to bootstrap multimodal perceptual abstractions associated to motor skills, in particular in the context of modelling language acquisition as a developmental process grounded in action.

Highlight: The Ergo-Robot Experiment

Project web page: English, français