Pierre-Yves Oudeyer: the man who made robots curious
Date:
Changed on 25/03/2020
Pierre-Yves Oudeyer received a grant worth €1.5 million for his project. An unusual career between high-tech fundamental research and experimental applications accessible to the general public, a revival of robotics and a journey to the sources of human language. Pierre-Yves Oudeyer had some weighty plus points when he presented his project to the ERC. Through the grant obtained in autumn 2009, the head of the FLOWERS team was promoted European "scientific leader" in robotics.
On leaving the prestigious university, ENS-Lyon, Pierre-Yves Oudeyer was enthusiastic about the question of language in robots. A question pulling together robotics, linguistics and biology as well as the brand new sciences of complexity.
At Sony's Computer Science Laboratory, Pierre-Yves Oudeyer spent 8 years of his career working on speech and its original structuring into a shared system of sounds. The tools he uses are the methods supplied by work on robotics, artificial intelligence and computer science modelling. Alongside Luc Steels and Frédéric Kaplan, Pierre-Yves Oudeyer therefore attempted to define the emergence of spoken language and bring new arguments to the scientific debate between linguists. He tested his hypotheses by getting robot dogs to talk to each other, a rather entertaining experiment which proved very popular with the general public.
Enhancing his team's efforts thanks to the ERC grant
Leading the Inria FLOWERS team since 2008, he continued his research into developmental and social robotics. A new structure and new challenges, this time involving designing robots able to learn new things over the long term, with no engineer intervention "after they leave the factory", by combining artificial curiosity and social interaction with humans… With the motto "humanising machines rather than robotising humans", this time he has used new, related disciplines, namely neurosciences and developmental psychology. This has led him to observe learning behaviour in children in order to reproduce it.
Are robot pets imminent?
Thanks to the ERC grant, Pierre-Yves Oudeyer will be able to enhance his team's efforts on two crucial subjects for his approach, i.e. motor learning (perception and self-awareness) and lexicon learning (naming things). With their accompanying related problems around object recognition and the design of appropriate human-robot interfaces. Offering the promise of results that could be decisive in the development of personal robots, and in particular personal support and assistance technologies. To successfully complete his research, Pierre-Yves Oudeyer has also already launched a talent call on his personal pages…"