Summary: The theory and methods behind machine learning is now converging to a whole scientific corpus: the science of data. Following a similar naming, roboticists are promoting a science of movement. Indeed, in order to move, a robot must solve at each new control cycle a problem composed of several thousand of variables; this problem may, or may not, be explicitly set up. Robotics research aims at understanding how to artificially reproduce the processes that the biological brain uses to solve these problems which such a natural easiness. For all that, should robotics be reduced to one of the many problems of artificial intelligence? We would rather claim that only some of the tools used by these researchers belong to AI. In this presentation, we will show why the tools that we are developing at LAAS are different from the data-based approaches; how these tools take inspiration from the science of data; and how much we are expecting from stronger collaborations in this field.
Bio: Nicolas Mansard is a roboticist, Directeur de Recherche CNRS at LAAS in Toulouse. His research activities lie at the intersection of the fields of robotics, automatic control, signal processing and numerical mathematics. They deal with sensor-based control, and more specifically the integration of sensor-based schemes into humanoid robot applications. He received the 2015 CNRS Bronze Medal. His project ANR Entracte received the Grand Prix de l'ANR in 2016.
