BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Teaching Robots to Drink\, Relax and Play Catch
DTSTART:20120119T171500
DTSTAMP:20260929T050418Z
UID:e974d1ac51bfc84161c521a403a931475323b53afe24514daaf2a078
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. Aude Billard\nTeaching robots to do skills is a more cha
 llenging task than it may appear at a first glance\, even when these robot
 s resemble us. In this presentation\, I will attempt to shed some light on
  some of the major challenges we face when transferring skills from humans
  to robots and will review recent works we did at LASA to allow robots to 
 learn tasks that require careful control of both the kinematic and dynamic
 s of motion. The algorithms are generic and can hence be ported to several
  types of robots\, from classical industrial robotic arms\, to multi-degre
 es of freedom humanoid robots. I will also emphasize the fact that learnin
 g is not just play and record. A key component of learning and adaptation 
 resides in the ability to generalize. In imitation learning\, generalizati
 on must come from very few examples to ensure that the duration of the tra
 ining phase is bearable to the human teacher. Moreover\, most approaches d
 iscard erroneous training examples\, hence forcing the human teacher to be
  very skilled at the task. There is an advantage to include errors in the 
 training to ensure robustness and better generalization. I will hence conc
 lude this presentation by discussing some promising avenues in learning fr
 om failed demonstrations. 
LOCATION:polydome
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR