BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Bayesian Machine Learning for Controlling Autonomous Systems - Mar
 c Deisenroth\, Imperial College
DTSTART:20131217T140000Z
DTEND:20131217T150000Z
UID:TALK48890@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:Autonomous learning has been a promising direction in control 
 and robotics for more than a decade since learning models and controllers 
 from data allows us to reduce the amount of engineering knowledge that is 
 otherwise required.\nDue to their flexibility\, autonomous reinforcement l
 earning (RL) approaches typically require many interactions with the syste
 m to learn controllers. However\, in real systems\, such as robots\, many 
 interactions can be impractical and time consuming. To address this proble
 m\, current learning approaches typically require task-specific knowledge 
 in form of expert demonstrations\, pre-shaped policies\, or specific knowl
 edge about the underlying dynamics.\n\nIn the first part of the talk\, we 
 follow a different approach and speed up learning by efficiently extractin
 g information from sparse data. In particular\, we learn a probabilistic\,
  non-parametric Gaussian process dynamics model. By explicitly incorporati
 ng model uncertainty into long-term planning and controller learning our a
 pproach reduces the effects of model errors\, a key problem in model-based
  learning. Compared to state-of-the art RL our model-based policy search m
 ethod achieves an unprecedented speed of learning. We demonstrate its appl
 icability to autonomous learning in real robot and control tasks.\n\nIn th
 e second part of my talk\, we will discuss an alternative method for learn
 ing controllers based on Bayesian Optimization\, where it is no longer pos
 sible to learn models of the underlying dynamics. We successfully applied 
 Bayesian optimization to learning controller parameters for a bipedal robo
 t\, where modelling the dynamics is very difficult due to ground contacts.
  Using Bayesian optimization\, we sidestep this modelling issue and direct
 ly optimize the controller parameters without the need of modelling the ro
 bot's dynamics.\n
LOCATION:Auditorium\, Microsoft Research Ltd\, 21 Station Road\, Cambridge
 \, CB1 2FB
END:VEVENT
END:VCALENDAR