BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Data efficient reinforcement learning | Rowan McAllister (UC Berke
 ley and Toyota) - Rowan McAllister (UC Berkeley and Toyota)
DTSTART:20210308T160000Z
DTEND:20210308T170000Z
UID:TALK157894@talks.cam.ac.uk
CONTACT:74143
DESCRIPTION:**ABSTRACT**\nData-efficiency is useful in robotic learning\, 
 where real-world data can be expensive and time-consuming to acquire. Prob
 abilistic dynamics models can help accelerate learning by mitigating overf
 itting and providing richer supervision signals than model-free control me
 thods. An additional benefit of probabilistic models is their ability to d
 etect out-of-distribution events\, useful in certain safety-critical setti
 ngs where control should not deviate from demonstration data. This talk in
 vestigates how deep probabilistic models can benefit learning safe control
  fast\, when either learning from scratch\, or from imitation data.\n\n**S
 PEAKER BIO**\nRowan McAllister is a research scientist at Toyota Research 
 Institute. His research is concerned with probabilistic modelling for data
 -efficient learning of control\, often with autonomous vehicle application
 s in mind. Rowan received a PhD from Cambridge in 2017 and was a postdocto
 ral scholar at UC Berkeley until 2020.
LOCATION:https://cern.zoom.us/j/69762859104?pwd=ekxRSERPYUN6QVc4QnoydHZOc2
 hGdz09
END:VEVENT
END:VCALENDAR