BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Computational Neuroscience Journal Club - Yul Kang (CBL)
DTSTART:20180626T150000Z
DTEND:20180626T160000Z
UID:TALK107764@talks.cam.ac.uk
CONTACT:Rodrigo Echeveste
DESCRIPTION:Yul Kang will cover:\n\n• The successor representation in hu
 man reinforcement learning\n\n• I. Momennejad\, E. M. Russek\, J. H. Che
 ong\, M. M. Botvinick\, N. D. Daw & S. J. Gershman\n\n• Nature Human Beh
 aviour (2017)\n\n• https://www.nature.com/articles/s41562-017-0180-8\n\n
 \nAbstract: Theories of reward learning in neuroscience have focused on tw
 o families of algorithms thought to capture deliberative versus habitual c
 hoice. ‘Model-based’ algorithms compute the value of candidate actions
  from scratch\, whereas ‘model-free’ algorithms make choice more effic
 ient but less flexible by storing pre-computed action values. We examine a
 n intermediate algorithmic family\, the successor representation\, which b
 alances flexibility and efficiency by storing partially computed action va
 lues: predictions about future events. These pre-computation strategies di
 ffer in how they update their choices following changes in a task. The suc
 cessor representation’s reliance on stored predictions about future stat
 es predicts a unique signature of insensitivity to changes in the task’s
  sequence of events\, but flexible adjustment following changes to rewards
 . We provide evidence for such differential sensitivity in two behavioural
  studies with humans. These results suggest that the successor representat
 ion is a computational substrate for semi-flexible choice in humans\, intr
 oducing a subtler\, more cognitive notion of habit. 
LOCATION:Cambridge University Engineering Department\, CBL\, BE4-38 (http:
 //learning.eng.cam.ac.uk/Public/Directions)
END:VEVENT
END:VCALENDAR