BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Computational Neuroscience Journal Club - Laurence Aitchison (CBL)
DTSTART:20171128T160000Z
DTEND:20171128T170000Z
UID:TALK96304@talks.cam.ac.uk
CONTACT:Rodrigo Echeveste
DESCRIPTION:Laurence Aitchison will cover:\n\n• The hippocampus as a pre
 dictive map\n\n• Kimberly L Stachenfeld\, Matthew M Botvinick & Samuel J
  Gershman\n\n• Nature Neuroscience (October 2017)\n\n• https://www.nat
 ure.com/articles/nn.4650\n\nAbstract: A cognitive map has long been the do
 minant metaphor for hippocampal function\, embracing the idea that place c
 ells encode a geometric representation of space. However\, evidence for pr
 edictive coding\, reward sensitivity and policy dependence in place cells 
 suggests that the representation is not purely spatial. We approach this p
 uzzle from a reinforcement learning perspective: what kind of spatial repr
 esentation is most useful for maximizing future reward? We show that the a
 nswer takes the form of a predictive representation. This representation c
 aptures many aspects of place cell responses that fall outside the traditi
 onal view of a cognitive map. Furthermore\, we argue that entorhinal grid 
 cells encode a low-dimensionality basis set for the predictive representat
 ion\, useful for suppressing noise in predictions and extracting multiscal
 e structure for hierarchical planning.
LOCATION:Cambridge University Engineering Department\, CBL\, BE4-38 (http:
 //learning.eng.cam.ac.uk/Public/Directions)
END:VEVENT
END:VCALENDAR