BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Learning and memory in neural networks: statistically optimal comp
 utations - Dr Mate Lengyel (Computational & Biological Learning Lab\, Dept
  Engineering)
DTSTART:20081023T135000Z
DTEND:20081023T141500Z
UID:TALK13259@talks.cam.ac.uk
CONTACT:Duncan Simpson
DESCRIPTION:How do networks of neurons in our brain endow us with the capa
 city to learn from experience and remember our past? \n\nWe are using Baye
 sian statistical theory to answer a different but related question first: 
 what is the best way in which networks of neurons could behave in order to
  endow us with the capacity of learning and memory? Such normative theorie
 s were then turned into specific predictions about the spike timing-depend
 ent neural dynamics of hippocampal pyramidal cells\, and about the relatio
 nship between spontaneous and stimulus-evoked activity in visual cortical 
 cells and its change with visual experience. \n\nExperimental data collect
 ed by our collaborators confirmed our predictions in both cases\, thereby 
 suggesting that the brain may implement highly efficient computational str
 ategies for learning and memory.\n
LOCATION:Kaetsu Centre\, New Hall
END:VEVENT
END:VCALENDAR