Prefrontal cortex as a meta-reinforcement learning system
Over the past twenty years, neuroscience research on reward-based learning has converged on a canonical model, under which the neurotransmitter dopamine ‘stamps in’ associations between situations, actions and rewards by modulating the strength of synaptic connections between neurons. However, a growing number of recent findings have placed this standard model under strain. In the present work, we draw on recent advances in artificial intelligence to introduce a new theory of reward-based learning. Here, the dopamine system trains another part of the brain, the prefrontal cortex, to operate as its own free-standing learning system. This new perspective accommodates the findings that motivated the standard model, but also deals gracefully with a wider range of observations, providing a fresh foundation for future research.
Date:
29 November 2017, 13:30 (Wednesday, 8th week, Michaelmas 2017)
Venue:
Le Gros Clark Building, off South Parks Road OX1 3QX
Venue Details:
Lecture theatre
Speaker:
Dr Matthew Botvinick (DeepMind and Gatsby Computational Neuroscience Unit, UCL)
Organisers:
Dr Friedemann Zenke (University of Oxford),
Dr Rafal Bogacz (University of Oxford),
Dr Tim Vogels
Organiser contact email address:
friedemann.zenke@cncb.ox.ac.uk
Part of:
Oxford Neurotheory Forum
Booking required?:
Not required
Audience:
Members of the University only
Editor:
Friedemann Zenke
