A self-organising model of the Bayesian Brain with a robust Infomax cherry on the top - 2015 HMNS Seminar Series
Event Details
- Date:
-
Friday, 06 March 2015 - Friday, 06 March 2015
- Time:
-
1:00 pm - 2:00 pm
- Room:
- 127
- UQ Location:
-
Human Movement Studies Building (St Lucia)
- Event category(s):
-
Event Contact
Event Description
- Full Description:
- A self-organising model of the Bayesian Brain with a robust Infomax cherry on the top.
In order to interact with our environment, our brain is tasked with making accurate estimates about the current and future state of the world, such as where objects are relative to us and where they will be in the near future.
Although our various sensory systems can offer important cues upon which to build such an estimate, they are far from infallible and may even offer inconsistent estimates. One of the outstanding questions in neuroscience is how the brain integrates these sources of information to arrive at an optimal estimate. One thing that is clear is that such estimates are affected by both current sensory input and prior, often context-dependent, cues derived from past experience.
Bayesian inference has been offered as a framework within which to optimally integrate these myriad cues. Although the power and generality of its use is unquestioned, scientists currently lack a biologically plausible model of how the framework might be realised in the brain. I describe the basis for such a model. I go on to explain how the model offers a range of additional attractive emergent properties including robustness, adaptability and the maximal retention of information for a given neural resource. I close by describing the types of predictions such a model makes, and summarise behavioural evidence consistent with these predictions.
Presenter:
Guy Wallis is an Associate Professor in the School of Human Movement and Nutrition Sciences. He is also currently Director, Centre for Sensorimotor Performance. He studied electronic engineering and neural systems at Imperial College, London; followed by a PhD in visual neuroscience from Oxford University.
After completing his PhD, A/Professor Wallis spent three years as a visiting Research Fellow at the Max Planck Institute for Biological Cybernetics in S.W. Germany where he studied models of learning using computer-simulated environments. Upon his arrival in Australia, he established a virtual environments lab within the School.
His research work, and infrastructure to support it, has been funded by the Wellcome Trust, Human Frontiers Science Program, Australian Research Council and several industrial partners.
A/Professor Wallis held an ARC QEII Fellowship from 2003-2007 and an ARC Future Fellowship from 2010-2014. He was also one of seven recipients of a UQ Foundation Research Excellence Award in 2003.
Directions to UQ
Event Tools
Share This Event
Print
Email
Share
Rate This Event
Tweet This Event
Calendar Tools
Featured Calendars
Subscribe via RSS