Event Details:
Zoom link center to Center members. Join the Center.
Rosa Cao
Stanford University
Abstract
How do computational approaches to understanding the brain relate to the kinds of evidence we have available? Neuroscientific tools allow us to probe the physical mechanisms by which phenomena of interest are generated. But what is the relationship between these physical mechanisms and computational models? For example: can learning the detailed anatomy and physiology of some part of the brain tell us what computations its performing? If so, how? Conversely, can discovering what computations must be performed provide a guide to what we observe in the brain? If so, what are the principles that connect computation to its neural implementation? One proposal is that representations provide a way to connect mechanism to computation. We seem to have conventions for assigning representational functions to certain neural activities, and perhaps also conventions for relating representations to computations. I’ll look at different ways to flesh out this proposal, and some challenges.
Related papers
[1] A Teleosemantic Approach to Information in the Brain