Cosyne 2007 Workshops
February 26-27, 2007
The Canyons, Utah
Workshop Title
Neurally plausible statistical inference
Organizer(s)
Chris Eliasmith (Universty of Waterloo): celiasmith@uwaterloo.ca
Charles H. Anderson (Washington University, St. Louis): cha@wustl.edu
Brian Fischer (Caltech): fischerb@caltech.edu
Abstract
The workshop will present recent advances in the modeling of statistical representations and transformations in neural systems. We think this topic will be of interest at Cosyne because the neural plausiblity of many of the statistical methods adopted from machine learning is questionable. Given the effectiveness of biological inference, it seems worthwhile to incorporate biological constraints to discover the algorithms employed by real systems. In addition, there has recently been a dramatic increase in interest in understanding neural systems as centrally involved in statistical inference. The workshop will provide an opportunity to present a variety of perspectives on this idea, as well as address the issues of appropriate approximations to machine learning methods.
Speakers
| Charles Anderson (Washington University) | Principles of Computing with Redundant Population Codes |
| Jeff Beck (University of Rochester) | Bayesian inference with probabilistic population codes |
| Erick Chastain and Rajesh Rao (University of Washington) | Embodied Population Codes: Bayes-optimal combination of Inference and Action Selection in a neural population |
| Sophie Deneve (CNRS) | Bayesian inference and learning with networks of integrate and fire neurons |
| Chris Eliasmith and James Martens (Universty of Waterloo) | Challenges for Biological Bayes: Solving normalization |
| Brian Fischer (Caltech) | A model of sound localization in the barn owl using populations of spiking neurons |
| Mehrdad Jazayeri (New York University) | On the readout of sensory neural signals in perceptual tasks |
| Maneesh Sahani (University College London) | Title TBD |
