home · mobile · calendar · bactac · 2006-2007 · 

BACTAC - Fan

Unsupervised Learning in Neural Nets: Infomax and Others
Adrian Fan
PhD Candidate
10/24/2006
3:30pm-4:30pm

Neural Networks derive much of their design from biology. However, most error update functions used in supervised neural nets are biologically implausible. We will review some approaches that have looked at the neural network structure from the unsupervised perspective, mainly Linsker's Infomax principle and many derived works, as well as current applications in signal deconvolution, computer vision, and future avenues.

Department of Computer Science
University of Colorado Boulder
Boulder, CO 80309-0430 USA
webmaster@cs.colorado.edu
www.cs.colorado.edu
May 5, 2012 (14:24)
XHTML 1.0/CSS2
©2012