Description: NEURAL NETWORKS AND INFORMATION THEORY I
*****
Review of major Artificial Neural Network paradigms. Analytical discussion of supervised and unsupervised learning. Emphasis on state-of-the-art Hebbian (biologically most plausible) learning paradigms and their relation to information theoretical methods. Applications to data analysis such as pattern recognition, clustering, classification, blind source separation, non-linear PCA. More at http://www.ece.rice.edu/~erzsebet/ANNcourse.html.
*****
Also offered as COMP 502.
*****
Prerequisite(s):ELEC 430, 431, or equivalent or consent of instructor. Cross-list: COMP 502.