Tools for Modeling the Information Processing Flow in the Brain

Project Goals

ICA Approach To Image And Signal Processing. Independent component analysis (ICA) is a method allowing finding of underlying components from multivariate statistical data. It is a recently developed technique that in many cases characterizes the data in a natural way. ICA-based neural networks present a powerful technology for extraction of hidden predictive information from large databases and for image noise removal. Our current research is focused on:

ü  Development of a two-layer neural network ICA architecture maximizing the entropy of the outputs with logistic transfer function. Design of a neural architecture for nonlinear three-layer ICA network for estimating the respective basis vectors. They are counterparts of the PCA eigenvectors, but characterize the data in many cases better. The purpose of the three layers is: a) whitening of the input data for yielding good separation results; b) separation of independent sources (components); c) estimation of the basis vectors.

ü  Design, modification and implementation of nonlinear constrained Hebbian rules – one-unit case and multi-unit case, bigradient learning rule of Wang etc. for learning with different criterion functions. Setting the experimental parameters: image data, sampling, data preprocessing, the algorithm and its parameters, estimation of statistical quantities.

The implementation of ICA requires selecting suitable parameterizations and estimation of parameters. Testing all possible combination of parameters would require a huge amount of experiments. In order to restrict their number we test the influence of one or two parameters at a time. We also propose a graphical user interface (GUI), which greatly improves the time efficiency of the testing phase.