Sanger’s Rule and BCM

For my final project in Computational Neuroscience, my team chose to investigate image feature extraction using neural networks using models such as the Generalized Hebbian Algorithm and the BCM rule.

This video shows the convergence of synaptic weights to the first four eigenvectors, or principle components. Also known as Sanger’s rule, the Generalized Hebbian Algorithm offers a much faster way to calculate principle components and is supported by biology.

You can read our paper here: Iterative face image feature extraction with Generalized Hebbian Algorithm and a Sanger-like BCM rule

Our code is here also: NEUR 1680 Final Code

We used the FG-NET aging database to test our neural network. Team members were Carl OIsson, Clayton Aldern, and Tyler Benster.

Leave a Reply

Your email address will not be published. Required fields are marked *