An Information Maximization Approach to Overcomplete and Recurrent Representations
Shriki, Oren, Sompolinsky, Haim, Lee, Daniel D.
–Neural Information Processing Systems
The principle of maximizing mutual information is applied to learning overcomplete and recurrent representations. The underlying model consists of a network of input units driving a larger number of output units with recurrent interactions. In the limit of zero noise, the network is deterministic and the mutual information can be related to the entropy of the output units.
Neural Information Processing Systems
Dec-31-2001