Methods for encoding in multilayer feed-forward neural networks

Abstract

Neural network techniques for encoding-decoding processes have been developed. The net we have devised can work like a memory retrieval system in the sense of Hopfield, Feinstein and Palmer. Its behaviour for 2^R (R ∈ N) input units has some special interesting features. In particular, the accessibilities for each initial symbol may be explicitly computed. Although thermal noise may muddle the code, we show how it can statistically rid the result of unwanted sequences while maintaining the network accuracy within a given bound.