On the use of neural networks for Hamming coding

An examination is made of the possible solutions to the problem of Hamming coding with reference to two different classes of neural networks: CPN and BPN. In the case of BPNs the authors discuss a general methodological approach based on subdivision of the problem into sub-problems. The CPN is well suited to solve the error-correcting codes for the way in which the Kohonen slab processes its input. More interesting is the application of the BPN. The particular approach adopted divides the problem into two subproblems: classification and codification. This has the advantage of allowing the designer to utilize well-tested theories relative to two well-known problems.