Invariant Operations on Generalized Neurofunctions

摘要

The research studied the spectral properties of logic algebra functions that are implemented by one generalized neural element. There are specified criteria of logic algebra functions implementation by one generalized neural element relating to a given system of characters. The research introduces the concept of characteristic vector for Boolean function relating to a given system of characters and demonstrates that Boolean functions, implemented by one generalized neural element, are uniquely defined by their characteristic vectors. The particularity mentioned for eigenvectors can be effectively applied for compression of generalized neural functions in the spectrum range of Walsh-Hadamard basis functions system, including the description either of invariant operations or compact representation over these functions.