R. Folk et al., SYMMETRIES IN NEURAL NETWORKS - A LINEAR GROUP ACTION APPROACH, Journal of physics. A, mathematical and general, 26(13), 1993, pp. 3159-3164
Neural networks storing Hadamard patterns have been completely classif
ied with respect to permutation symmetry. The symmetry group of the Ha
damard patterns is found to be isomorphic to GL(n, F2), and the symmet
ry groups of the networks are explicitly constructed for the most impo
rtant classes. The volumes of different equivalence classes have been
calculated.