Boolean neural network design using set covering in hamming geometrical space

Some novel learning strategies based on set covering in Hamming geometrical space are presented and proved, which are related to the three-layer Boole an neural network (BNN) for implementing an arbitrary Boolean function with low-complexity. Each hidden neuron memorizes a set of learning patterns, t hen the output layer combines these hidden neurons for explicit output as a Boolean function. The network structure is simple, reliable and can be eas ily implemented by hardware.