An information-based approach towards neuro-evolution
Behzad Behzadan, Robert E. Smith · 2012
A novel self-governing system, which is theoretically founded on information theory, is introduced with the ability of determining the optimal quantity and connectivity of the hidden-layer of a three layer feed-forward neural network. The system - called MINES - simultaneously links parameter learning (performed by back-propagation) to structural learning (performed by genetic algorithm) with the aid of mutual information between the error-space and the hidden-layer.