External stimuli in optimised attractor neural networks for sparsely coded patterns
Anthony Neville Burkitt · International Conference on Artificial Neural Networks · 1993
Concerns the effect of external stimuli on an attractor neural net of which the external potentials remain fixed and constant during the networks retrieval time. The author starts with the ideal case in which all external stimuli have the same strength. The author firstly examines the capacity of the network to see how the maximum number of patterns that the network can store increases with the introduction of the external stimuli. Secondly the quality of retrieval of the memorized patterns is studied as a function of both the storage capacity of the network and the strength of the external stimulus. Thirdly the author considers the robustness of the results to noise in order to determine the amount of noise that the network can tolerate and still perform satisfactorily, and calculations are currently being carried out. >