Title :
Random attractors for initial memory codes
Abstract :
The problem of how the temporal relations among events are converted to the spatial structure in neural connections are considered. Random coding is one hypothesis to this problem, in which items to be stored are associated with randomly chosen network attractors, independent of spatial relationships between the items. We demonstrate the idea with a simple network model, and discuss the relevant issues
Keywords :
content-addressable storage; neural nets; neurophysiology; initial memory codes; neural connections; randomly chosen network attractors; spatial structure; temporal relations; Agricultural engineering; Agriculture; Brain modeling; Hebbian theory; Image recognition; Mathematics; Neural networks; Neurons; Retina; Shape;
Conference_Titel :
Neural Networks,1997., International Conference on
Conference_Location :
Houston, TX
Print_ISBN :
0-7803-4122-8
DOI :
10.1109/ICNN.1997.616207
