Title :
Principal interconnections in higher order Hebbian-type associative memories
Author_Institution :
Dept. of Electr. Eng., Nat. Taiwan Ocean Univ., Keelung
Abstract :
The existence of principal interconnections useful in solving the proliferation problem in higher order Hebbian-type associative memories is introduced. Among all legal interconnections, the author proves there exists a subset Tpr that carries more information than the others. Regardless of the network order p, the elements in Tpr are shown to be those interconnections T that fall within the range of √m, ⩽|T|⩽2√m, where ms equals the number of stored code words. Memories that use only Tpr can maintain original generalization performance, using less than 50 percent of the total number of interconnections
Keywords :
content-addressable storage; generalisation (artificial intelligence); neural nets; generalization performance; higher order Hebbian-type associative memories; information carrying; legal interconnections; network order; principal interconnections; proliferation problem; stored code words; Associative memory; Capacity planning; Convergence; Integrated circuit interconnections; Law; Legal factors; Neurons; Optical interconnections; Optical modulation; Optical recording;
Journal_Title :
Knowledge and Data Engineering, IEEE Transactions on
