Title :
The asymmetric Hopfield model for associative memory
Author_Institution :
Dept. of Math., Shantou Univ., Guangdong, China
Abstract :
The theoretical research on associative memory of the asymmetric Hopfield model is presented. The perceptron learning scheme is proposed to store the sample vectors (patterns) in the neural network. For this generalized Hopfield model of n neurons, an upper and a lower bounds for asymptotic memory capacity are obtained respectively to be 2n and (n-1).
Keywords :
Hopfield neural nets; associative processing; content-addressable storage; learning (artificial intelligence); perceptrons; associative memory; asymmetric Hopfield model; asymptotic limit theorem; lower bound; neural network; perceptron learning; sample vectors; upper bound; Associative memory; Capacity planning; Computer networks; Discrete time systems; H infinity control; Hopfield neural networks; Mathematical model; Neural networks; Neurons; Performance evaluation;
Conference_Titel :
Neural Networks, 1993. IJCNN '93-Nagoya. Proceedings of 1993 International Joint Conference on
Print_ISBN :
0-7803-1421-2
DOI :
10.1109/IJCNN.1993.714259
