This paper provides a new architecture of neural network, called Loop Architecture Neural Network (LANN), and its learning rules. One of its distinguished features from other network, such as Hopfield and bidirectional assiociative memories, is that it can perform the associative memory among multiple categories. Analysis and simulated results have proved that it is an effective network with excellent convergence.
Farhat N H, Psaltis P. Prata A, et al[J].Applied Optics.1985, 24(10):1469-1475[2]Hopfield J J.[3]Neurons with graded response have collective computational properities like those of two states neurons. Proceedings of the National Academy of Sciences, USA; 31, May,1984. 3088-3092.[4]Anderson J A. IEEE Trans. on Systems, Man, and Cybernetics, 1983, SMC-13(5): 799-815.[5]Kosko B. Applied Optics, 1981, 26(23): 4974-4979.