An error correcting algorithm for Hopfield network

The principle and the weakness of the Hopfield network are discussed. It is found that the assumption that the Hopfield network made on the noise effect of input patterns is inappropriate and an adaptive training algorithm that minimizes the noise effect of the input patterns is presented. This algorithm alters the connection weights of the network. It is shown that the storage capacity of the resultant model increases from 0.16n to greater than 1.14n, where n is the number of neurons in the network. Moreover, the model has a higher error tolerance level than the original model.<<ETX>>

[1]  Anders Krogh,et al.  Introduction to the theory of neural computation , 1994, The advanced book program.

[2]  J J Hopfield,et al.  Neural networks and physical systems with emergent collective computational abilities. , 1982, Proceedings of the National Academy of Sciences of the United States of America.

[3]  Teuvo Kohonen,et al.  Self-Organization and Associative Memory , 1988 .

[4]  David J. Burr,et al.  Experiments on neural net recognition of spoken and written text , 1988, IEEE Trans. Acoust. Speech Signal Process..