A new approach to the design of Hopfield associative memory

The authors present a novel method for constructing the weight matrix for the Hopfield associative memory. The most important feature of this method is the explicit introduction of the size of the attraction basin to be a main design parameter, and the weight matrix is obtained as a result of optimizing this parameter. Another feature is that all the connection weights can only assume three different values, -1, +1, and 0, which facilitates the VLSI implementation of the weights. Compared to the widely used Hebbian rule, the method can guarantee all the given patterns to be stored at least as fixed points, regardless of the internal structure of the patterns. The proposed design method is illustrated by a few examples.<<ETX>>

[1]  Joos Vandewalle,et al.  Pattern Storage and Hopfield Neural Associative Memory with Hidden Structure , 1992, Int. J. Neural Syst..

[2]  Demetri Psaltis,et al.  Linear and logarithmic capacities in associative neural networks , 1989, IEEE Trans. Inf. Theory.

[3]  Jehoshua Bruck,et al.  A study on neural networks , 1988, Int. J. Intell. Syst..

[4]  Yaser S. Abu-Mostafa,et al.  Information capacity of the Hopfield model , 1985, IEEE Trans. Inf. Theory.

[5]  Amir Dembo,et al.  On the capacity of associative memories with linear threshold functions , 1989, IEEE Trans. Inf. Theory.

[6]  Joos Vandewalle,et al.  Determination of weights for Hopfield associative memory by error back propagation , 1991, 1991., IEEE International Sympoisum on Circuits and Systems.

[7]  J S Denker,et al.  Neural network models of learning and adaptation , 1986 .

[8]  Santosh S. Venkatesh,et al.  The capacity of the Hopfield associative memory , 1987, IEEE Trans. Inf. Theory.

[9]  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.

[10]  Robert Fischl,et al.  Design of the fully connected binary neural network via linear programming , 1990, IEEE International Symposium on Circuits and Systems.