Non-Greedy Adaptive Vector Quantizers

Kohonen's Learning Vector Quantization (LVQ) technique easily gets trapped in local minima of the distortion surface, resulting suboptimal vector quantizers. The reason is that the behavior of competitive learning on which the LVQ bases is greedy, that is, it only accepts new solutions which maximally reduce the distortion. In this paper, a new and non-greedy adaptive vector quantization scheme is developed which applied a simulated annealing — a randomized search algorithm to the learning procedure and has the capabilities of hill-climbing and approaching global optima. Therefore, this scheme has the advantage of global optimization over the Kohonen's LVQ scheme. The adaptation (learning) equations are derived and the design schedule procedure is presented.

[1]  Donald Geman,et al.  Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images , 1984, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[2]  Teuvo Kohonen,et al.  Improved versions of learning vector quantization , 1990, 1990 IJCNN International Joint Conference on Neural Networks.

[3]  Zhicheng Wang,et al.  Code-excited neural vector quantization , 1993, 1993 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[4]  J. V. Hanson,et al.  Design optimization of code-excited neural vector quantizers , 1993, Proceedings of 1993 International Conference on Neural Networks (IJCNN-93-Nagoya, Japan).

[5]  Robert M. Gray,et al.  An Algorithm for Vector Quantizer Design , 1980, IEEE Trans. Commun..

[6]  Huang,et al.  AN EFFICIENT GENERAL COOLING SCHEDULE FOR SIMULATED ANNEALING , 1986 .

[7]  Bruce Hajek,et al.  A tutorial survey of theory and applications of simulated annealing , 1985, 1985 24th IEEE Conference on Decision and Control.

[8]  C. D. Gelatt,et al.  Optimization by Simulated Annealing , 1983, Science.

[9]  Teuvo Kohonen,et al.  An introduction to neural computing , 1988, Neural Networks.

[10]  H. Szu.  Fast simulated annealing , 1987 .

[11]  J. Hanson,et al.  A neural vector quantizer , 1992, [Proceedings] 1992 IEEE International Symposium on Circuits and Systems.

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