A Hierarchy of Hyperbolic Macrodynamic Equations as a Model for Network Training

Melnik, R.V.N.

Proceedings of the IEEE International Symposium on Information Theory, Ulm, Germany, 322, 1997

Abstract:

We propose mathematical models of hyperbolic type for training of neural networks. We analyze such models and develop their computational implementation by using the Markov chain approximation method.

Key words: complex systems; T-computable functions; network training as a problem in optimal control; perturbed Markov chains; singular stochastic control problems; generalized dynamic systems; discrete Markovian decision processes.

Publication is abstracted/indexed in INSPEC, Compendex, IEEE Xplore, and other major databases.