Algebraic training of a neural network

A novel algebraic neural network training technique is developed and demonstrated on two well-known architectures. This approach suggests an innovative, unified framework for analyzing neural approximation properties and for training neural networks in a much simplified way. Various implementations show that this approach presents numerous practical advantages; it provides a trouble-free non-iterative systematic procedure to integrate neural networks in control architectures, and it affords deep insight into neural nonlinear control system design.

Duke Authors

Cited Authors

  • Ferrari, S; Stengel, RF

Published Date

  • 2001

Published In

  • Proceedings of the American Control Conference

Volume / Issue

  • 2 /

Start / End Page

  • 1605 - 1610