Title :
Implementation of H∞-learning and its analysis
Author :
Nishiyama, Kiyoshi
Author_Institution :
Dept. of Comp. & Inf. Sci., Iwate Univ., Japan
Abstract :
This paper studies implementation of the H∞-learning and unified approach to analyze the backpropagation and H2-learning as well as the H∞-learning. Various forms of H∞-learning algorithms are developed from the tradeoff between the learning performance and computational complexity. Also, an unified update formula of weight vector is derived.
Keywords :
backpropagation; computational complexity; feedforward neural nets; state-space methods; backpropagation; computational complexity; learning algorithm; multilayered feedforward network; multilayered neural networks; state-space model; supervised learning; Algorithm design and analysis; Artificial neural networks; Backpropagation algorithms; Computational complexity; Multi-layer neural network; Neural networks; Neurons; Robustness; Supervised learning; Uncertainty;
Conference_Titel :
Neural Information Processing, 2002. ICONIP '02. Proceedings of the 9th International Conference on
Print_ISBN :
981-04-7524-1
DOI :
10.1109/ICONIP.2002.1202197
