Robustness of feedforward neural networks
Ching-tai Chiu, K. Mehrotra, Chilukuri Krishna Mohan, Sanjay Ranka · 2002
Methods are developed for measuring the sensitivity of links and nodes of a feedforward neural network, and for implementing a technique to ensure the development of neural networks that satisfy well-defined robustness criteria. Experimental observations indicate that performance degradation in the authors' robust feedforward network is significantly less than a randomly trained feedforward network of the same size by an order of magnitude.>