(Not) Bounding the True Error

We present a new approach to bounding the true error rate of a continuous valued classifier based upon PAC-Bayes bounds. The method first constructs a distribution over classifiers by determining how sensitive each parameter in the model is to noise. The true error rate of the stochastic classifier found with the sensitivity analysis can then be tightly bounded using a PAC-Bayes bound. In this paper we demonstrate the method on artificial neural networks with results of a

• Type: Conference
• Year: 2001
• Where: NIPS
• Authors: John Langford, Rich Caruana
• Keywords: Neural Net Bounds, NIPS 2001, NIPS 2007, Pac-bayes Bound, True Error Rate