Bayesian Learning for Neural Networks

1. Front Matter

   Pages i-xiv

   PDF

2. Introduction

   • Radford M. Neal

   Pages 1-28

3. Priors for Infinite Networks

   • Radford M. Neal

   Pages 29-53

4. Monte Carlo Implementation

   • Radford M. Neal

   Pages 55-98

5. Evaluation of Neural Network Models

   • Radford M. Neal

   Pages 99-143

6. Conclusions and Further Work

   • Radford M. Neal

   Pages 145-152

7. Back Matter

   Pages 153-185

   PDF

Artificial "neural networks" are widely used as flexible models for classification and regression applications, but questions remain about how the power of these models can be safely exploited when training data is limited. This book demonstrates how Bayesian methods allow complex neural network models to be used without fear of the "overfitting" that can occur with traditional training methods. Insight into the nature of these complex Bayesian models is provided by a theoretical investigation of the priors over functions that underlie them. A practical implementation of Bayesian neural network learning using Markov chain Monte Carlo methods is also described, and software for it is freely available over the Internet. Presupposing only basic knowledge of probability and statistics, this book should be of interest to researchers in statistics, engineering, and artificial intelligence.

• Fitting
• Likelihood
• algorithms
• artificial intelligence
• classification
• intelligence
• learning
• statistics

• Department of Statistics and Department of Computer Science, University of Toronto, Toronto, Canada

  Radford M. Neal

Policies and ethics