Gaussian process models

A Gaussian process model is a data interpolation model which can be used to infer the relationship between a set of input parameters and a set of target, or output, parameters. The following papers give a brief introduction to the models, and show their application in the context of material processing. They are ordered in order of decreasing order of information on Gaussian Processes: the first is the most recommended.

• Prediction of Deformed and Annealed Microstructures Using Bayesian Neural Networks and Gaussian Processes (1997)
• Gaussian Process Modelling of Austenite Formation in Steel (1999)
• Static and Dynamic Modelling of Materials Forging (1998)
• Accelerated learning using Gaussian process models to predict static recrystalisation in an Al-Mg alloy (2000)

• A mathematical description and discussion of the covariance function and training (PDF file)
• A link to David MacKay's Gaussian Process web page, including his review article
• Tpros, a software implementation of Gaussian Processes, by Mark Gibbs