Structured Models with Gaussian Processes

Markus Kaiser

Abstract: We formulate Bayesian structured models using composite and hierarchical Gaussian process models that reproduce knowledge, are understandable for domain-experts and make physically plausible predictions. Using real-world industrial applications such as the detection of faulty sensors and the prediction of power generation in a wind-farm as examples, we show how to use structured models to factorize uncertainties, achieve interpretability, and generalize to unobserved inputs.