Alexis Boukouvalas, Dan Cornford, Milan Stehlik,
"Approximately Optimal Experimental Design for Heteroscedastic Gaussian Process Models"
, Serie Technical Report, Neural Computing Research Group, Birmingham B4 7ET, UK, 11-2009
Approximately Optimal Experimental Design for Heteroscedastic Gaussian Process Models
Sprache des Titels:
This paper presents a greedy Bayesian experimental design criterion for heteroscedastic Gaussian
process models. The criterion is based on the Fisher information and is optimal in the sense of
minimizing parameter uncertainty for likelihood based estimators. We demonstrate the validity
of the criterion under different noise regimes and present experimental results from a rabies
simulator to demonstrate the effectiveness of the resulting approximately optimal designs.