In computer experiments, statistical models are commonly used as surrogates for slow-running codes. In this talk, the usually ubiquitous Gaussian process models are nowhere to be seen, however. Instead, an adaptive nonparametric regression model (BART) is used to deal with nonstationarities in the response surface. By providing both point estimates and uncertainty bounds for prediction, BART provdes a basis for sequential design criteria to ?nd optima with few function evaluations. Similar ideas will also be illustrated in other active learning problems, such as identification of active compounds in drug discovery.
Group for Research in Decision Analysis