Bounds optimization of model response moments: a twin-engine Bayesian active learning method · Study