ABSTRACT
A surrogate is a substitute for the real thing. In statistics, draws from predictive equations derived from a fitted model can act as a surrogate for the data-generating mechanism. Surrogate modeling is statistical modeling of computer experiments. Computer simulations are generally cheaper than physical observation, so the former could be entertained as an alternative or precursor to the latter. Facebook uses surrogates to tailor its web portal and apps to optimize engagement; Uber uses surrogates trained to traffic simulations to route pooled ride-shares in real-time, reducing travel and wait time. Advances in high performance computing have facilitated distribution of solvers to an unprecedented degree, allowing thousands of runs where only tens could be done before. That’s helpful with big input spaces, but shifts the burden to big models and big training data which bring their own computational challenges.
