API for deterministic and stochastic surrogates.

Given data $((x_1, y_1), \ldots, (x_N, y_N))$ obtained by evaluating a function $y_i = f(x_i)$ or sampling from a conditional probability density $p_{Y|X}(Y = y_i|X = x_i)$, a deterministic surrogate is a function $s(x)$ (e.g. a radial basis function interpolator) that uses the data to approximate $f$ or some statistic of $p_{Y|X}$ (e.g. the mean), whereas a stochastic surrogate is a stochastic process (e.g. a Gaussian process approximation) that uses the data to approximate $f$ or $p_{Y|X}$ and quantify the uncertainty of the approximation.