The Content Optimizer applies a two-stage policy that combines Stage 1 diversity selection across rank factor classes with Stage 2 Bayesian acquisition to iteratively select the most promising content hypotheses.
Referenced by
