Diversity weighting adjusts the sampling probability of each Rank Factor Class by raising its inverse test count to an exponential power \(d\), shifting exploration toward under-tested mechanism classes before running the intra-class Bayesian acquisition policy.
