No-prior Bayes reIMagined: probabilistic approximations of inferential models

Ryan Martin

When prior information is lacking, the go-to strategy for probabilistic inference is to combine a "default prior" and the likelihood via Bayes's theorem. Objective Bayes, (generali…