{"id":"arxiv/2603.05335","url":"https://arxiv.org/abs/2603.05335","tags":["stat.ML","cs.LG","math.ST"],"score":0,"title":"Bayes with No Shame: Admissibility Geometries of Predictive Inference","domain":"stat.ML","source":"arxiv","authors":["Nicholas G. Polson","Daniel Zantedeschi"],"reviews":0,"subject":"stat.ML","summary":"Modern predictive systems combine predictors, sequential monitors, prediction sets, and online strategies, each with a different certificate of optimality. We study four criterion-relative geometries: Blackwell risk dominance, anytime-valid admissibility, fixed-level marginal coverage with expected-length efficiency within a declared rank-indexed family, and choice-based approachability (CApp) boundary-feasibility. We embed the four procedure types in a common product space and prove witness-based pairwise non-nesting: for every ordered pair of criterion classes, an explicit predictive…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-03-05","reactions":{}}