{"id":"arxiv/2610.32546","url":"https://arxiv.org/abs/2610.32546","tags":["stat.ML","cs.LG","math.ST"],"score":0,"title":"Exact Global Minimax Risk near Nonidentification","domain":"stat.ML","source":"arxiv","authors":["Kaiyue Yang","Yunan Wu"],"reviews":0,"subject":"stat.ML","summary":"We study squared-error estimation when a nuisance parameter is unidentified on a submodel but affects a target that vanishes there. For models with higher-order profile expansions, we give sufficient conditions under which the largest critical-path minimax risk is the exact asymptotic minimax risk over the full parameter space. An efficient observable feature, negligible calibration error and stable recovery yield one estimator attaining the leading constant uniformly, with the profile coordinates unknown. The estimator uses calibration estimates, a local feature mean and an empirical…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-10-23","reactions":{}}