{"id":"arxiv/2610.08069","url":"https://arxiv.org/abs/2610.08069","tags":["cs.LG","math.ST","stat.ML"],"score":0,"title":"Detecting a Shift Is Not Enough: Exact Minimax Limits of Linear Representation Repair","domain":"cs.LG","source":"arxiv","authors":["Anuar Aimoldin","Yankai Chen","Ayana Mussabayeva","Xue Liu"],"reviews":0,"subject":"cs.LG","summary":"A mean shift between two data sources can be easy to detect but hard to remove without substantially changing their representations. We cast its removal as a statistical decision problem: from noisy differences between paired calibration measurements in $\\mathbb{R}^d$, learn one linear map, applied to both sources under a hard distortion budget, that leaves as little of the shift as possible on fresh data. We derive the exact finite-sample minimax risk over all such maps, $(d-k) \\mathbb{E}[1/(d+2J)]$ with $J\\sim\\mathrm{Pois}(κ/2)$, where the budget allows deleting $k$ directions and $κ$ is…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-10-06","reactions":{}}