{"id":"arxiv/2610.12358","url":"https://arxiv.org/abs/2610.12358","tags":["math.PR","cs.IT","cs.LG","math.CO"],"score":0,"title":"Subspace Uncertainty and Sharp Sampling Thresholds on the Boolean Cube","domain":"math.PR","source":"arxiv","authors":["Thomas Weinberger"],"reviews":0,"subject":"math.PR","summary":"We study Gaussian regression under squared population $L_2$ loss in a known $m$-dimensional subspace of degree-at-most-$k$ functions on the $d$-dimensional Boolean cube. Random inputs can undersample regions essential for prediction, delaying the parametric rate even when the model is known. For fixed $q_0<1/2$, $1\\le k\\le q_0d$, and sufficiently large fixed $A$, the worst-subspace sample threshold for minimax error $Aσ^2(m+t)/n$ with confidence $1-e^{-t}$, $t\\ge\\log4$, is \\[ N=(m+t)\\exp\\{E_{d,k}+O(k^{1/3})\\}, \\quad E_{d,k}=dΨ(k/d), \\] where $Ψ(q)=\\log2-\\mathsf H(\\tfrac12-\\sqrt{q(1-q)})$ and…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-08","reactions":{}}