{"id":"arxiv/2606.11263","url":"https://arxiv.org/abs/2606.11263","tags":["math.ST","cs.LG","math.NA","math.PR"],"score":0,"title":"Geometric bias in eigenspace perturbation under random heterogeneous noise","domain":"math.ST","source":"arxiv","authors":["Fengkai Liu","Ke Wang","Wanjie Wang"],"reviews":0,"subject":"math.ST","summary":"Spectral methods rely on the stability of principal eigenspaces under random perturbations. Classically, this is quantified by the Davis-Kahan and Wedin theorems, which bound the eigenspace error via the operator norm of the noise and the relevant spectral gaps. While sharp for arbitrary deterministic perturbations, these worst-case bounds can be wasteful in the low-rank signal-plus-noise setting, as they fail to capture the interaction between the signal geometry and the noise distribution. We study the spectral perturbation of signal-plus-noise matrices corrupted by sparse random noise…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-06-09","reactions":{}}