High-dimensional mean testing under truncation has an information-theoretic detectability floor from moment-based bias O(ν_{P,p} ε^{1-1/p}), with near-optimal second-order tests above it, and an escape to linear bias O(ε) under median regularity that recovers classical √d sample complexity.
Private statistical estimation via truncation
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Mean Testing under Truncation beyond Gaussian
High-dimensional mean testing under truncation has an information-theoretic detectability floor from moment-based bias O(ν_{P,p} ε^{1-1/p}), with near-optimal second-order tests above it, and an escape to linear bias O(ε) under median regularity that recovers classical √d sample complexity.