Statistics Paper Reveals New Lower Bound Results
lihua_lei_stat · x · 2026-07-17
This post discusses findings from a statistics paper where the author notes their obtained lower bound is q sqrt{log(1/q)}, a surprisingly unexpected form.
They compare this to @weijie444's FDR linking theorem, which provides an upper bound on the order of log(1/q). However, the properties differ: the latter applies to more general non-Gaussian scenarios, whereas these results are specifically tied to Gaussian settings.
A reply adds that if q is allowed to approach 0, they have proven supn C(n) = infty and derived stronger rate conclusions. The paper has been submitted to arXiv and will be published the next day.
Related event: Lihua Lei sharpens worst-case FDR bounds in common-factor Gaussian models(5 posts)→
More from Research
- Nat Lambert shares a reading list on synthetic data and agentic SFT data — natolambert · 2026-07-22
- WeirdChat catalogs strange model behaviors from more than 100 million sampled responses — JacobSteinhardt · 2026-07-22
- New agentic benchmark shows AI managers escalate to coercion and fake success — Jasmine Brazilek · 2026-07-22
- Ai2’s Asta adds one-click handoff and self-checking deep paper search — allen_ai · 2026-07-22
- DepthART pushes monocular depth to tiny models at 1000 FPS on RTX A6000 — kwangmoo_yi · 2026-07-22
- Meta says SAM 3 and DINOv3 cut 3D volume labeling from a month to 15 minutes — AIatMeta · 2026-07-22