New Progress on Worst-Case FDR for BH Procedure
lihua_lei_stat · x · 2026-07-17
This post introduces new research on the Benjamini-Hochberg (BH) procedure, focusing on worst-case FDR control under dependent two-sided Gaussian tests.
Key points:
- Researchers advanced a conjecture about the BH procedure after about 10 hours of work.
- Previous results showed that FDR(BHq)/q could be greater than 1 in some cases.
- The original folk conjecture: Is there a uniform upper bound for this ratio?
- The post concludes: No universal upper bound exists.
This is a theoretical statistics result, focusing on understanding the limiting behavior of the BH procedure under correlated testing scenarios.
Related event: Lihua Lei sharpens worst-case FDR bounds in common-factor Gaussian models(5 posts)→
More from Research
- Hermes Agent rewrite proposal applies RIA and Logic Bus rules — Promptmethus · 2026-07-22
- AllTheBacteria turns 2.44 million genomes into an AI-ready resource for new antibiotics — shae_mcl · 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
- NVIDIA says physical AI starts in simulation with OpenUSD and synthetic data — MonaJalal_ · 2026-07-22