VibeMathedMath problems solved by AI

Universal Multiplicative FDR Bound for Benjamini-Hochberg

The Benjamini-Hochberg procedure is known not to control the false discovery rate at its nominal level under arbitrary dependence. A folklore conjecture in the FDR literature held that it must at least control the FDR up to a universal multiplicative constant. It does not: there are finite Gaussian models whose FDR divided by qq diverges as q0q \downarrow 0, with an explicit two-sided lower bound qlog(1/q)/(2π)+0.6493q+o(q)q\sqrt{\log(1/q)}/(2\sqrt{\pi}) + 0.6493 q + o(q).

Result
Disproved
Status
Resolved
AI contribution
AI-assisted
Method
Construction
Field
Statistics
Posed by
folklore in the FDR literature
Year posed
Years open
Solved
2026-07-20
Model
GPT-5.6 Sol
Vendor
OpenAI
Collaborators
Lihua Lei
Verification
Unreviewed
Publication
Preprint
Significance
30 / 100
Disclosed cost
Wikipedia
No dedicated article

What the AI did

The disclosure says only that the author used the model for assistance with proof exploration, exposition and editing. It does not attribute any specific step, so the lowest tier applies.

Verification

Single-author arXiv preprint; not yet peer-reviewed.

Source

arXiv:2607.14812 - How Much Can Gaussian Dependence Inflate the Benjamini-Hochberg Procedure's FDR?

Discussion