🚨New conjecture dropped💥
Amid the excitement over using AI to solve conjectures, I’m going to announce a new one on
FDR(BHq) for correlated Gaussian tests
Formal statement is in the attached screenshot from the latest version (v4) of a paper I’ve been working on since 7/15:
The paper has taken ~180 hours of human work and ~100 hours of AI work by GPT-5.6 Sol. It was motivated by @EdgarDobriban’s amazing counterexample to a nearly 30-year-old conjecture in the @stats literature, and @weijie444’s elegant FDR-linking theorem. I’ve learned a ton from this interesting human-AI collaboration!
Show more
Episode3️⃣ of the FDR(BHq) saga: we now have promising lower bounds for ONE-sided Gaussian tests:
The wild part: GPT-5.6 Sol got the correct answer in ONE SHOT 🥳
I attached my exact prompt. It took me 1.5h to verify and another 6h to edit.
🧵(1/4)
Show more
New paper🚨After 10h w/ GPT-5.6 Sol, I made exciting progress on worst-case FDR of BH for correlated two-sided Gaussian tests
@edgardobriban showed FDR(BH_q)/q can exceed 1
Q (a folklore conjecture): Is the ratio universally bounded?
A: No👇
🧵(1/8)
Show more
