A new paper demonstrates that the Benjamini--Hochberg procedure can fail to control the false discovery rate (FDR) for correlated two-sided Gaussian tests. Researchers constructed a factor model where, at a 0.01 significance level, the FDR exceeded 0.0104 for a large number of hypotheses. This finding disproves a long-standing conjecture and was generated with the assistance of GPT-5.6 Pro, then meticulously verified by the author. AI
IMPACT Challenges assumptions in statistical testing, potentially impacting AI model evaluation and research reproducibility.
RANK_REASON Academic paper presenting novel theoretical findings that disprove a long-standing conjecture. [lever_c_demoted from research: ic=1 ai=0.7]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
