Dictators are most informative

Vu Khac Ky, Tuan Tran

Ky and Tran give a separate proof of the Courtade–Kumar inequality for every Boolean function under independent binary symmetric noise.

What was already known

The original conjecture was known in several restricted regimes. This proof uses an entropy-production argument with Fourier and spectral bounds.

AI involvement AI-centered

Ky and Tran formulated the approach and credit ChatGPT with developing the proofs, preparing the verification code and helping revise the exposition.

Importance Major

The proof establishes the full inequality by combining entropy production with Fourier and spectral estimates.

Technical verification Partial

Appendix A describes exact rational and outward-rounded interval checks for the finite coefficient and parameter estimates used in the argument.

Formal checking

Checked the finite inequalities and coefficient estimates listed in Appendix A.

Checker
The authors’ rational and Arb verification routines
Scope
Ky and Tran give a separate proof of the Courtade–Kumar inequality for every Boolean function under independent binary symmetric noise.
Basis
Reported by the linked checker

Human understanding Few

The authors devised the approach. Their disclosure does not separately document a human reconstruction of the AI-assisted technical proof.

Other useful sources

Correction and publication history