The Cohn–Elkies exponent and sign uncertainty

1. Introduction🔗

This blueprint follows Chapter 1 of the report Ten proofs (OpenAI), which determines the exact exponential growth rate of the Cohn–Elkies sphere-packing linear program and the sharp asymptotics of the two Fourier sign-uncertainty constants. The chapters follow the report. This introduction fixes the setting and states the main results. The next chapter proves the Cohn–Elkies bound itself, which the report cites as an external input and which the Lean development proves from Poisson summation. The chapter on preliminaries collects the gamma-function identities, the radial reduction, the radial Mellin transform, and the subharmonic functions and half-plane Poisson inequality behind the report's proof of Lemma 3.2; the chapter on the lower bound proves the universal obstruction (Proposition 3.1) and the packing lower bound (Theorem 3.8); the chapter on the upper bound constructs the asymptotically optimal functions (Theorem 4.1). The appendix compares the two sign-uncertainty constants (Proposition A.1 and \mathsf{A}_+(d) < \mathsf{A}_-(d), through the existence of extremizers, Cohn–Gonçalves 2019, Theorem 1.4), and a final chapter records where the Lean formalization deviates from the report.

  1. 1.1. The sphere-packing problem and the Cohn–Elkies linear program
  2. 1.2. Fourier sign uncertainty
  3. 1.3. Strategy