The Fourier uniformity conjecture for the Liouville function

Let λ:N{+1,1}\lambda:\mathbb{N}\to\{+1,-1\} be the Liouville function, defined by λ(n)=(1)Ω(n)\lambda(n)=(-1)^{\Omega(n)}, where Ω(n)\Omega(n) counts prime factors with multiplicity. Let H=H(X)H=H(X) be any function tending to infinity with XX, and let e(α)=exp(2πiα)e(\alpha)=\exp(2\pi i\alpha). Fourier uniformity conjecture. One has

Xx<2XsupαR/Zxn<x+Hλ(n)e(nα)=o(HX)\sum_{X\leqslant x<2X}\sup_{\alpha\in\mathbb{R}/\mathbb{Z}}\left\lvert\sum_{x\leqslant n<x+H}\lambda(n)e(n\alpha)\right\rvert=o(HX)

as XX\to\infty. The conjecture asserts negligible correlation between λ\lambda and linear phases in almost all short intervals, with the frequency allowed to depend arbitrarily on the interval. It was previously known for Hexp((logX)1/2+ε)H\geqslant\exp((\log X)^{1/2+\varepsilon}), and this paper proves it unconditionally for Hexp((logX)2/5+ε)H\geqslant\exp((\log X)^{2/5+\varepsilon}); the conjecture remains open below the stated range.

Sources & referencesView supporting material

Primary source

Cédric Pilatte, “Improved bounds for the Fourier uniformity conjecture”, arXiv:2604.26564 (2026).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.