paper

Linear and quadratic uniformity of the Möbius function over

arXiv:1711.05358

Abstract

We examine correlations of the Möbius function over with linear or quadratic phases, that is, averages of the form \begin{equation} \label{eq:average} \frac{1}{q^n}\sum_{\text{deg }f<n} μ(f)χ(Q(f)) \end{equation} for an additive character over and a polynomial of degree at most 2 in the coefficients of . Like in the integers, it is reasonable to expect that, due to the random-like behaviour of , such sums should exhibit considerable cancellation. In this paper we show that the above correlation is bounded by for any if is linear and for some absolute constant if is quadratic. The latter bound may be reduced to ) for some when is a linear form in the coefficients of , that is, a Hankel quadratic form, whereas for general quadratic forms, it relies on a bilinear version of the additive-combinatorial Bogolyubov theorem.

24 pages, 1 figure. Second version optimises the saving in the linear case. Third version uses the proof by Hosseini and Lovett of the author's conjecture to make the main result unconditional

Cited by in corpus (1)