Mean-square shifted convolution conjecture for non-holomorphic cusp-form coefficients

Let t(n)t(n) denote the Fourier coefficients of the non-holomorphic cusp form considered in the paper, and let mMm\sim M mean M<m2MM<m\leq 2M. Let N1N\geq 1, 1LN1\leq L\leq N, and 1FN1ε1\leq F\ll N^{1-\varepsilon}. Shifted convolution conjecture. One should have

fFnNlLt(n+l)t(n+l+f)2N2+ε+N1+εLF.\sum_{f\sim F}\sum_{n\sim N}\left|\sum_{l\sim L}t(n+l)t(n+l+f)\right|^{2}\ll N^{2+\varepsilon}+N^{1+\varepsilon}LF.

This is the conjectured analogue of the established mean-square estimate for the divisor function and the corresponding result for holomorphic cusp forms. It is proposed because the spectral decomposition for non-holomorphic cusp forms contains an additional arithmetic correction term that obstructs the proof; the source gives no resolution of the conjecture.

Sources & referencesView supporting material

Primary source

Eeva Suvitie, “On the shifted convolution problem in mean”, arXiv:1202.3906 (2012).

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.