Twisted pair correlation conjecture

Let n=qan=q^a be a prime power, define γ~=(logT)/(2π)γ\widetilde\gamma=(\log T)/(2\pi)\,\gamma, and let C(R)\mathfrak C(\mathbb R) be the set of continuous L1(R)L^1(\mathbb R) functions rr whose Fourier transform r^\hat r is integrable and Lipschitz continuous with r^(a)a3\hat r'(a)\ll|a|^{-3}. Let En=(a1)logq/logT+1/logT\mathcal E_n=(a-1)\log q/\log T+1/\log T, and define mnm_n by

mn(α)={1,α<1Λ(n)/logT,(logT/Λ(n))(α1),1Λ(n)/logTα<1,0,1α<1logn/logT,(logT/Λ(n))(α1+logn/logT),1logn/logTα<1(lognΛ(n))/logT,1,α1(lognΛ(n))/logT.m_n(\alpha)=\begin{cases}1,&\alpha<-1-\Lambda(n)/\log T,\\(\log T/\Lambda(n))(-\alpha-1),&-1-\Lambda(n)/\log T\leq\alpha<-1,\\0,&-1\leq\alpha<1-\log n/\log T,\\(\log T/\Lambda(n))(\alpha-1+\log n/\log T),&1-\log n/\log T\leq\alpha<1-(\log n-\Lambda(n))/\log T,\\1,&\alpha\geq1-(\log n-\Lambda(n))/\log T. \end{cases}

Twisted pair correlation conjecture. Fix ε>0\varepsilon>0 and let rC(R)r\in\mathfrak C(\mathbb R). Uniformly for all prime powers n=qaT1εn=q^a\leq T^{1-\varepsilon},

(T2πΛ(n)n)1Tγ,γ2Tniγr(γ~γ~)=Rr^(α)+r^(αlogn/logT)2(δ(α)+mn(α))dα+O(En).-\left(\frac{T}{2\pi}\frac{\Lambda(n)}{\sqrt n}\right)^{-1}\sum_{T\leq\gamma,\gamma'\leq2T}n^{i\gamma}r(\widetilde\gamma-\widetilde\gamma')=\int_{\mathbb R}\frac{\hat r(\alpha)+\hat r(-\alpha-\log n/\log T)}2\bigl(\delta(\alpha)+m_n(\alpha)\bigr)\,d\alpha+O(\mathcal E_n).

This weaker, integrated formulation follows from the strong twisted pair correlation conjecture by convolution with a suitable kernel. It is introduced because it is the form needed for the later argument; the source supports it conditionally from the strong conjecture but does not establish it in full.

Sources & referencesView supporting material

Primary source

Alessandro Fazzari and Maxim Gerspach, “The third moment of the logarithm of zeta and a twisted pair correlation conjecture”, arXiv:2412.20099 (2024).

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.