English

Integral presentations of the shifted convolution problem and subconvexity estimates for $\operatorname{GL}_n$-automorphic $L$-functions

Number Theory 2023-11-14 v3 Representation Theory

Abstract

Fix n2n \geq 2 an integer, and FF be a totally real number field. We reduce the shifted convolution problem for LL-function coefficients of GLn(AF)\operatorname{GL}_n({\bf{A}}_F)-automorphic forms to the better-understood setting of GL2(AF)\operatorname{GL}_2({\bf{A}}_F). The key idea behind this reduction is to use the classical projection operator P1nφ\mathbb P^n_1 \varphi together with properties of its Fourier-Whittaker expansion. This allows us to derive novel integral presentations for the shifted convolution problem as Fourier-Whittaker coefficients of certain L2L^2-automorphic forms on the mirabolic subgroup P2(AF)P_2({\bf{A}}_F) of GL2(AF)\operatorname{GL}_2({\bf{A}}_F) or its two-fold metaplectic cover P2(AF)\overline{P}_2({\bf{A}}_F). We then construct liftings of these mirabolic forms to GL2(AF)\operatorname{GL}_2({\bf{A}}_F) and its two-fold metaplectic cover G(AF)\overline{G}({\bf{A}}_F) to justify expanding the underlying forms into linear combinations of Poincar\'e series. Decomposing each of the Poincar\'e series spectrally then allows us to derive completely new bounds for the shifted convolution problem in dimensions n3n \geq 3. As an application, we derive a uniform subconvexity bound for GLn(AF)\operatorname{GL}_n({\bf{A}}_F)-automorphic LL-functions twisted by Hecke characters. This uniform level-aspect subconvexity estimate appears to the the first of its kind for dimensions n3n \geq 3.

Keywords

Cite

@article{arxiv.1903.07284,
  title  = {Integral presentations of the shifted convolution problem and subconvexity estimates for $\operatorname{GL}_n$-automorphic $L$-functions},
  author = {Jeanine Van Order},
  journal= {arXiv preprint arXiv:1903.07284},
  year   = {2023}
}

Comments

This paper is withdrawn, at least temporarily, due to a gap in deriving bounds from the L^2-decomposition of the non-\Z-finite lifted mirabolic forms \Phi for the shifted convolution problem in ranks n \geq 3. While the setup leading to integral presentations (+ applications) is correct, the derivation of bounds via decompositions starting in {\S}4.3 is not. We intend to post a revised version later