Gromov-Witten invariants of local \(\mathbb{P}^2\) and modular forms (Q2231672)

The paper under review studies the Gromov-Witten invariants of local \(\mathbb{P}^2\) which is of special importance in the sense that it can not be computed by strip technique of topological vertex formalism from algorithm viewpoint and it is a canonical example corresponding to non-Lagrangian physics. \textit{E. Witten} [``Quantum background independence in string theory'', Preprint, \url{arXiv:hep-th/9306122}] observed that topological string partition function behaves as a ``wave function'' of quantum-mechanical system. Based on Witten's work, \textit{M. Aganagic} et al. [Commun. Math. Phys. 277, No. 3, 771--819 (2008; Zbl 1165.81037)] conjectured that the Gromov-Witten potential of \(K_{\mathbb{P}^2}\cong \mathcal{O}_{\mathbb{P}^2}(-3)\) is a meromorphic quasi-modular form for the congruence subgroup \(\Gamma_1(3)< SL(2, \mathbb{Z})\). Towards this conjecture, many different groups including \textit{S. Guo} and \textit{J. Zhou} [Adv. Math. 268, 1--61 (2015; Zbl 1343.14045)], \textit{M. Alim} et al. [Adv. Theor. Math. Phys. 18, No. 2, 401--467 (2014; Zbl 1314.14081)], \textit{H. Lho} and \textit{R. Pandharipande} [Proc. Lond. Math. Soc. (3) 119, No. 3, 781--813 (2019; Zbl 1465.14053)], \textit{B. Fang} et al. [J. Am. Math. Soc. 33, No. 1, 135--222 (2020; Zbl 1444.14093)] made their own contributions. Different from the above endeavors, the goal of this paper is to understand the physics picture for local \(\mathbb{P}^2\) from mathematical Givental's formalism. The story consists of two parts: the first part is about the case of genus 0 Gromov-Witten invariants, the genus 0 Gromov-Witten potential \(F_0\) can define a Lagrangian submanifold \(\mathcal{L}\) in some cotangent bundle as the graph of \(dF_0\), the celebrated crepant resolution conjecture at genus 0 states that the graphs of \(dF_0^{K_{\mathbb{P}^2}}\) and \(dF_0^{[\mathbb{C}^3 / \mathbb{Z}_3]}\) coincide under some affine symplectic transformation. Furthermore, the family of tangent spaces \(T_t{\mathcal{L}}\) along \(\mathcal{L}\) defines a variation of Hodge structures (VHS) of weight 1. Under mirror symmetry, \(F_0^{K_{\mathbb{P}^2}}\) and \(F_0^{[\mathbb{C}^3 / \mathbb{Z}_3]}\) describe the local behaviors of the VHS near large radius limit and orbifold point respectively; the second part is about the case of higher genus Gromov-Witten invariants, inspired by the idea of quantization and Aganagic-Bouchard-Klemm's Feymann rule description, the authors introduce a Fock sheaf over moduli space of elliptic curves with \(\Gamma_1(3)\)-level structure associated with Gromov-Witten potential, then near appropriate limit point, the global section of this Fock sheaf coincides with the Gromov-Witten potential of local \(\mathbb P^2\) or the orbifold \([\mathbb{C}^3 / \mathbb{Z}_3]\). Under crepant resolution, the higher genus Gromov-Witten potentials \(F_g^{K_{\mathbb{P}^2}}\) and \(F_g^{[\mathbb{C}^3 / \mathbb{Z}_3]}\) are related by an explicit Feymann diagram expansion with the helps of mirror symmetry, K. Saito's structure and global sections of Fock sheaf, hence it proves Aganagic-Bouchard-Klemm's conjecture and crepant resolution conjecture for \([\mathbb{C}^3 / \mathbb{Z}_3]\) and its crepant resolution \(K_{\mathbb{P}^2}\). In conclusion, the paper under review presents a significant Fock sheaf understanding of Gromov-Witten potentials and use this interpretation to prove Aganagic-Bouchard-Klemm's quasi-modular form conjecture and all genus Crepant Resolution Conjecture (CRC) together with conifold gap conjecture in terms of Givental formalism, holomorphic anomaly equation and mirror symmetry. Compared with Lho-Pandharipande's stable quotient method, Coates-Iritani's work can be regarded as a different, impressive Givental-style's flavor.