Quasimap wall-crossings and mirror symmetry (Q2186092)

To fully appreciate the results of the present paper, let us first consider the case of the quintic threefold. Givental genus 0 enumerative mirror symmetry for the quintic states that, after an identification of the parameters by the mirror map, the \(I\)-function equals the \(J\)-function. The (genus 0) \(J\)-function is a generating function of genus 0 Gromov-Witten invariants of the quintic, and the \(I\)-function is built from periods of the mirror map. In 1993, Bershadsky-Cecotti-Ooguri-Vafa (BCOV) made a physics proposal of how to extend this to higher genus invariants. Namely (see Equation \((1.5.1)\) in the present article), they proposed that \[ I_0(q)^{2g-2} \mathcal{F}^B_g(q)= \sum_{d\geq0} Q^d \langle \; \rangle^{\infty}_{g,0,d}, \] where \(\langle \; \rangle^{\infty}_{g,0,d}\) is the Gromov--Witten invariant of curves of genus \(g\) and degree \(d\), \(Q=Q(d)\) is the mirror map, \(I_0(q)\) is the zeroth component of the \(I\)-function and \(\mathcal{F}^B_g(q)\) is the holomorphic limit of the genus \(g\) partition function of the B-model associated to the mirror family of the quintic. While there had been some progress in other (simpler) cases, a mathematically rigorous definition of \(\mathcal{F}^B_g(q)\) for the quintic was missing. An important consequence of the main result of the present paper is that \(\mathcal{F}^B_g(q)\) is a generating function of quasimap invariants for the stability condition \(\epsilon=0+\), where the stability parameter \(\epsilon\) is an element of \((0,\infty)\) and there are walls at \(\{1/n \, : \, n\in \mathbb{Z}\}\). Going from \(\epsilon=0+\) and crossing the walls (finitely many for each curve class) leads to the Gromov-Witten invariants \(\langle \; \rangle^{\infty}_{g,0,d}\) associated to the chamber \(\epsilon\in(1,\infty)\). This thus yields the first rigorous definition of \(\mathcal{F}^B_g(q)\). Moreover, it is defined for any Calabi-Yau variety \(X\) that is realized as a smooth projective GIT quotient and the analogous result to the above formula holds if \(X\) is a complete intersection in projective space. The main conceptual advance of the present paper is the conjectured wall-crossing formula (Conjecture 1.1 in the article) for virtual fundamental classes of quasimap invariants associated to GIT quotients. The main result of the present paper then consists of a proof of this virtual class wall-crossing formula in the case of complete intersections in projective spaces. For complete intersection Fano varieties of index at least 2, the authors show that the wall-crossing is trivial, i.e.~the invariants are independent of \(\epsilon\). For Fano index 1, the wall-crossing is simple and explicit. For complete intersection Calabi-Yau varieties \(X\), the author's work firmly establishes quasimap invariants as the \(B\)-model invariants of \(X\) and enumerative mirror symmetry as a consequence of wall-crossing between Gromov-Witten and quasimap invariants. Future directions likely will consist in mapping the physics BCOV theory to the quasimap theory. For complete intersection projective varieties of general type, the wall-crossing is new and might lead to interesting future directions.