Landau-Ginzburg/Calabi-Yau correspondence for the complete intersections \(X_{3,3}\) and \(X_{2,2,2,2}\) (Q502632)

The Landau-Ginzburg/Calabi-Yau (LG/CY) correspondence was proposed and understood in physical terms to correspond to the study of two (or more) different ``phases'' of certain two dimensional gauge theories [\textit{E. Witten}, Nucl. Phys., B 403, No. 1--2, 159--222 (1993; Zbl 0910.14020)]. Mathematically, the context for the LG/CY correspondence -- at least for CYs realized as complete intersections in weighted projective spaces -- is understood to be, respectively, FJRW theory [\textit{H. Fan} et al., Commun. Pure Appl. Math. 61, No. 6, 745--788 (2008; Zbl 1141.58012)] and genus zero Gromov-Witten (GW) theory. This paper generalizes the notion of FJRW theory for LG models corresponding to a collection of singularities (i.e. the threefolds are complete intersections in projective space), focusing on the only two non-quintic examples, \(X_{3, 3} \subset \mathbb{P}^5\) and \(X_{2, 2, 2, 2} \subset \mathbb{P}^7\). The main result of the paper is to establish the genus zero LG/CY correspondence for these two examples. This is accomplished by constructing the I-function of the hybrid model, whose coefficients expanded in powers of \(H^{(i)}\) span the solution space of a Picard-Fuchs equation associated to the corresponding Calabi-Yau (after a simple identification). The author establishes that the hybrid I-function and the usual genus zero hybrid J-function are related by an explicit change of variables (the mirror map) given for these two examples in Theorem 1.1.1. The I- and J-functions of the CY side of the correspondence are similarly related to one another in [\textit{A. Givental}, Prog. Math. 160, 141--175 (1998; Zbl 0936.14031)]. The latter is the generating function of genus zero GW invariants for the CY, explicitly \(J_{GW}(\mathbf{t}, z) = z + \mathbf{t} + \sum_{n, \beta}{1 \over n!} \langle \mathbf{t}(\psi), \ldots, \mathbf{t}(\psi), {\varphi_a \over z - \psi} \rangle^{GW}_{0, n+1, \beta} \varphi^a\) with \(\mathbf{t}(z) = t_0 + t_1 z + \ldots \in H_{GW}[z]\) and \(\varphi_a\) running over a basis for \(H_{GW}\). The small J-function is the restriction of the J-function to the degree-two component; for the genus zero correspondence to hold, the small \(J_{\mathrm{hyb}}\) and \(J_{GW}\) functions must coincide. The author explicitly verifies a degree-preserving isomorphism between the state spaces, \(H_{\mathrm{hyb}}\) and \(H_{GW} = H^*(X)\) for these examples. From this isomorphism between the state spaces in which the respective I-functions take values, and the content of Theorem 1.1.1, it then follows that \(I_{\mathrm{hyb}}\) and the analytic continuation of \(I_{GW}\) in a certain coordinate patch are given in terms of a basis of solutions to the same differential equation and are related by a linear isomorphism that performs the change of basis. This establishes the genus zero LG/CY correspondence for these examples (Corollary 1.1.2).