Mirror symmetry for exceptional unimodular singularities (Q522513)

Broadly speaking, mirror symmetry is an equivalence, suitably stated, between \(A\)-model data associated to a symplectic geometric object and \(B\)-model data associated to the mirror complex geometric object. In its original form, it was concerned with an equivalence between the symplectic geometry of a Calabi-Yau (CY) manifold, in particular its associated Gromov-Witten invariants, and the complex geometry of the mirror Calabi-Yau manifold, in particular the variation of Hodge structures thereof. Since this early formulation, mirror symmetry has evolved to include other settings. The paper under consideration is concerned with Landau-Ginzburg (LG) models of a special type. Recall that a LG model is simply an affine manifold with a choice of function (though we will also include a group of symmetries of the function). Note also that there is a LG/CY correspondence, more on that below. Start with an invertible weighted homogeneous polynomial \(W\) on \(\mathbb{C}^n\) with isolated critical point at the origin and a finite abelian symmetry group \(G\) of \(W\). In this paper, \(G\) is assumed to be the group of diagonal symmetries of \(W\). \(G\) plays a role for the definition of the state space \(H_W\) and enters the LG/CY correspondence. The Berglund-Hübsch-Krawitz mirror to the Landau-Ginzburg model \((W,G)\) is another pair \((W^T,G^T)\). With the assumption on \(G\), \(G^T\) is trivial and \[ W^T= \sum_{i=1}^n \prod_{j=1}^n x_j^{a_{ji}} \; \text{ if } \; W= \sum_{i=1}^n \prod_{j=1}^n x_j^{a_{ij}}. \] As choices of \(W\), the authors consider a list of 14 exceptional unimodular singularities. This is the first list of examples considered with central charge \(\hat{c}_W>1\) and the authors prove for these singularities the LG/LG mirror conjecture, which we outline now. The \(A\)-model associated to \((W,G)\) is the FJRW-theory developed by Fan-Jarvis-Ruan based on a proposal by Witten. It solves the moduli problem for the Witten equations of a LG model \((W,G)\). The output consists of the FJRW invariants which depend on genus, tautological classes on the moduli space of stable curves and classes in the state space \(H_W\) for \((W,G)\). The state space has a multiplication \(\bullet\) defined by an intersection pairing with the genus 0 primary FJRW invariants with three marked points. The invariants at all genera are collected into the total ancestor potential \(\mathscr{A}_W^{FJRW}\). Note that the FJRW invariants are related to Gromov-Witten invariants of a Calabi-Yau manifold via the LG/CY correspondence. We now describe the \(B\)-model Saito-Givental theory for \((W^T,\{1\})\). Denote by \(\text{Jac}(W^T)\) the Jacobian ring of \(W^T\). Following Krawitz, there is a ring isomorphism \(\text{Jac}(W^T)\cong (H_W,\bullet)\). Consider the germ \(f\) of \(W^T\) at the origin, i.e. \[ f : (\mathbb{C}^n,\mathbf{0}) \to (\mathbb{C},0). \] For the Milnor number \(\mu = \dim_\mathbb{C}\text{Jac}(f)_\mathbf{0}\), consider the universal unfolding \[ (\mathbb{C}^n\times\mathbb{C}^\mu,\mathbf{0}\times\mathbf{0}) \to (\mathbb{C}\times\mathbb{C}^\mu,0\times\mathbf{0}), \] where \(\mathbb{C}^\mu\) parametrizes the deformations. The next step involves choosing a primitive form \(\zeta\), which is a relative holomorphic volume form at the germ \((\mathbb{C}^n\times\mathbb{C}^\mu,\mathbf{0}\times\mathbf{0})\) and which induces a Frobenius manifold structure at the germ \((\mathbb{C}^\mu,\mathbf{0})\). From this data is extracted the Saito-Givental potential \(\mathscr{A}_{W^T}^{SG}\). The LG/LG mirror conjecture, the theorem proven by the authors, now states that there is a ring isomorphism \(\text{Jac}(W^T)\cong (H_W,\bullet)\) and a choice of primitive form \(\zeta\) such that the potentials \(\mathscr{A}_W^{FJRW}\) and \(\mathscr{A}_{W^T}^{SG}\) are identified. The proof, as often in mirror symmetry, proceeds by calculating the potentials on both sides. This involves the WDVV equations. The paper is the continuation of the work by Fan-Jarvis-Ruan on ADE singularities (\(\hat{c}_W <1\)) and by Krawitz-Milanov-Shen on simple elliptic singularities (\(\hat{c}_W =1\)). It is well-written, contains an ample introduction and clarifies the connections to other aspects of mirror symmetry.