The canonical wall structure and intrinsic mirror symmetry (Q2162741)

Mirror symmetry is a phenomenon discovered by physicists and used by mathematicians apriori as a tool for making predictions on enumerative invariants of Calabi-Yau varieties. However, in the last thirty years, apart from verifying such predictions, several mathematical techniques have been developed by pure mathematicians to investigate relationships between complex and symplectic geometries of Calabi-Yau varieties that are manifested by mirror symmetry. One of the most celebrated conjectures in mirror symmetry, due to \textit{A. Strominger} et al. [Nucl. Phys., B 479, No. 1--2, 243--259 (1996; Zbl 0896.14024)], relates the complex geometry of a Calabi-Yau threefold with integral affine geometry on a three-dimensional real manifold \(B\). In rough terms, the manifestation of mirror symmetry in this context is that mirror pairs of Calabi-Yau varieties \((X,\breve{X})\) admit dual special Lagrangian torus fibrations \(f\colon X\to B\), and \(\breve{f} \colon \breve{X}\to B\). with singular fibers over a discriminant locus \(\Delta \subset B\). The mirror to \(X\), in the sense of Strominger-Yau-Zaslow (SYZ), can be obtained by first constructing a semi-flat mirror to the restriction of \(X\) to \(B\setminus \Delta\) obtained by dualizing the non-singular torus fibers and then compactifying it by considering appropriate corrections of the complex structure. These corrections are expected to be captured by counts of holomorphic discs in \(X\) with boundaries on torus fibers [\textit{K. Fukaya}, Proc. Symp. Pure Math. 73, 205--278 (2005; Zbl 1085.53080)]. However, constructing the mirror to a Calabi-Yau following this strategy is a notoriously difficult problem, in particular, as showing the existence of special Lagrangian fibrations is difficult. Moreover, it is technically challenging to provide a precise description of the counts of holomorphic disks with boundaries on the torus fibers. Nonetheless, following early insights of Kontsevich-Soibelman [\textit{M. Kontsevich} and \textit{Y. Soibelman}, Prog. Math. 244, 321--385 (2006; Zbl 1114.14027)], Gross-Siebert show that algebro-geometric analogues of such counts of holomorphic disks can be recovered from the combinatorics of a normal crossing limit \(X_0\) of a large complex structure degeneration \((X_t)\) of \(X\). In particular, they reduce the problem to determine such counts to a combinatorial problem encoded on a \textit{canonical wall structure} on the affine manifold \(B\) which is viewed as a tropical limit of \(X_0\). The construction of the canonical wall structure in this paper, builds on ideas of the Gross-Siebert program, which the authors have been developing since early 2000's. The Gross-Siebert program provides the most general mirror construction algebro-geometrically. In particular, this approach allows one to overcome the aforomentioned challenges in the context of SYZ mirror symmetry. Here, one does not need the special Lagrangian fibration \(X \rightarrow B\), but only the base \(B\) with a structure of integral affine manifold with singularities. Further, the theory of algebro-geometric counter-parts of the holomorphic discs appearing in SYZ mirror symmetry has been developed by the authors and their collaborators recently. These counterparts are given in terms of rational curves in the total space of the family \((X_t)\) with tangency conditions along the normal crossings divisor defined by the special fiber \(X_0\). Such counts are examples of so-called punctured Gromov-Witten invariants of \textit{D. Abramovich} et al. [``Punctured logarithmic maps'', Preprint, \url{arXiv:2009.07720}]. Building on punctured Gromov-Witten theory, Gross and Siebert recently provided a general mirror construction, known as ``intrinsic mirror symmetry'' [\textit{M. Gross} and \textit{B. Siebert}, Proc. Symp. Pure Math. 97, 199--230 (2018; Zbl 1448.14039)], in which the homogeneous coordinate ring of the mirror is directly constructed in terms of an explicit linear basis of so-called ``theta functions'', and where the structure constants of the product are given by punctured Gromov-Witten invariants. Whereas the ``intrinsic mirror symmetry'' construction is very direct, it is a bit far from the original SYZ picture. In the current paper, Gross and Siebert use punctured Gromov-Witten theory to give an algebro-geometric construction of the ``canonical wall structure'', which is the expected combinatorial structure on the base \(B\) of the SYZ fibration defined by the counts of Maslov index zero holomorphic disks. Moreover, they show that the mirror which can be constructed from this wall structure agrees with the mirror obtained by intrinsic mirror symmetry. Technically, the contributions of the paper are as follows. Given a maximal log Calabi-Yau variety or a large complex structure degeneration of Calabi-Yau varieties satisfying some technical conditions, the authors in this article provide: \begin{itemize} \item[1.] The definition of the canonical wall structure using punctured Gromov-Witten invariants, \item[2.] The definition of logarithmic theta functions using punctured Gromov-Witten invariants, \item[3.] The proof that the logarithmic theta functions agree with combinatorially defined theta functions constructed from the canonical wall structure (Theorem A), \item[4.] The proof that the canonical wall structure is consistent (Theorem B), and; \item[5.] The proof that the mirror constructed from the canonical wall structure agrees with the mirror constructed by intrinsic mirror symmetry (Theorem C). \end{itemize} The proof of Theorem A uses the gluing formula for punctured Gromov-Witten invariants recently proved by \textit{Y. Wu} [``Splitting of Gromov-Witten invariants with toric gluing strata'', Preprint, \url{arXiv:2103.14780}]. The proof of Theorem B uses Theorem A and a version of deformation invariance in punctured Gromov-Witten theory, and more precisely follows from an explicit rational equivalence between cycles. Finaly, the proof of Theorem C also relies on the gluing formula for punctured Gromov-Witten invariants. This paper symmetry is a key contribution in mirror symmetry, as it provides a general algebro-geometric mirror construction in the spirit of the more than 20 years old SYZ conjecture. Explicit examples of this construction building on this paper have recently been worked out by \textit{H. Argüz} and \textit{M. Gross} [Geom. Topol. 26, No. 5, 2135--2235 (2022; Zbl 07632769)].