Topological recursion for the conifold transition of a torus knot (Q2414598)

The conifold transition is one of the basic building blocks of mirror symmetry going back to [\textit{M. Reid}, Math. Ann. 278, 329--334 (1987; Zbl 0649.14021)], in that it can be used to construct mirrors: the operation of smoothing is expected to be mirror to the operation of resolution. A torus knot \(K\) is a knot that is realized on a real torus in \(S^3\). Trough the conifold transition relating \(T^*S^3\) to the total space \(\mathcal{X}\) of \(\mathcal{O}_{\mathbb{P}^1}(-1)\oplus\mathcal{O}_{\mathbb{P}^1}(-1)\to\mathbb{P}^1\), \(K\) determines a Lagrangian \(L_K\) in \(\mathcal{X}\). Consider the \(A\)-model all genus open-closed Gromov-Witten invariants of \((\mathcal{X},L_K)\). Mirror symmetry predicts that there is a mirror \(B\)-model geometry, in this case a spectral curve, and that Eynard-Orantin topological recursion on the spectral curve computes under the mirror map the open-closed Gromov-Witten invariants of \((\mathcal{X},L_K)\). This was conjectured in [\textit{A. Brini} et al., Ann. Henri Poincaré 13, No. 8, 1873--1910 (2012; Zbl 1256.81086)] and [\textit{D. E. Diaconescu} et al., Commun. Math. Phys. 319, No. 3, 813--863 (2013; Zbl 1309.57011)]. The main result of the present article is a proof of that conjecture. By the results of [\textit{G. Borot} et al., Commun. Number Theory Phys. 9, No. 1, 51--187 (2015; Zbl 1329.14074)], these invariants are in turn equivalent to the colored HOMFLY polynomial of \(K\). The authors start by reviewing the construction of \(L_K\). Then they define the open-closed Gromov-Witten invariants via localization and compute their potential as a graph-sum formula. Of particular interest is their definition of the open Gromov-Witten invariants of \((\mathcal{X},L_K)\) through relative Gromov-Witten invariants of formal toric Calabi-Yau threefolds as in [\textit{J. Li} et al., Geom. Topol. 13, No. 1, 527--621 (2009; Zbl 1184.14084)] and their proof of equivalence to the definition in terms of localization. After giving an overview of Eynard-Orantin topological recursion for the mirror spectral curve and recalling the expression of its \(B\)-model higher genus invariants, the authors end by comparing the invariants on both sides of the mirror finding agreement under the mirror map. The article, while technically challenging, is proving an important result in mirror symmetry. Sufficient detail is provided to introduce the results used and the reader might view it as an introduction to the conifold transition and the enumerative theories associated to it.