Knot invariants from topological recursion on augmentation varieties (Q2018313)

This paper studies the following question: How must we extend remodeling and topological recursion methods to accommodate the calculation of colored HOMFLY polynomials? The paper observes that the usual approach, using topological recursion based directly on the Bergman kernel, cannot suffice. Instead, the paper argues, it is crucial to replace the Bergman kernel by the ``calibrated annulus kernel'' introduced in the paper. The main evidence for the approach taken in the paper are detailed calculations of low-genus amplitudes for torus knots, but the authors optimistically suggest that their approach is correct to all orders for all knots. This leads them to some remarkable conjectures about knot theory, most notably that the full data of the colored HOMFLY polynomials of a knot for arbitrary representation can be extracted from the data just of the colored HOMFLY polynomials corresponding to Young diagrams with at most two rows. Connections are also made to knot contact homology augmentation varieties.