\(sl(N)\)-link homology \((N\geq 4)\) using foams and the Kapustin-Li formula (Q1006148)

The \(\mathfrak{sl}(N)\) link polynomial \(P_{N}(q)\) is given by evaluating the 2-variable HOMFLY polynomial at \((q^{N},q)\). This gives the Alexander polynomial when \(N=0\), the trivial polynomial when \(N=1\), and the Jones polynomial when \(N=2\). Categorifications of these polynomials, i.e.\ bigraded homology theories with graded Euler characteristic equal to \(P_{N}(q)\), were first given by \textit{M. Khovanov} [Duke Math. J. 101, No.~3, 359--426 (2000; Zbl 0960.57005)] in the case \(N=2\) and \textit{P. Ozsváth} and \textit{Z. Szabó} [Adv. Math. 186, No.~1, 58--116 (2004; Zbl 1062.57019)] and \textit{J. Rasmussen} [Floer homology and knot complements; \url{arXiv:math.GT/0306378}] for \(N=0\). \textit{M. Khovanov} [Algebr. Geom. Topol. 4, 1045--1081 (2004; Zbl 1159.57300)] later used webs (planar trivalent graphs with single and double edges) and foams (singular cobordisms between webs) to categorify \(P_{3}\), and \textit{M. Khovanov} and \textit{L. Rozansky} [Fundam. Math. 199, No.~1, 1--91 (2008; Zbl 1145.57009)] used a construction involving matrix factorisations to categorify \(P_{N}\) for \(N\geq 4\). In their paper, Khovanov and Rozansky state that an approach using foams should shed additional light on these link invariants. Such a topological approach is taken in the paper under review. A link homology \(H^{*}({\mathcal F}(L))\) with graded Euler characteristic \(P_{N}\) is defined from a link diagram using webs and foams and shown to be invariant under Reidemeister moves, and also to extend to a functor from the category of links and link cobordisms. For any link \(L\) the complex \({\mathcal F}(L)\) is shown to be isomorphic to the corresponding Khovanov-Rozansky complex, and the authors conjecture that their functor is isomorphic to that of Khovanov-Rozansky. The authors note that the foams for \(N\geq 4\) contain more complicated singularities than in the \(N=3\) case. They remark that their construction is almost completely combinatorial but not quite, as ``the introduction of the new singularities makes it much harder to evaluate closed foams and we do not know how to do it combinatorially. Instead we use the Kapustin-Li formula adapted by Khovanov and Rozansky\dots Although we have not completely achieved our final goal we believe we have made good progress towards it.''