Perversity equals weight for Painlevé spaces (Q2020380)

Let \(M\) be a compact Kähler manifold and \(G\) a connected reductive affine algebraic group. The non-abelian Hodge correspondence is a homeomorphism (stratified diffeomorphism) between the moduli space of (twisted) \(G\)-representations of the fundamental group of \(M\) (called a ``character variety'') and the moduli space of \(G\)-Higgs bundles over \(M\). This correspondence can be extended to the so-called ``wild'' moduli spaces of representations and Higgs bundles, respectively. Very loosely speaking, this means allowing certain irregular singularities on one side of the correspondence, and Higgs fields with certain poles on the other side. The \(P=W\) conjecture of de Cataldo, Hausel and Migliorini says that the weight filtration \(W\) on the character variety should correspond, via the non-ableian Hodge correspondence, to the perverse Leray filtration \(P\) on the moduli space of Higgs bundles. See [\textit{M. A. A. De Cataldo} et al., Ann. Math. (2) 175, No. 3, 1329--1407 (2012; Zbl 1375.14047)] for the first established cases of this conjecture. This is surprising since it implies \(W\), despite only depending on the topology of \(M\), keeps track of topological properties of the Hitchin map, which depends on the complex structure of \(M\). Following the notation of the paper under review, let \(PX\) be a Painlevé system where \(X\) is any of \(I, II, III(D6), III(D7), III(D8), IV, V_{\deg} , V, VI\). To each such system there is a wild character variety \(\mathcal{M}_B^{PX}(G)\) and a wild moduli space of Higgs bundles \(\mathcal{M}_{Dol}^{PX}(G)\). The main theorem of this well-written and timely paper establishes the \(P=W\) conjecture between \(\mathcal{M}_B^{PX}(G)\) and \(\mathcal{M}_{Dol}^{PX}(G)\) for all \(X\) when \(G=\mathrm{GL}(2,\mathbb{C})\). Let \(D =\sum_i m_i p_i\) be a divisor on \(\mathbb{C}P^1\) such that \(\sum_i m_i = 4\). Then the moduli spaces \(\mathcal{M}_{Dol}^{PX}(\mathrm{GL}(2,\mathbb{C}))\) are rank 2 meromorphic parabolic Higgs bundles over \(\mathbb{C}P^1\) with poles at \(D\). On the other hand, the moduli spaces \(\mathcal{M}_B^{PX}(\mathrm{GL}(2,\mathbb{C}))\) are smooth affine cubic surfaces. In Section 3 of the paper, the author determines the perverse filtrations on \(\mathcal{M}_{Dol}^{PX}(\mathrm{GL}(2,\mathbb{C}))\), and in Section 4 the weight filtrations on \(\mathcal{M}_B^{PX}(\mathrm{GL}(2,\mathbb{C}))\) are established. The author does a direct comparison of these computations after making numerous simplifying observations; which establishes the main theorem. This by no means is a small task, and is a significant and interesting result about a deep conjecture.