Dolbeault S-duality matching conjecture for tempered magical triples

Let ρ:sl2g\rho:\mathfrak{sl}_2\to\mathfrak{g} be a tempered magical sl2\mathfrak{sl}_2-triple. Assume that MρM_\rho is hyperspherical, equivalently XρX_\rho has connected generic stabilizers, and that the representation SρVρVρS_\rho\simeq V_\rho\oplus V_\rho^* of Gˇρ\check{G}_\rho is of cotangent type. Let GG be the adjoint group with Lie algebra g\mathfrak{g}, let Gˇ\check{G} be its Langlands dual group, and regard the Cayley morphism for GρRG^\mathbf{R}_\rho as a morphism of sheaves

ADol(ϕ):ADol(Mρ)ADol(Mρ)QC(HiggsG).A_{\mathrm{Dol}}(\phi):A_{\mathrm{Dol}}(M_\rho)\longrightarrow A_{\mathrm{Dol}}(M_\rho')\in\mathrm{QC}(\mathrm{Higgs}_G).

Dolbeault S-duality matching conjecture. Over the \diamondsuit-locus, the following matching data hold under SDol\mathbf{S}_{\mathrm{Dol}}: the dual of ADol(Mρ)A_{\mathrm{Dol}}(M_\rho') is BDol(Mˇρ)=ω(HiggsGˇρ)B_{\mathrm{Dol}}(\check{M}_\rho')=\omega(\mathrm{Higgs}_{\check{G}_\rho}), up to tensoring with a finite-rank vector bundle WρW_\rho in the \ast-cases; the dual of ADol(Mρ)A_{\mathrm{Dol}}(M_\rho) is

BDol(Mˇρ)=k0(SymkDHVρ)[k],B_{\mathrm{Dol}}(\check{M}_\rho)=\bigoplus_{k\geq 0}(\operatorname{Sym}^k\mathrm{DH}_{V_\rho})[k],

again tensored with WρW_\rho in the \ast-cases; and the dual of ADol(ϕ)A_{\mathrm{Dol}}(\phi) is the natural morphism BDol(Mˇρ)BDol(Mˇρ)B_{\mathrm{Dol}}(\check{M}_\rho)\to B_{\mathrm{Dol}}(\check{M}_\rho') induced by projection of the Dirac--Higgs bundle to its zero section,

DHVρHiggsGˇρ,\mathrm{DH}_{V_\rho}\longrightarrow\mathrm{Higgs}_{\check{G}_\rho},

with the same tensoring convention. The claim is a proposed matching under relative Langlands/S-duality over the specified locus; no resolution is supplied in the source.

Sources & referencesView supporting material

Primary source

Eric Y. Chen, Enya Hsiao and Mengxue Yang, “(BAA)-branes from higher Teichmüller theory”, arXiv:2508.09562 (2025).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.