The cornering and GIT-specialisation diagram for Kleinian quiver varieties

Let 0JJI0 \in J \subseteq J' \subseteq I, set KIJK \coloneqq I \setminus J and KIJK' \coloneqq I \setminus J', and let σK,K\sigma_{K,K'} be the cone defined by

\sigma_{K,K'} \coloneqq \left\\{\theta \in \Theta_{n\delta} \\; \Bigg| \\; \begin{matrix*}[l] \theta(\delta) \geq 0 \\;, & \\ \\\\ \theta(\alpha_j) \geq (n-1)\theta(\delta) & \text{for} \\; j \in J \setminus \\{0\\} \\;, \\\\ \theta(\alpha_k) \geq 0 & \text{for} \\; k \in K \setminus K' \\;, \\\\ \theta(\alpha_k) = 0 & \text{for} \\; k \in K' \\;. \end{matrix*} \right\\}.

Let hEJh_*\operatorname{\mathcal{E}}_{J'} and hVJh_*\operatorname{\mathcal{V}}_{J'} be the cornered algebra and module, and write QuotJ,Jv\operatorname{Quot}_{J,J'}^v for the corresponding Quot scheme. The cornering and GIT-specialisation conjecture. If θCK\theta \in C_K, θ\theta' lies in the relative interior of σK\sigma_K, and θ\theta” lies in the relative interior of σK,K\sigma_{K,K'}, then there is a commutative diagram

\begin{tikzcd} \operatorname{Quot}_J^{n \delta} \arrow[r, "\sim"] \arrow[d] & \operatorname{\mathfrak{M}}_{\theta}(n \delta,w) \arrow[d] \\\\ \operatorname{Quot}_{J,J'}^{n \delta} \arrow[r, "\sim"] \arrow[d] & \operatorname{\mathfrak{M}}_{\theta”}(n \delta,w) \arrow[d] \\\\ \operatorname{Hilb}^n(X_J) \arrow[r, "\sim"] & \operatorname{\mathfrak{M}}_{\theta'}(n \delta,w), \end{tikzcd}

in which all morphisms are birational and projective, the horizontal morphisms are isomorphisms induced by gg_*, the left vertical morphisms are induced by cornering at JJ' and JJ, respectively, and the right vertical morphisms arise from GIT specialisation. This conjectures an interpolation between the Quot scheme, the intermediate cornered Quot scheme, and the Hilbert scheme, matching the corresponding Nakajima quiver varieties.

Sources & referencesView supporting material

Primary source

Lukas Bertsch and Ruth Wye, “Stacky Resolutions of Kleinian Singularities and Nakajima Quiver Varieties”, arXiv:2509.22569 (2026).

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.