Finiteness of the Grassmann distance degree

Determine whether the Grassmann distance (GD) degree is finite for every irreducible model \mathcal{M}\subseteq \mathrm{pGr}(k,n) with Zariski-dense real points, beyond the established finiteness result for models in \mathrm{pGr}(2,n).

Background

The paper defines the GD degree as the number of complex critical points of the Euclidean trace optimization problem when the data point is a generic rank-k orthogonal projection matrix in \mathrm{pGr}(k,n). Unlike the Euclidean distance degree, whose finiteness follows from the dimension of the Euclidean distance correspondence, the GD correspondence can be reducible and may have positive-dimensional generic fibers.

The authors prove in Theorem 3.6 that every model in \mathrm{pGr}(2,n) has finite GD degree. The unresolved question is whether this finiteness extends to arbitrary ranks kk and all irreducible models satisfying the paper’s standing hypotheses.

References

It is still unknown whether the GD degree is always a finite number.

— Euclidean Distance Optimization Within the Grassmannian  (2610.02018 - Friedman et al., 1 Oct 2026) in Section 1, paragraph beginning “This article is organized as follows”; Section 3 (Cut Locus and GD Degree)