Stable High-Order Interpolation on the Grassmann Manifold by Maximum-Volume Coordinates and Arnoldi Orthogonalization

Qiang Niu, Wen Jiang, Jie Fei, Ruoyu Xiong, Yuxuan Li · 2026

Read Stable High-Order Interpolation on the Grassmann Manifold by Maximum-Volume Coordinates and Arnoldi Orthogonalization by Qiang Niu, Wen Jiang, Jie Fei, Ruoyu Xiong, Yuxuan Li (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-23858.txt.

Introduction

Niu and co-authors build a stable way to interpolate at high order on the Grassmann manifold. Standard maps based on singular value decomposition cost much and can make derivatives unstable. Their method first maps data into maximum-volume coordinates, a well-conditioned flat domain. It then builds polynomial bases with an Arnoldi process, avoiding ill-conditioned linear systems. A confluent version handles derivative data for Hermite interpolation. They prove stability bounds for the mapping and the approximation, and tests show high accuracy at high degree.

Catalog record

CREATORQiang Niu, Wen Jiang, Jie Fei, Ruoyu Xiong, Yuxuan Li
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE73563
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.23858

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Stable High-Order Interpolation on the Grassmann Manifold by Maximum-Volume Coordinates and Arnoldi Orthogonalization Qiang Niu, Wen Jiang, Jie Fei, Ruoyu Xiong, Yuxuan Li Abstract High-order interpolation on the Grassmann manifold Gr ⁡ ( n , p ) \Gr(n,p) is often hindered by the computational overhead and derivative instability of SVD-based geometric mappings. To solve the challenges, we propose a stabilized framework that combines Maximum-Volume (MV) local coordinates with Arnoldi-orthogonalized polynomial bases. First, manifold data are mapped to a well-conditioned Euclidean domain via MV coordinates. The approach bypasses the costly matrix factorizations inherent to traditional Riemannian normal coordinates. Within the coordinate space, we use the Vandermonde-with-Arnoldi (V+A) method for Lagrange interpolation and its confluent extension (CV+A) for derivative-enriched Hermite interpolation. By constructing discrete orthogonal bases directly from the parameter nodes, the solution of ill-conditioned linear system is avoided. Theoretical bounds are established to verify the stability of the geometric mapping and the polynomial approximation. Extensive numerical experiments demonstrate that the proposed MV-(C)V+A framework can produce highly accurate approximation in high-degree polynomial interpolation. Keywords. Grassmann manifold, interpolation, Arnoldi method, Maximum-Volume coordinates, QR decomposition. MSC(2020). 15B10, 65D05, 65F99, 53C30, 53C80 1 Introduction Interpolation of manifold-valued data is important in scientific and engineering applications where the geometric structure must be preserved [ 8 , 2 , 12 , 32 , 3 , 39 ] . Interpolation on the Grassmann manifold \Gr(n,p) is widely used in parametric model order reduction [ 5 , 3 , 34 ] , computer vision [ 22 , 29 ] , and machine learning [ 20 , 23 ] . In parametric model order reduction, reduced bases computed via proper orthogonal decomposition at sampled parameters are treated as points on the Grassmann manifold, which enables efficient approximation at unseen parameters [ 27 , 16 ] . Since Euclidean interpolation of matrix entries fails to respect the underlying quotient geometry, manifold-aware techniques are often needed [ 4 , 1 ] . High-order interpolation on manifolds usually has two numerical difficulties. The geometric challenge comes from traditional normal-coordinate formulations, which rely on matrix decompositions such as SVD-based logarithm and exponential maps [ 8 , 18 , 10 , 21 ] . For Hermite interpolation, where derivative data must be transported, it requires differentiating through matrix decompositions, an operation that is computationally costly and numerically sensitive [ 41 , 40 , 37 ] . The algebraic challenge arises once data is mapped to the coordinate space: monomial-based interpolation leads to Vandermonde systems whose condition numbers grow exponentially with polynomial degree, which makes unstable coefficient recovery and error amplification [ 6 , 17 , 30 ] . The numerical instability often limits practitioners to low-degree approximations, even when the underlying function is smooth. Recently, Zimmermann [ 41 , 36 ] developed a general approach for Hermite interpolation on Riemannian manifolds and quantified error propagation from tangent space to manifold. Jensen and Zimmermann [ 19 ] introduced maximum-volume coordinates, a matrix-decomposition-free local coordinate for \Gr(n,p) . By selecting Stiefel subblocks with maximized determinant, the approach ensures well-conditioned geometric mappings without SVD computations. The MV approach also provides an algebraic derivative mapping formula that avoids differentiating matrix factorizations. The maximum-volume principle, originally developed for pseudoskeleton approximations and rank-revealing factorizations, has proven effective for constructing well-conditioned submatrices [ 15 , 14 ] . Motivated by the recent success of the Vandermonde-with-Arnoldi (V+A) method and its confluent (CV+A) and multivariate extensions in eliminating algebraic instability [ 6 , 25 , 33 , 35 ] , this paper develops a stable polynomial interpolation framework for data on the Grassmann manifold. By integrating these Arnoldi-based techniques with maximum-volume (MV) coordinates, our approach simultaneously resolves two critical numerical bottlenecks: MV coordinates geometrically stabilize the domain mapping without costly SVDs, while the (C)V+A method avoids the ill-conditioning of high-degree polynomial construction. Building on recent advances in manifold error analysis [ 38 , 24 , 28 ] , the MV-(C)V+A framework consists of the following steps. First, geometric preprocessing is performed via Householder-stabilized coordinates, which controls the geometric condition number and ensures a well-conditioned local representation of the manifold. Second, stable Euclidean interpolation is carried out using Arnoldi-orthogonalized polynomial bases and then the interpolated coordinates are mapped back to the manifold through a reconstruction procedure. Finally, a rigorous convergence and error analysis of the MV-(C)V+A method is established, which shows that the total interpolation error admits a decomposition into geometric and algebraic components and achieves high precision. The remainder of this paper is organized as follows. Section 2 reviews MV coordinates and the associated derivative mappings. Section 3 recalls the V+A and CV+A methods. Section 4 introduces the integrated MV-(C)V+A framework together with complete algorithmic formulations. Section 5 develops the convergence theory and provides explicit error bounds. Section 6 reports numerical experiments and discusses performance of the proposed method. Section 7 concludes the paper and outlines future work. 2 Interpolation on the Grassmann Manifold Let \Gr(n,p) denote the Grassmann manifold of p -dimensional subspaces in \mathbb{R}^{n} , represented numerically through Stiefel matrices U\in\St(n,p) with orthonormal columns. Each subspace \mathcal{U}\in\Gr(n,p) corresponds to an equivalence class [U]=\{UQ:Q\in O(p)\} . Let f:D\rightarrow\Gr(n,p) be a differentiable manifold-valued function defined on a parameter domain D\subset\mathbb{R} . Given a discrete data set \{(t_{i},\mathcal{U}_{i})\}_{i=1}^{m} sampled from the function, where t_{i}\in\mathbb{R} are parameter nodes and \mathcal{U}_{i}\in\Gr(n,p) are the corresponding subspaces, the goal is to construct an interpolant \hat{f}:D\rightarrow\Gr(n,p) such that \hat{f}(t_{i})=\mathcal{U}_{i} for all i . For Hermite interpolation, the problem is extended by requiring the interpolant to also match manifold velocities at the sample points. In numerical implementations, each subspace \mathcal{U}_{i} is specified by a Stiefel representative matrix U_{i}\in\mathrm{St}(n,p) such that \operatorname{span}(U_{i})=\mathcal{U}_{i} , and its velocity is represented by a tangent lift \dot{U}_{i}\in T_{U_{i}}\mathrm{St}(n,p) . To remove the ambiguity in the choice of representative, we take \dot{U}_{i} to be a horizontal lift satisfying U_{i}^{T}\dot{U}_{i}=0 . The interpolant matrix function \hat{U}(t) must therefore satisfy \operatorname{span}(\hat{U}(t_{i}))=\mathcal{U}_{i},\quad i=1,\dots,m, and the corresponding local coordinate derivatives must match the specified tangent data at each node t_{i} . The maximum-volume (MV) coordinate [ 19 ] is a matrix-decomposition-free alternative to Riemannian normal coordinates for representing subspaces on the Grassmann manifold. The approach exploits the natural block structure of Stiefel representatives to construct local coordinates using only algebraic operations, and avoids the singular value decompositions. Given a sample basis U\in\St(n,p) and a row permutation P , we partition PU=\begin{bmatrix}\widetilde{U}_{1}\\ \widetilde{U}_{2}\end{bmatrix},\qquad\widetilde{U}_{1}\in\mathbb{R}^{p\times p},\quad\widetilde{U}_{2}\in\mathbb{R}^{(n-p)\times p}. The MV coordinates are then defined as \Xi(U)=\widetilde{U}_{2}\widetilde{U}_{1}^{-1}\in\mathbb{R}^{(n-p)\times p},\quad N_{\text{coord}}=(n-p)p. (1) Unlike Riemannian normal coordinates, the mapping avoids SVD, matrix logarithms, and matrix exponentials entirely within the interpolation loop, requiring only matrix inversion and multiplication. The numerical quality of ( 1 ) depends critically on the conditioning of \widetilde{U}_{1} . The MV strategy selects permutations to maximize |\det(\widetilde{U}_{1})| , which is equivalent to minimizing \|\widetilde{U}_{1}^{-1}\|_{F} . Jensen and Zimmermann [ 19 ] established a bound guaranteeing \|\Xi\|_{F}\leq\sqrt{p(n-p)} , ensuring that the coordinate representation remains well-conditioned. The maximum-volume principle, originally developed for pseudoskeleton approximations and rank-revealing factorizations, has proven effective for constructing well-conditioned submatrices [ 15 , 14 ] . In the original MV framework [ 19 ] , for multiple samples \{U_{i}\}_{i=1}^{m} , a single global permutation is typically chosen to balance conditioning across all points: P_{\star}\approx\arg\min_{P}\max_{1\leq i\leq m}\|\tilde{U}_{i,1}^{-1}\|_{F}. (2) Geometrically, selecting a single global permutation

Read the complete text