pith. sign in

arxiv: 1907.07119 · v1 · pith:MAUXVQBKnew · submitted 2019-07-16 · 🧮 math.NA · cs.NA

On the smallest singular value of multivariate Vandermonde matrices with clustered nodes

Stefan Kunis , Dominik Nagel This is my paper

Pith reviewed 2026-05-24 20:37 UTC · model grok-4.3

classification 🧮 math.NA cs.NA
keywords Vandermonde matrixsmallest singular valueclustered nodesunit circlemultivariatelower boundconditioning
0
0 comments X

The pith

The smallest singular value of rectangular multivariate Vandermonde matrices with clustered nodes on the unit circle is bounded below by the product of inverse distances within each cluster.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper proves lower bounds on the smallest singular value for rectangular multivariate Vandermonde matrices generated by nodes on the complex unit circle that form clusters. The bound equals the product, for each node, of the reciprocals of its distances to the other nodes in the same cluster. Upper bounds are derived as well, which together resolve the univariate case completely and the case of two-node clusters in higher dimensions, including explicit constants. For clusters larger than two nodes the value further depends on the precise geometric arrangement of points inside the cluster.

Core claim

For rectangular multivariate Vandermonde matrices generated by distinct nodes on the unit circle that form clusters, the smallest singular value is bounded below by the product of the inverted distances of a node to all other nodes in its specific cluster. Matching upper bounds hold in the univariate setting and for pairs of nodes in the multivariate setting, with the bound depending on internal cluster geometry for larger groups.

What carries the argument

The product of inverted intra-cluster distances, which supplies the explicit lower bound on the minimal singular value of the Vandermonde matrix.

If this is right

  • The univariate case receives a complete characterization with reasonable sharp constants.
  • Pairs of nodes receive a complete characterization in the multivariate setting with reasonable sharp constants.
  • Larger clusters require an additional factor that accounts for the geometric configuration inside the cluster.
  • The matrix remains invertible with controlled conditioning whenever inter-cluster separations are large compared with intra-cluster distances.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The bound isolates the effect of intra-cluster proximity from the placement of the clusters themselves on the circle.
  • A direct numerical check for a three-node cluster would show whether the geometry dependence can be expressed by a simple additional factor.
  • The same distance-product form may govern conditioning for other polynomial bases that admit a Vandermonde structure.

Load-bearing premise

The nodes lie on the complex unit circle, are distinct, and form identifiable clusters whose internal distances control the bound.

What would settle it

Compute the smallest singular value of the Vandermonde matrix for any chosen pair of nodes on the unit circle separated by distance d and verify whether it lies above or below the value 1/d up to the paper's stated constants.

Figures

Figures reproduced from arXiv: 1907.07119 by Dominik Nagel, Stefan Kunis.

Figure 2
Figure 2. Figure 2 [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 6
Figure 6. Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p014_6.png] view at source ↗
Figure 6
Figure 6. Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p015_6.png] view at source ↗
Figure 6
Figure 6. Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p016_6.png] view at source ↗
read the original abstract

We prove lower bounds for the smallest singular value of rectangular, multivariate Vandermonde matrices with nodes on the complex unit circle. The nodes are ``off the grid'', groups of nodes cluster, and the studied minimal singular value is bounded below by the product of inverted distances of a node to all other nodes in the specific cluster. By providing also upper bounds for the smallest singular value, this completely settles the univariate case and pairs of nodes in the multivariate case, both including reasonable sharp constants. For larger clusters, we show that the smallest singular value depends also on the geometric configuration within a cluster.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 3 minor

Summary. The manuscript proves lower bounds on the smallest singular value of rectangular multivariate Vandermonde matrices whose nodes lie on the complex unit circle and form identifiable clusters. The lower bound is given by the product of the reciprocals of the intra-cluster distances from each node to the others in its cluster. Matching upper bounds are derived that completely settle the univariate case and the bivariate case of node pairs (including explicit constants). For clusters larger than pairs, the smallest singular value is shown to depend additionally on the specific geometric configuration inside the cluster.

Significance. If the stated bounds and proofs hold, the work supplies sharp, explicit characterizations of the conditioning of Vandermonde matrices under clustering, which is directly useful in numerical analysis, polynomial interpolation, and exponential fitting. The provision of both lower and upper bounds with constants for the univariate and pair cases, together with the explicit identification of geometric dependence for larger clusters, constitutes a clear advance over purely asymptotic or non-sharp estimates.

minor comments (3)
  1. [§4] The transition from the univariate proof to the bivariate pair case in §4 could include a short remark on how the separation assumption between clusters is used to control the error terms.
  2. [Definition 2.3] Notation for the rectangular multivariate Vandermonde matrix (Definition 2.3) would be clearer if an explicit low-dimensional example were added showing the block structure induced by the clusters.
  3. [Theorem 5.1] The statement that the bound for larger clusters depends on intra-cluster geometry (Theorem 5.1) is correct but would benefit from a brief sentence indicating which geometric quantities appear in the dependence.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive assessment of the manuscript, the accurate summary of its contributions, and the recommendation of minor revision. No major comments were provided in the report.

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The paper derives lower and upper bounds on the smallest singular value of rectangular multivariate Vandermonde matrices directly from matrix properties and geometric distances between clustered nodes on the unit circle. These bounds are established via standard analytic techniques for such matrices, without any reduction to fitted parameters, self-definitional constructions, or load-bearing self-citations. The univariate case and bivariate pairs are settled completely with explicit constants, while larger clusters explicitly require additional intra-cluster geometry; all steps remain consistent with the stated assumptions of distinct nodes and separated clusters. The derivation chain is self-contained against external benchmarks and exhibits none of the enumerated circularity patterns.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard linear-algebra facts about singular values of Vandermonde matrices and on the geometric assumption that nodes lie on the unit circle; no free parameters or new entities are introduced.

axioms (2)
  • standard math Standard properties of singular values and determinants for rectangular Vandermonde matrices
    Used to obtain both lower and upper bounds on the minimal singular value.
  • domain assumption Nodes are distinct points on the complex unit circle that form identifiable clusters
    Required for the distance-product expression and for the distinction between univariate and multivariate geometry.

pith-pipeline@v0.9.0 · 5619 in / 1337 out tokens · 27863 ms · 2026-05-24T20:37:24.221929+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages

  1. [1]

    Akinshin, G

    A. Akinshin, G. Goldman, and Y. Yomdin. Geometry of error amplification in solving Prony system with near-colliding nodes. ArXiv e-prints, Jan. 2017

    work page 2017

  2. [2]

    Aubel and H

    C. Aubel and H. B¨ olcskei. Vandermonde matrices with nodes in the unit disk and the large sieve. Appl. Comput. Harmon. Anal. , 47(1):53–86, jul 2019

    work page 2019

  3. [3]

    Batenkov, L

    D. Batenkov, L. Demanet, G. Goldman, and Y. Yomdin. Stability of partial Fourier matrices with clustered nodes. ArXiv e-prints, Sept. 2018. 16

    work page 2018

  4. [4]

    Batenkov, G

    D. Batenkov, G. Goldman, and Y. Yomdin. Super-resolution of near-colliding point sources. arXiv e-prints, page arXiv:1904.09186, Apr 2019

    work page arXiv 1904

  5. [5]

    F. S. V. Baz´ an. Conditioning of rectangular Vandermonde matrices with nodes in the unit disk. SIAM J. Matrix Anal. Appl. , 21(2):679–693, 1999

    work page 1999

  6. [6]

    B. G. R. de Prony. Essai ´ experimental et analytique: sur les lois de la dilatabilit´ e de fluides ´ elastique et sur celles de la force expansive de la vapeur de l’alkool, a diff´ erentes temp´ eratures.Journal de l’´ ecole polytechnique, 1(22):24–76, 1795

    work page

  7. [7]

    Demanet and N

    L. Demanet and N. Nguyen. The recoverability limit for superresolution via sparsity. ArXiv e-prints, Feb. 2015

    work page 2015

  8. [8]

    Diederichs

    B. Diederichs. Well-Posedness of Sparse Frequency Estimation. arXiv e-prints , May 2019

    work page 2019

  9. [9]

    D. L. Donoho. Superresolution via sparsity constraints. SIAM J. Math. Anal. , 23(5):1309–1331, 1992

    work page 1992

  10. [10]

    R. A. Horn and C. R. Johnson. Matrix Analysis. Cambridge University Press, New York, USA, 2nd edition, 2013

    work page 2013

  11. [11]

    Hua and T

    Y. Hua and T. K. Sarkar. Matrix pencil method for estimating parameters of exponen- tially damped/undamped sinusoids in noise. IEEE Trans. Acoust. Speech Signal Process., 38(5):814–824, 1990

    work page 1990

  12. [12]

    Kunis, H

    S. Kunis, H. M. M¨ oller, T. Peter, and U. von der Ohe. Prony’s method under an almost sharp multivariate Ingham inequality. J. Fourier Anal. Appl. , 24(5):1306–1318, 2018

    work page 2018

  13. [13]

    Kunis and D

    S. Kunis and D. Nagel. On the condition number of Vandermonde matrices with pairs of nearly-colliding nodes. arXiv e-prints, page arXiv:1812.08645, Dec 2018

    work page arXiv 2018

  14. [14]

    Kunis and D

    S. Kunis and D. Potts. Stability results for scattered data interpolation by trigonometric polynomials. SIAM J. Sci. Comput. , 29:1403–1419, 2007

    work page 2007

  15. [15]

    Li and W

    W. Li and W. Liao. Stable super-resolution limit and smallest singular value of restricted Fourier matrices. ArXiv e-prints, Sept. 2017

    work page 2017

  16. [16]

    W. Li, W. Liao, and A. Fannjiang. Super-resolution limit of the ESPRIT algorithm. arXiv e-prints, page arXiv:1905.03782, May 2019

    work page arXiv 1905

  17. [17]

    Liao and A

    W. Liao and A. Fannjiang. MUSIC for single-snapshot spectral estimation: stability and super-resolution. Appl. Comput. Harmon. Anal. , 40(1):33–67, 2016

    work page 2016

  18. [18]

    Mattner and B

    L. Mattner and B. Roos. Maximal probabilities of convolution powers of discrete uniform distributions. Statist. Probab. Lett., 78(17):2992 – 2996, 2008

    work page 2008

  19. [19]

    A. Moitra. Super-resolution, extremal functions and the condition number of Vander- monde matrices. In STOC’15—Proceedings of the 2015 ACM Symposium on Theory of Computing, pages 821–830. ACM, New York, 2015

    work page 2015

  20. [20]

    V. I. Morgenshtern and E. J. Cand` es. Super-resolution of positive sources: the discrete setup. SIAM J. Imaging Sci. , 9(1):412–444, 2016. 17

    work page 2016

  21. [21]

    Potts and M

    D. Potts and M. Tasche. Error estimates for the ESPRIT algorithm. In Large truncated Toeplitz matrices, Toeplitz operators, and related topics , volume 259 of Oper. Theory Adv. Appl., pages 621–648. Birkh¨ auser/Springer, Cham, 2017

    work page 2017

  22. [22]

    Roy and T

    R. Roy and T. Kailath. ESPRIT - estimation of signal parameters via rotational invari- ance techniques. IEEE Trans. Acoust. Speech Signal Process., 37(7):984–995, 1989

    work page 1989

  23. [23]

    R. O. Schmidt. Multiple emitter location and signal parameter estimation. IEEE Trans. Antennas and Propagation, 34(3):276–280, 1986. 18

    work page 1986