pith. sign in

arxiv: 1907.00790 · v2 · pith:ONKRQDJAnew · submitted 2019-06-28 · 🧮 math.FA · math.AG· math.PR· math.ST· stat.TH

The multidimensional truncated Moment Problem: Shape and Gaussian Mixture Reconstruction from Derivatives of Moments

Philipp J. di Dio This is my paper

Pith reviewed 2026-05-25 13:46 UTC · model grok-4.3

classification 🧮 math.FA math.AGmath.PRmath.STstat.TH
keywords truncated moment problemGaussian mixture representationderivatives of momentsmultidimensional momentspolynomial functionalsshape reconstructionmoment functionals
0
0 comments X

The pith

Some truncated moment functionals on polynomials in n variables of degree 2d require exactly binom(n+2d,n) minus n times binom(n+d,n) plus binom(n,2) Gaussians and no fewer.

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

The paper introduces the theory of derivatives of moments to represent moment functionals as Gaussian mixtures or characteristic functions of polytopes. It proves that certain functionals on the space of polynomials up to degree 2d in n variables cannot be written with fewer than the stated number of Gaussians. This lower bound approaches the full dimension of the polynomial space for large n, while an upper bound of one less than that dimension always holds. A reader cares because the result quantifies the minimal complexity of Gaussian-mixture representations recoverable from moments alone.

Core claim

There exist moment functionals L from the space of polynomials in n variables of degree at most 2d to the reals that admit a representation as a sum of binom(n+2d,n) minus n binom(n+d,n) plus binom(n,2) Gaussians but cannot be represented by any smaller number; the same functionals always admit a representation with at most binom(n+2d,n) minus 1 Gaussians. The argument proceeds by constructing explicit functionals via the newly defined derivatives of moments and then counting the linear independence constraints that each additional Gaussian can satisfy.

What carries the argument

The theory of derivatives of moments and moment functionals, which converts the problem of representing a linear functional on polynomials into the task of matching values and derivatives of a Gaussian mixture or polytope indicator.

If this is right

  • Reconstruction of a moment functional from its moments is always possible with at most one fewer Gaussian than the dimension of the polynomial space.
  • For any fixed degree d the proportion of Gaussians required approaches 1 as the number of variables grows.
  • The same derivative construction yields representations by characteristic functions of polytopes and by simple functions on polytopes.
  • The lower-bound count is achieved by explicit linear independence of the moment derivatives contributed by each Gaussian.

Where Pith is reading between the lines

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

  • The result implies that generic high-dimensional moment data cannot be explained by sparse Gaussian mixtures.
  • One could test whether the same lower-bound construction applies when the Gaussians are required to have identical covariance matrices.
  • The derivative machinery may extend to other radial basis functions beyond Gaussians.

Load-bearing premise

The derivatives-of-moments construction actually produces valid representations by the claimed numbers of Gaussians and polytopes.

What would settle it

For concrete small values such as n=3 and d=2, exhibit a moment functional whose minimal Gaussian count lies strictly outside the interval from binom(3+4,3) minus 3 binom(3+2,3) plus binom(3,2) to binom(3+4,3) minus 1.

read the original abstract

In this paper we introduce the theory of derivatives of moments and (moment) functionals to represent moment functionals by Gaussian mixtures, characteristic functions of polytopes, and simple functions of polytopes. We study, among other measures, Gaussian mixtures, their reconstruction from moments and especially the number of Gaussians needed to represent moment functionals. We find that there are moment functionals $L:\mathbb{R}[x_1,\dots,x_n]_{\leq 2d}\to\mathbb{R}$ which can be represented by a sum of $\binom{n+2d}{n} - n\cdot \binom{n+d}{n} + \binom{n}{2}$ Gaussians but not less. Hence, for any $d\in\mathbb{N}$ and $\varepsilon>0$ we find an $n\in\mathbb{N}$ such that $L$ can be represented by a sum of $(1-\varepsilon)\cdot\binom{n+2d}{n}$ Gaussians but not less. An upper bound is $\binom{n+2d}{n}-1$.

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

1 major / 2 minor

Summary. The paper introduces the theory of derivatives of moments and moment functionals to represent moment functionals by Gaussian mixtures, characteristic functions of polytopes, and simple functions of polytopes. It studies the reconstruction from moments and the number of Gaussians needed, finding that there exist moment functionals L on R[x1,...,xn]≤2d that can be represented by a sum of binom(n+2d,n) - n binom(n+d,n) + binom(n,2) Gaussians but not fewer, with an upper bound of binom(n+2d,n)-1 and asymptotic density (1-ε) for large n.

Significance. If the constructions using the new derivative operators are valid and the independence of the resulting conditions is established, this provides sharp lower bounds on the number of Gaussians in mixture representations of moment functionals. This advances the multidimensional truncated moment problem by quantifying the minimal complexity of such representations.

major comments (1)
  1. [The section introducing the theory of derivatives of moments] The lower bound calculation subtracts n·binom(n+d,n) for the means and binom(n,2) for a quadratic correction. The manuscript must explicitly show that the derivative operators on moment functionals yield binom(n,2) linearly independent conditions that are not implied by the ordinary moment conditions; otherwise the codimension argument does not hold. This is central to the main theorem.
minor comments (2)
  1. Clarify the notation for the space of polynomials R[x1,...,xn]≤2d in the abstract and introduction.
  2. Provide a small-dimensional example (e.g., n=2, d=1) to illustrate the derivative operators and the resulting bound.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the careful reading of our manuscript and for the positive evaluation of its significance in advancing the multidimensional truncated moment problem. We address the major comment point by point below.

read point-by-point responses

  1. Referee: [The section introducing the theory of derivatives of moments] The lower bound calculation subtracts n·binom(n+d,n) for the means and binom(n,2) for a quadratic correction. The manuscript must explicitly show that the derivative operators on moment functionals yield binom(n,2) linearly independent conditions that are not implied by the ordinary moment conditions; otherwise the codimension argument does not hold. This is central to the main theorem.

    Authors: We agree that establishing the linear independence of the binom(n,2) conditions arising from the quadratic correction is essential for the codimension argument in the lower bound. In the section on derivatives of moments, the operators are introduced and the resulting conditions are derived from the representation by Gaussian mixtures. The independence from ordinary moment conditions is addressed by showing that these correspond to distinct second-order derivative functionals on the space of polynomials of degree at most 2d. However, we acknowledge that a more explicit verification of the rank (i.e., that these binom(n,2) conditions are not implied by the lower-order ones) would strengthen the presentation. In the revised manuscript we will add a dedicated lemma in that section, including an explicit basis for the dual space or a rank computation of the associated linear map, to make this independence fully rigorous and self-contained. revision: yes

Circularity Check

0 steps flagged

No circularity: bounds derived from newly introduced derivative theory without reduction to inputs by construction.

full rationale

The paper introduces the theory of derivatives of moments as a new framework and states the Gaussian-mixture lower bounds as a consequence of studying representations within that framework. The dimensional subtraction yielding binom(n+2d,n) - n·binom(n+d,n) + binom(n,2) is presented as a finding about the codimension achieved by the derivative operators, not as a definition or a fit to data that is then relabeled as a prediction. No self-citation chain, ansatz smuggling, or self-definitional loop is exhibited in the abstract or the described claims; the result is not forced by renaming a known pattern or by invoking an unverified uniqueness theorem from the same author. The derivation therefore remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 1 invented entities

Based solely on the abstract; no explicit free parameters, axioms, or invented entities are detailed enough to enumerate. The new concept of 'derivatives of moments' functions as an invented tool whose independent justification is not provided here.

invented entities (1)
  • derivatives of moments no independent evidence
    purpose: to represent moment functionals by Gaussian mixtures and polytopes
    Introduced in the abstract as the central new device; no independent evidence or definition supplied.

pith-pipeline@v0.9.0 · 5717 in / 1349 out tokens · 27000 ms · 2026-05-25T13:46:42.654049+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

14 extracted references · 14 canonical work pages · 2 internal anchors

  1. [1]

    Am´ endola, J.-C

    [AFS16] C. Am´ endola, J.-C. Faug` ere, and B. Sturmfels, Moment varieties of gaussian mix- tures, J. Alg. Stat. 7 (2016), 14–28. [Akh65] N. I. Akhiezer, The classical moment problem and some related questions in a nalysis, Oliver & Boyd, Edinburgh,

    work page 2016

  2. [2]

    [Ana06] G. A. Anastassiou, Applications of geometric moment theory related to optimal port- folio management , Comput. Math. Appl. 51 (2006), 1405–1430. [APST19] H. Ammari, M. Putinar, A. Streenkamp, and F. Triki, Identification of an algebraic domain in two dimensions from a finite number of its generaliz ed polarization ten- sors, Math. Ann. (2018/19), in...

    work page 2006

  3. [3]

    [dD19] P. J. di Dio, The multidimensional truncated Moment Problem: Gaussian a nd Log- Normal Mixtures, their Carath´ eodory Numbers, and Set of At oms, Proc. Amer. Math. Soc. 147 (2019), 3021–3038, arXiv:1804.07058. [dDK19] P. J. di Dio and M. Kummer, The multidimensional truncated moment problem: Carath´ eodory Numbers from Hilbert Functions, https://ar...

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  4. [4]

    [Har99] W. R. Harris, Real Even Symmetric Ternary Forms , J. Alg. 222 (1999), 204–245. [HK14] D. Henrion and M. Korda, Convex computation of the region of attraction of poly- nomial control systems , IEEE Trans. Aut. Control 59 (2014), 297–312. [Hu62] M.-K. Hu, Visual pattern recognition by moment invariants , IRE Trans. Inf. Theory 12 (1962), 179–187. [K...

    work page 1999

  5. [5]

    [KSS18] K. Kohn, B. Shapiro, and B. Sturmfels, Moment varieties of measures on polytopes , arXiv:1807.10258v1. [Kun14] A. Kunert, Facial Structure of Cones of non-negative Forms , Ph.D. thesis, Univer- stity of Konstanz,

    work page arXiv

  6. [6]

    Laurent, Sums of Squares, Moment Matrices and Polynomial over Optimi zation, Emerging application of algebraic geometry, IMA Vol

    [Lau09] M. Laurent, Sums of Squares, Moment Matrices and Polynomial over Optimi zation, Emerging application of algebraic geometry, IMA Vol. Math. Appl., vol. 149, Springer, New York, 2009, pp. 157–270. [Law91] J. Lawrence, Polytope volume computation , Math. Comput. 57 (1991), 259–271. [LPHT08] J.-B. Lasserre, C. Prieur, D. Henrion, and E. Tr´ el at, Non...

    work page 2009

  7. [7]

    [LR82] Y. T. Lee and A. A. G. Requicha, Algorithms for computing the volume and other integral properties of solids. I. known methods and open iss ues, Comm. ACM 25 (1982), 635–641. [Mar08] M. Marshall, Positive Polynomials and Sums of Squares , Mathematical Surveys and Monographs, no. 146, American Mathematical Society, Rhode Island,

    work page 1982

  8. [8]

    Martin, K

    [MMR05] J.-M. Martin, K. Mengersen, and C. P. Robert, Bayesian modelling and inference on mixtures of distributions , Handbook of Statistics 25 (2005), 459–507. [MN68] M. Maˇ nas and J. Nedoma, Finding all vertices of a convex polyhedron , Numer. Math. 12 (1968), 226–229. [Mot67] T. S. Motzkin, The arithmetic-geometric inequality , Inequalities (New York)...

    work page 2005

  9. [9]

    ¨Uber den Approximationssatz von W eierstrass, pp

    August 1914 gewidmet von Freunden und Sch¨ ulern., ch. ¨Uber den Approximationssatz von W eierstrass, pp. 303–312, Springer, Berlin,

    work page 1914

  10. [10]

    A moment approach for entropy solutions to nonlinear hyperbolic PDEs

    [MVKW95] P. Milanfar, G. Verghese, W. Karl, and A. Willsky, Reconstructing polygons from moments with connections to array processing , IEEE Trans. Signal Proc. 43 (1995), 432–443. [MWHL18] S. Marx, T. W eisser, D. Henrion, and J.-B. Lasserre , A moment approach for entropy solutions to nonlinear hyperbolic PDEs , arXiv:1807.02306v1. [Pea94] K. Pearson, C...

    work page internal anchor Pith review Pith/arXiv arXiv 1995

  11. [11]

    [PK92] A. V. Pukhlikov and A. G. Khovanskii, The Riemann–Roch theorem for integrals and sums of quasipolynomials on virtual polytopes , Algebra Anal. 4 (1992), 188–216. [Ric57] H. Richter, Parameterfreie Absch¨ atzung und Realisierung von Erwartungswerten, Bl. Deutsch. Ges. Versicherungsmath. 3 (1957), 147–161. [Rob69] R. M. Robinson, Some definite polynom...

    work page 1992

  12. [12]

    [Rog58] W. W. Rogosinski, Moments of non-negative mass , Proc. R. Soc. Lond. A 245 (1958), 1–27. [Ros52] P. C. Rosenbloom, Quelques classes de probl` eme extr´ emaux. II , Bull. Soc. Math. France 80 (1952), 183–215. [RS18] C. Riener and M. Schweighofer, Optimization approaches to quadrature: new char- acterizations of Gaussian quadrature on the line and q...

    work page 1958

  13. [13]

    Sommer, O

    [SMD+07] I. Sommer, O. M¨ uller, F. S. Domingues, O. Sander, J. W eic kert, and T. Lengauer, Moment invariants as shape recognition technique for compa ring protein binding sites, Bioinformatics 23 (2007), 3139–3146. [ST43] J. A. Shohat and J. D. Tamarkin, The Problem of Moments , Amer. Math. Soc., Providence, R.I.,

    work page 2007

  14. [14]

    [Sti94] T. J. Stieltjes, Recherches sur les fractions continues , Ann. Fac. Sci. Toulouse 8 (1894), no. 4, J1–J122. 28 DERIV ATIVES OF MOMENTS: SHAPE AND GAUSSIAN MIXTURE RECON STRUCTION [Sto16] J. Stoyanov, Moment properties of probability distributions used in sto chastic finan- cial models, Recent Advances in Financial Engineering 2014 Proceeding s of t...

    work page 2014