REVIEW 3 minor 38 references
Rigidity and Quantitative Stability of the Sliced Wasserstein Deficit
T0 review · 0 major / 3 minor · reviewed 2026-06-29 · grok-4.3
Pith's one-line read The sliced Wasserstein deficit is zero exactly when the Brenier map between the measures is homothetic affine.
desk verdict The paper gives a precise rigidity result for when the sliced Wasserstein deficit vanishes and introduces a new SPK constant for stability, which looks like a genuine addition worth checking in detail. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The sliced Wasserstein deficit D(μ,ν) = (1/d)W₂²(μ,ν) - SW₂²(μ,ν), whose vanishing is equivalent to the Brenier map being homothetic affine.
What would settle it
A pair of measures in dimension two or higher, with the source absolutely continuous, whose Brenier map is known to be nonlinear, yet whose sliced Wasserstein deficit is exactly zero, would disprove the rigidity statement.
Extended reading notes
Core claim
We prove that D(μ,ν)=0 if and only if the Brenier map T=∇φ from μ to ν is homothetic affine, T(x)=λx+b μ-a.e., for some λ≥0 and b∈R^d. For quantitative stability, we introduce the sliced Poincaré-Korn constant κ_SPK(μ) as a spectral gap of an averaged ridge-projection quadratic form on gradient fields modulo the family {λx+b}. Whenever this constant is positive, we prove a stability estimate for the sliced Wasserstein deficit, up to a one-dimensional Lipschitz scale for the projected monotone transports. We obtain the sharp SPK constant for the Gaussian measures, and establish positive SPK bounds for bounded perturbations of the Gaussian and compact classes of gradient fields for fixed sourc
Load-bearing premise
The source measure must be absolutely continuous with respect to Lebesgue measure so that a unique Brenier map exists.
Editorial extensions
If this is right
- The inequality SW₂²(μ,ν) ≤ (1/d)W₂²(μ,ν) becomes equality precisely when the transport map is homothetic affine.
- A positive sliced Poincaré-Korn constant yields a quantitative bound relating the size of the deficit to the distance of the Brenier map from the affine class, up to one-dimensional Lipschitz factors.
- The sliced Poincaré-Korn constant is positive and sharp for centered Gaussians and remains positive under small bounded perturbations of the Gaussian.
- Anisotropic Gaussians demonstrate that neither Bakry-Émery curvature bounds nor standard Poincaré inequalities are sufficient to guarantee a positive sliced Poincaré-Korn constant.
Reading between the lines
- The rigidity result may be used to characterize when high-dimensional measures differ only by an affine transformation through projection-based distances.
- Numerical approximation of the sliced Poincaré-Korn constant for a given source measure could provide practical error controls when replacing Wasserstein distances by their sliced versions in computations.
- The obstruction examples suggest that any attempt to derive sliced Poincaré-Korn inequalities from curvature must incorporate dimension-dependent or anisotropy-dependent corrections.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper proves that the sliced Wasserstein deficit D(μ,ν) := (1/d)W₂²(μ,ν) − SW₂²(μ,ν) vanishes if and only if the Brenier map T = ∇φ from μ to ν is homothetic affine (T(x) = λx + b μ-a.e.) for μ ≪ ℒ^d in dimension d ≥ 2. It introduces the sliced Poincaré–Korn constant κ_SPK(μ) as a spectral gap on gradient fields modulo affine homotheties and derives quantitative stability estimates whenever this constant is positive. Explicit sharp values are obtained for Gaussian measures, positive bounds are established for bounded perturbations of Gaussians and compact classes of gradient fields, and anisotropic Gaussians are shown to obstruct both Bakry–Émery curvature and standard Poincaré inequalities from implying a global SPK inequality.
Significance. The rigidity characterization clarifies the equality case in the elementary comparison between sliced and full Wasserstein distances, which underpins the use of sliced Wasserstein distances as computationally tractable proxies in statistics and machine learning. The SPK constant supplies a new, explicitly computable stability modulus with sharp Gaussian examples; the obstruction result for anisotropic Gaussians is a useful negative result that delineates the reach of curvature-based methods.
minor comments (3)
- Abstract, line 3: 'an new spectral gap' should read 'a new spectral gap'.
- The stability statement is phrased 'up to a one-dimensional Lipschitz scale for the projected monotone transports'; a precise statement of the dependence on this scale (or a reference to the relevant proposition) would improve readability.
- The paper invokes μ ≪ ℒ^d to guarantee uniqueness of the Brenier map; a brief reminder of the classical theorem (e.g., Brenier or McCann) in the setup section would help readers who are not optimal-transport specialists.
Simulated Author's Rebuttal
We thank the referee for their accurate summary of the paper, the positive significance assessment, and the recommendation for minor revision. No specific major comments were provided in the report.
Circularity Check
No significant circularity identified
full rationale
The paper establishes a rigidity theorem stating that the sliced Wasserstein deficit D(μ,ν) vanishes if and only if the Brenier map is homothetic affine, under the standard assumption μ ≪ ℒ^d that guarantees existence and uniqueness of the map via classical optimal transport theory. The SPK constant is introduced as a new spectral gap quantity on gradient fields modulo affine homotheties and used to derive stability estimates when positive, with explicit computations for Gaussians. No load-bearing steps reduce by construction to fitted inputs, self-definitions, or unverified self-citations; the derivation chain relies on standard properties of Wasserstein distances and Brenier maps without circular reduction to the paper's own inputs.
Assumptions & free parameters
assumptions (2)
- standard math Existence and uniqueness of the Brenier map T=∇φ when the source measure is absolutely continuous with respect to Lebesgue measure
- standard math The inequality SW_2^2(μ,ν) ≤ (1/d) W_2^2(μ,ν) that defines the non-negative deficit
invented entities (1)
-
sliced Poincaré-Korn (SPK) constant κ_SPK(μ)
Cite this review
Pith. "Pith review of Rigidity and Quantitative Stability of the Sliced Wasserstein Deficit." pith.science (2026). https://pith.science/paper/SPVLDZAO
@misc{pith2026260525453,
author = {Pith},
title = {Pith review of: Rigidity and Quantitative Stability of the Sliced Wasserstein Deficit},
year = {2026},
howpublished = {\url{https://pith.science/paper/SPVLDZAO}},
note = {Machine review of arXiv:2605.25453}
}
abstract
The sliced Wasserstein distance $SW_2(\mu,\nu)$ compares high-dimensional probability measures by averaging one-dimensional optimal transport distances over linear projections. Although sliced Wasserstein distances are now standard computational tools in statistics, imaging, and machine learning, the rigidity behind the elementary comparison \[ SW_2^2(\mu,\nu)\leq \frac1d W_2^2(\mu,\nu) \] has not been systematically studied. Let $\mu,\nu\in\mathcal P_2(\mathbb R^d)$, $d\ge2$, with $\mu\ll\mathcal L^d$, and define the sliced Wasserstein deficit by \[ {\mathrm D}(\mu,\nu):=\frac1d W_2^2(\mu,\nu)-SW_2^2(\mu,\nu)\geq 0. \] We prove that ${\mathrm D}(\mu,\nu)=0$ if and only if the Brenier map $T=\nabla\varphi$ from $\mu$ to $\nu$ is homothetic affine, \[ T(x)=\lambda x+b \qquad \mu\text{-a.e.}, \] for some $\lambda\ge0$ and $b\in \mathbb R^d$. For quantitative stability, we introduce the sliced Poincar\'e--Korn (SPK) constant $\kappa_{\mathrm{SPK}}(\mu)$, defined as an new spectral gap of an averaged ridge-projection quadratic form on gradient fields modulo the family $\{\lambda x+b\}$. Whenever this constant is positive, we prove a stability estimate for the sliced Wasserstein deficit, up to a one-dimensional Lipschitz scale for the projected monotone transports. We obtain the sharp SPK constant for the Gaussian measures as the most important example, and establish positive SPK bounds for bounded perturbations of the Gaussian and compact classes of gradient fields for fixed source measures. Finally, we show that anisotropic Gaussians give a sharp obstruction: neither a Bakry--\'Emery lower curvature bound nor a usual Poincar\'e inequality alone can imply a global sliced Poincar\'e--Korn inequality.
Reference graph
Works this paper leans on
-
[1]
Bayraktar and G
E. Bayraktar and G. Guo, Strong equivalence between metrics of Wasserstein type,Electron. Commun. Probab.26(2021), Paper No. 13, 1–13. 3
2021
-
[2]
Bonneel, J
N. Bonneel, J. Rabin, G. Peyr´ e, and H. Pfister, Sliced and Radon Wasserstein barycenters of measures,J. Math. Imaging Vision51(2015), no. 1, 22–45. 2, 3
2015
-
[3]
Bonnotte,Unidimensional and Evolution Methods for Optimal Transportation, Ph.D
N. Bonnotte,Unidimensional and Evolution Methods for Optimal Transportation, Ph.D. thesis, Universit´ e Paris-Sud and Scuola Normale Superiore, 2013. 3
2013
-
[4]
L. A. Caffarelli, Monotonicity properties of optimal transportation and the FKG and related inequalities,Comm. Math. Phys.214(2000), no. 3, 547–563. 14
2000
-
[5]
Carrapatoso, J
K. Carrapatoso, J. Dolbeault, F. H´ erau, S. Mischler, Cl. Mouhot, Weighted Korn and Poincar´ e–Korn inequalities in the Euclidean space and associated operators,Arch. Ration. Mech. Anal.243(2022), no. 3, 1565–1596. 3, 11
2022
-
[6]
G. Carlier, A. Figalli, Q. M´ erigot, and Y. Wang, Sharp comparisons between sliced and standard 1-Wasserstein distances, arXiv preprint arXiv:2510.16465, 2025. 3
-
[7]
Cavalletti and A
F. Cavalletti and A. Mondino, Sharp and rigid isoperimetric inequalities in metric-measure spaces with lower Ricci curvature bounds,Invent. Math.208(2017), no. 3, 803–849. 4
2017
-
[8]
P. G. Ciarlet,Mathematical Elasticity. Vol. I. Three-Dimensional Elasticity, Studies in Mathematics and its Applications, vol. 20, North-Holland, Amsterdam, 1988. 11
1988
Show all 38 references
-
[9]
K. J. Ciosmak, Leaves decompositions in Euclidean spaces,J. Math. Pures Appl. (9)154 (2021), 212–244. 6
2021
-
[10]
T. A. Courtade, M. Fathi, Stability of the Poincar´ e–Korn inequality,Arch. Ration. Mech. Anal.249(2025), no. 5, Paper No. 55, 17 pp. 3, 11
2025
-
[11]
T. A. Courtade, M. Fathi, and A. Pananjady, Quantitative stability of the entropy power inequality,IEEE Trans. Inform. Theory64(2018), no. 8, 5691–5703. 4
2018
-
[12]
Cozzi and F
G. Cozzi and F. Santambrogio, Long-time asymptotics of the sliced-Wasserstein flow,SIAM J. Imaging Sci.18(2025), no. 1, 1–19. 3
2025
-
[13]
Dacorogna,Direct Methods in the Calculus of Variations, 2nd ed., Applied Mathematical Sciences, vol
B. Dacorogna,Direct Methods in the Calculus of Variations, 2nd ed., Applied Mathematical Sciences, vol. 78, Springer, New York, 2008. 11
2008
-
[14]
Delalande and Q
A. Delalande and Q. M´ erigot, Quantitative stability of optimal transport maps under variations of the target measure,Duke Math. J.172(2023), no. 17, 3321–3357. 4
2023
-
[15]
Eldan, Thin shell implies spectral gap up to polylog via a stochastic localization scheme, Geom
R. Eldan, Thin shell implies spectral gap up to polylog via a stochastic localization scheme, Geom. Funct. Anal.23(2013), no. 2, 532–569. 6
2013
-
[16]
Fathi, E
M. Fathi, E. Indrei, and M. Ledoux, Quantitative logarithmic Sobolev inequalities and stability estimates,Discrete Contin. Dyn. Syst.36(2016), no. 12, 6835–6853. 4
2016
-
[17]
Fathi, N
M. Fathi, N. Gozlan, and M. Prod’homme, A proof of the Caffarelli contraction theorem via entropic regularization,Calc. Var. Partial Differential Equations59(2020), no. 3, Paper No. 96, 23 pp. 14
2020
-
[18]
Figalli, F
A. Figalli, F. Maggi, and A. Pratelli, A mass transportation approach to quantitative isoperimetric inequalities,Invent. Math.182(2010), no. 1, 167–211. 4 27
2010
-
[19]
Friesecke, R
G. Friesecke, R. D. James, and S. M¨ uller, A theorem on geometric rigidity and the derivation of nonlinear plate theory from three-dimensional elasticity,Comm. Pure Appl. Math.55 (2002), no. 11, 1461–1506. 11
2002
-
[20]
Higuchi, Symmetric tensor spherical harmonics on the N-sphere and their application to the de Sitter groupSO(N,1),J
A. Higuchi, Symmetric tensor spherical harmonics on the N-sphere and their application to the de Sitter groupSO(N,1),J. Math. Phys.28(1987), no. 7, 1553–1566. 23
1987
-
[21]
A. T. James and A. G. Constantine, Generalized Jacobi polynomials as spherical functions of the Grassmann manifold,Proc. London Math. Soc. (3)29(1974), 174–192. 24, 25
1974
-
[22]
Janson,Gaussian Hilbert Spaces, Cambridge Tracts in Mathematics, vol
S. Janson,Gaussian Hilbert Spaces, Cambridge Tracts in Mathematics, vol. 129, Cambridge University Press, Cambridge, 1997. 15
1997
-
[23]
Kannan, L
R. Kannan, L. Lov´ asz, and M. Simonovits, Isoperimetric problems for convex bodies and a localization lemma,Discrete Comput. Geom.13(1995), no. 3–4, 541–559. 6
1995
-
[24]
Kitagawa and A
J. Kitagawa and A. Takatsu, Sliced optimal transport: is it a suitable replacement?,Indiana Univ. Math. J.(2025), to appear; arXiv:2311.15874. 3
2025
-
[25]
Klartag and J
B. Klartag and J. Lehec, Affirmative resolution of Bourgain’s slicing problem using Guan’s bound,Geom. Funct. Anal.35(2025), 1147–1168. 6
2025
-
[26]
Kolouri, S
S. Kolouri, S. R. Park, M. Thorpe, D. Slepˇ cev, and G. K. Rohde, Optimal mass transport: signal processing and machine-learning applications,IEEE Signal Process. Mag.34(2017), no. 4, 43–59. 3
2017
-
[27]
Kolouri, K
S. Kolouri, K. Nadjahi, U. S ¸im¸ sekli, R. Badeau, and G. K. Rohde, Generalized sliced Wasserstein distances, inAdvances in Neural Information Processing Systems 32, Curran Associates, Red Hook, NY, 2019. 3
2019
-
[28]
T. Le, M. Yamada, K. Fukumizu, and M. Cuturi, Tree-sliced variants of Wasserstein distances, inAdvances in Neural Information Processing Systems 32, Curran Associates, Red Hook, NY, 2019. 3
2019
-
[29]
Y. T. Lee and S. Vempala, Eldan’s stochastic localization and the KLS conjecture: isoperime- try, concentration and mixing,Ann. of Math. (2)199(2024), no. 3, 1043–1092. 6
2024
-
[30]
M´ erigot, A
Q. M´ erigot, A. Delalande, and F. Chazal, Quantitative stability of optimal transport maps and linearization of the 2-Wasserstein space, inProceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics, Proc. Mach. Learn. Res., vol. 108, PM...
2020
-
[31]
Nadjahi, A
K. Nadjahi, A. Durmus, L. Chizat, S. Kolouri, S. Shahrampour, and U. S ¸im¸ sekli, Statistical and topological properties of sliced probability divergences, inAdvances in Neural Infor- mation Processing Systems 33, Curran Associates, Red Hook, NY, 2020, pp. 20802–20812. 3
2020
-
[32]
Nadjahi,Sliced-Wasserstein Distance for Large-Scale Machine Learning: Theory, Method- ology and Extensions, Ph.D
K. Nadjahi,Sliced-Wasserstein Distance for Large-Scale Machine Learning: Theory, Method- ology and Extensions, Ph.D. thesis, Institut Polytechnique de Paris, 2021. 3
2021
-
[33]
Nguyen, N
K. Nguyen, N. Ho, T. Pham, and H. Bui, Distributional sliced-Wasserstein and applications to generative modeling, inInternational Conference on Learning Representations, 2021. 3
2021
-
[34]
Nualart,The Malliavin Calculus and Related Topics, 2nd ed., Probability and Its Applications, Springer-Verlag, Berlin, 2006
D. Nualart,The Malliavin Calculus and Related Topics, 2nd ed., Probability and Its Applications, Springer-Verlag, Berlin, 2006. 15
2006
-
[35]
Park and D
S. Park and D. Slepˇ cev, Geometry and analytic properties of the sliced Wasserstein space, J. Funct. Anal.289(2025), no. 7, Paper No. 110975. 3 28
2025
-
[36]
Rabin, G
J. Rabin, G. Peyr´ e, J. Delon, and M. Bernot, Wasserstein barycenter and its application to texture mixing, inScale Space and Variational Methods in Computer Vision, Lecture Notes in Comput. Sci., vol. 6667, Springer, Berlin, 2012, pp. 435–446. 3
2012
-
[37]
Santambrogio,Optimal Transport for Applied Mathematicians: Calculus of Variations, PDEs, and Modeling, Progress in Nonlinear Differential Equations and Their Applications, vol
F. Santambrogio,Optimal Transport for Applied Mathematicians: Calculus of Variations, PDEs, and Modeling, Progress in Nonlinear Differential Equations and Their Applications, vol. 87, Birkh¨ auser/Springer, Cham, 2015. 2
2015
-
[38]
Villani,Optimal Transport: Old and New, Grundlehren der mathematischen Wis- senschaften, vol
C. Villani,Optimal Transport: Old and New, Grundlehren der mathematischen Wis- senschaften, vol. 338, Springer, Berlin, 2009. 2 Declaration.The author declare that he have no conflict of interest and that the manuscript has no associated data. During the preparation of this ma...
2009
Reviewed June 29, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.