Pith. sign in

REVIEW 3 major objections 4 minor 44 references

DPOT: A DeepParticle method for Computation of Optimal Transport with convergence guarantee

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper proves that a map-valued loss has a unique minimizer, and that minimizer is exactly the optimal transport map.

desk verdict A simple, genuinely new loss for Monge map estimation with a clean consistency proof; the quantitative stability bound has a fixable but real regularity gap. read the letter →

arxiv 2506.23429 v1 pith:BNDXIFEI submitted 2025-06-29 stat.ML cs.LG

Yingyuan Li , Aokun Wang , Zhongjian Wang This is my paper · ORCID
classification stat.MLcs.LG MSC 49Q2268T07
keywords optimaltransportMongemapWassersteindistancemin-minoptimizationneuralconvergenceguaranteeconditionalgenerativemodeling
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper proposes a way to learn the optimal transport (Monge) map between two continuous distributions from unpaired samples, using a loss that combines the map's own transport cost with the Wasserstein-2 distance between the pushed-forward distribution and the target. The central theoretical claim is that this loss has a unique minimizer and that minimizer is exactly the Monge map, so training never needs adversarial or input-convex architectures. It further proves that a small gap in the loss implies a small $L^2$ error to the true map, giving a quantitative certificate of accuracy. The authors demonstrate the method on synthetic maps, conditional families, inverse maps, posterior sampling, and image color transfer. A sympathetic reader would care because this turns optimal-transport-map learning into a stable min-min optimization with a direct convergence guarantee.

What carries the argument

The load-bearing object is the functional $P(T)=\lambda\,(I_\mu(T))^{1/2}+W_2(T_\sharp\mu,\nu)$ with $0<\lambda<1$, where $I_\mu(T)$ is the quadratic transport cost of $T$ and $W_2$ is the Wasserstein-2 distance. The consistency proof runs on one inequality chain: $P(T)\ge\lambda W_2(T_\sharp\mu,\mu)+W_2(T_\sharp\mu,\nu)\ge\lambda W_2(\mu,\nu)$, with equality only if $T$ is the Monge map. The quantitative proof decomposes the optimality gap $\epsilon$ into $\epsilon_1+\epsilon_2+\epsilon_3$ and bounds $\|T_\epsilon-\bar T\|_{L^2(\mu)}$ by the sum of a perturbative estimate and a potential-comparison estimate, yielding $\sqrt{\epsilon_1}+\sqrt{\epsilon_2}$.

What would settle it

Construct a family of smooth, uniformly convex densities satisfying Assumption 3.2 with $\operatorname{supp}(\nu)\subseteq\operatorname{supp}(\nu_\epsilon)$ and bounded $|\det(D^2\psi_\epsilon)|$, and compute the minimal eigenvalue of $D^2\psi_\epsilon$ as $\nu_\epsilon\to\nu$; if that eigenvalue tends to zero while the determinant bound stays finite, the asserted uniformity in Lemma 3.3 fails and the $\sqrt{\epsilon_1}+\sqrt{\epsilon_2}$ bound is not established by the given proof.

Watch

Extended reading notes

Core claim

The paper claims that a single map-valued functional, $\lambda$ times the transport cost of $T$ plus the Wasserstein-2 distance from $T_\sharp\mu$ to $\nu$, has a unique minimizer and that minimizer is exactly the Monge map between $\mu$ and $\nu$. Moreover, if a learned map $T_\epsilon$ has gap $\epsilon$ in this functional, its $L^2(\mu)$ distance to the true map is $O(\sqrt{\epsilon_1}+\sqrt{\epsilon_2})$, so a small training loss is a certificate of near-optimality. The training objective is min-min rather than adversarial, and it imposes no convexity or Lipschitz constraints on the neural network. The paper also extends the loss to conditional families of maps and to inverse maps, and validates the claims numerically.

Load-bearing premise

The quantitative error bound rests on assuming the source and target supports are smooth, uniformly convex, with densities bounded above and below, and on the asserted but not fully derived claim that the convex potentials stay uniformly strongly convex as the approximation improves; if these regularity conditions fail, only the weak convergence guarantee remains.

Editorial extensions

If this is right

  • Any standard neural architecture can be used for optimal-transport-map estimation, because the loss itself enforces push-forward correctness without Lipschitz or convexity constraints.
  • The loss gap $\epsilon$ functions as a training certificate, so reaching a small gap controls the $L^2(\mu)$ error by $\sqrt{\epsilon_1}+\sqrt{\epsilon_2}$.
  • Conditional maps $T(\cdot|\kappa)$ can be trained once and evaluated at unseen parameters $\kappa$, allowing the method to scale to families of related transport problems without retraining.
  • With added cycle-consistency terms, the same loss learns forward and inverse maps, enabling bidirectional transport in applications such as color transfer.
  • Because inference is a single forward pass with no entropic smoothing or iterative sampling, the trained map can generate large numbers of samples quickly, as demonstrated on posterior sampling.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper does not propose using the loss gap as a training diagnostic, but Theorem 3.6 makes that a direct corollary: plotting $\epsilon$ during training gives an upper-bound certificate for the map error, so one could use it for early stopping or model selection.
  • The quantitative proof's dependence on uniform strong convexity suggests the $\sqrt{\epsilon_1}+\sqrt{\epsilon_2}$ rate is tied to smooth, uniformly convex geometry; testing the method on a target with non-convex support, such as an annulus, would show whether the rate survives outside those assumptions even though weak convergence likely does.
  • The consistency argument is specific to the quadratic cost and the Wasserstein-2 triangle inequality; extending the same loss to other costs would require a new lower bound relating the map's cost to the distance between $T_\sharp\mu$ and $\mu$, so the min-min scheme should not be expected to transfer unchanged.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The manuscript proposes DPOT, a neural-network method for computing Monge maps between continuous distributions by minimizing P(T) = λ(I_µ(T))^{1/2} + W2(T♯µ, ν) over maps T, with 0 < λ < 1. Section 3 proves that the unique minimizer of P is the exact optimal transport map (Theorem 3.1), establishes weak convergence of approximately optimal maps (Theorem 3.5), and claims a quantitative L2 error bound of order √ε in terms of the optimality gap ε (Theorem 3.6). Section 3.3 gives a discrete min-min training objective, including a conditional version, and Section 4 reports experiments on synthetic OT problems, inverse maps, a compartmental epidemic model, and image color transfer.

Significance. Theorem 3.1 is a clean and correct elementary result: the term λ(I_µ(T))^{1/2} removes the trivial identity-map degeneracy of the pure Wasserstein objective while preserving the Monge solution. If Theorem 3.6 were fully established, the paper would provide a useful primal-loss-gap to map-error guarantee for neural OT without architectural constraints, which is a valuable complement to dual-potential analyses. The numerical study is broad, spanning multiple architectures and real-world tasks, and the experiments are generally consistent with the qualitative message of the theory. However, the advertised quantitative convergence guarantee is not yet fully supported: the proof of Theorem 3.6 relies on an unproved uniform strong-convexity regularity estimate, and the discretized loss in Eq. (3.21) does not match the continuous objective as written. These issues need to be resolved before the central claim can be accepted.

major comments (3)
  1. [§3.3, Eq. (3.21)] The quantitative bound (3.9) is not established as stated. The passage from (3.18) to (3.19) invokes Proposition A.5, which requires the Kantorovich potentials ψ_ε to belong to the class X(2K) with a constant K independent of ε. Assumption 3.2 and the Monge–Ampère equation give uniform bounds on det(D²ψ_ε) and interior regularity, but in dimension n > 1 a uniform lower and upper bound on det(D²ψ_ε) does not by itself imply a uniform lower bound on the smallest eigenvalue of D²ψ_ε unless a uniform upper bound on D²ψ_ε (or on its trace) is also available; that upper bound is not proved anywhere. The sentence following Lemma 3.3 and Remark 3.7 assert this uniformity by citing [17] and [40, Theorem 12.50], but they do not supply the required ε-independent constants. Consequently Lemma 3.3, the use of Lemma 3.3 in Theorem 3.5, and Theorem 3.6 all rest on an unproved regularity hypothesis. The authors should either prove the uniform strong convexity of ψ_ε under Assumption 3.2, or add it explicitly as an assumption, and should also verify that the pair (µ, ν) and the potentials satisfy the hypotheses of Proposition A.5 with a common K.
  2. [Theorem 3.6, Eq. (3.15), Remark 3.7] The discretized first term in (3.21) does not approximate I_µ(T) as defined in (3.1). The continuous functional uses I_µ(T) = ½∫|T(x) − x|²dµ(x), whose empirical counterpart is (2N)^{-1}∑_{i=1}^N |T_θ(x_i) − x_i|². As written, the first term is (2N)^{-1}∑_{i,j=1}^N |T_θ(x_i) − x_j|², which converges to ½∫∫|T(x) − x'|² dµ(x)dµ(x'), not to I_µ(T). Thus the objective actually minimized in the experiments is different from the loss analyzed in Theorems 3.1–3.6. If the double sum is a typo, it should be corrected; if not, the numerical validation does not test the theoretical objective, and the claimed empirical support for the convergence bound needs to be re-examined.
  3. The regularity of ψ*_ε on supp(ν) used in Eq. (3.15) is not justified by the stated assumptions. The proof applies the first-order convexity inequality for ψ*_ε at points y ∈ supp(ν) and bounds ∇ψ*_ε(y) by a constant R(µ) using only the fact that ∇ψ*_ε pushes ν_ε forward to µ. This requires ψ*_ε to be differentiable and its gradient to be uniformly bounded on supp(ν) (not merely ν_ε-a.e.), which is not derived from Assumption 3.2. The assumption supp(ν) ⊆ supp(ν_ε) is explicit but is an extra condition on T_ε that is not implied by a small gap ε, and Remark 3.7's proposed relaxation is only a sketch: the cited standard extension and [40, Theorem 12.50] are not shown to produce ε-independent constants. This affects the bound on the term I and hence the estimate (3.18).
minor comments (4)
  1. [Appendix A.1, Proposition A.5] The statement of Proposition A.5 is not self-contained: it defines the class M(K) but never specifies the hypotheses on the target measure Q, even though the conclusion refers to an optimal transport map from P to Q. Please either quote the full statement from [23] or state the needed assumptions explicitly.
  2. [Eq. (3.16)] The displayed inequality bounds M(ν) − M(ν_ε) from above, but the term II in the decomposition contains M(ν_ε) − M(ν). The needed upper bound follows by the same argument with the roles of ν and ν_ε swapped, but this should be stated to make the proof complete.
  3. [Section 4, Figures 3a and 3b] The text says that the L2 relative error 'correlates well' with the optimality gap ε, but no quantitative fit is reported. Since Theorem 3.6 predicts a √ε rate, a log-log plot with a fitted slope or a table of exponents would make the empirical validation more convincing.
  4. [Section 3.3] The sentence 'This optimization scheme maintains convergence while significantly lowering computational overhead' is not supported by any theoretical or experimental analysis of the alternating update of γ and θ; please soften or provide evidence for this claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the consistency proof is self-contained and the quantitative error bounds rely on external optimal-transport regularity theorems, not on the paper's own results.

full rationale

The consistency theorem (Theorem 3.1) is derived directly from the definition of the functional P(T), the triangle inequality for W2, and the uniqueness of the Monge map in Proposition 2.1 (Villani). The equality analysis in (3.2)-(3.4) shows that equality forces T to push mu onto nu and to be the optimal map between mu and T#mu, which is the standard definitional characterization of the Monge map, not an assumption of the conclusion. The quantitative bounds in Lemma 3.3 and Theorem 3.6 use external standard results: Proposition A.4 is Proposition 3.3 of Gigli [17], Proposition A.5 is Proposition 10 of Hutter-Rigollet [23], and the weak-convergence argument uses Villani [40,41]. No fitted parameter is renamed as a prediction: the epsilon decomposition (3.6) is an algebraic identity, and the constants in the bounds are tracked rather than calibrated. The self-citations to DeepParticle [42,43] appear only as the lambda=0 special case and as inspiration for the discrete loss (3.21); they are not load-bearing in any proof. The paper does contain a nontrivial regularity gap: the uniform strong convexity and uniform C^2 bounds for the Kantorovich potentials psi_epsilon are asserted via 'the same argument in [17]' and via Remark 3.7 rather than fully proved, and Proposition A.5 is invoked 'under the assumption that psi_epsilon is strongly convex as required by the proposition.' This is a correctness or rigor concern about the quantitative theorem, not circularity, because the assertion is supported by external sources [17,40] and does not presuppose the theorem being proved. Consequently, no circular step is present, and the appropriate score is 0.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The central claim rests on standard OT regularity results plus several ad hoc assumptions (C² uniformly convex supports, density bounds, support containment, uniform strong convexity of potentials) that the paper does not fully prove; the only hand-tuned free parameter is λ.

free parameters (1)
  • λ = 0.3
    Weight balancing the transport cost term and the Wasserstein-2 target term. The theory holds for any λ∈(0,1), but experiments fix λ=0.3, and Appendix B.3 notes λ=0.3 gives the best approximation among the tested values.
assumptions (5)
  • domain assumption µ and ν lie in Pac,2(Rn) with µ absolutely continuous and not charging sets of dimension n−1 (Proposition 2.1)
    Ensures existence and uniqueness of the Monge map; used throughout the proof of Theorem 3.1 and the convergence bounds.
  • ad hoc to paper Assumption 3.2: supp(µ) and supp(νϵ) are C² uniformly convex; µ and νϵ have C^{0,α} densities with uniform positive bounds.
    Invoked for Lemma 3.3 and Theorem 3.6; the paper claims these are achievable by regularity of neural-network maps without proof.
  • ad hoc to paper supp(ν)⊆supp(νϵ) in Theorem 3.6
    Needed for the first-order convexity inequality on ψ*_ϵ; the paper calls it purely technical and sketches a relaxation in Remark 3.7.
  • ad hoc to paper Uniform strong convexity of the Kantorovich potentials ψϵ with modulus independent of ϵ
    Used in Lemma 3.3 (via Proposition A.4 from [17]) and in the application of Proposition A.5 from [23]; the derivation is only cited, not shown.
  • standard math External results Proposition A.4 ([17]) and Proposition A.5 ([23]) are correct and applicable under the stated assumptions
    The paper relies on these as black-box stability and duality-gap estimates.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DPOT: A DeepParticle method for Computation of Optimal Transport with convergence guarantee." pith.science (2026). https://pith.science/paper/BNDXIFEI

@misc{pith2026250623429,
  author       = {Pith},
  title        = {Pith review of: DPOT: A DeepParticle method for Computation of Optimal Transport with convergence guarantee},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BNDXIFEI}},
  note         = {Machine review of arXiv:2506.23429}
}
read the original abstract

In this work, we propose a novel machine learning approach to compute the optimal transport map between two continuous distributions from their unpaired samples, based on the DeepParticle methods. The proposed method leads to a min-min optimization during training and does not impose any restriction on the network structure. Theoretically we establish a weak convergence guarantee and a quantitative error bound between the learned map and the optimal transport map. Our numerical experiments validate the theoretical results and the effectiveness of the new approach, particularly on real-world tasks.

Figures

Figures reproduced from arXiv: 2506.23429 by the authors.

Figure 1
Figure 1. Comparison of the transformed Cartesian mesh under [PITH_FULL_IMAGE:figures/full_fig_p011_1.png] view at source ↗
Figure 2
Figure 2. Transport map inferred by conditioned network [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. Total optimality gap ϵ and L 2 relative error vs. epochs      [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Comparison of source and prediction inferred by conditioned network [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]
Figure 5
Figure 5. Figure 5: Network predicted map for inverse transport [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]
Figure 6
Figure 6. Figure 6: Marginal density comparison of predicted posterior and ground truth (In [PITH_FULL_IMAGE:figures/full_fig_p016_6.png]
Figure 7
Figure 7. Figure 7: Comparison of color transferred results for the first task (The transferred sky [PITH_FULL_IMAGE:figures/full_fig_p018_7.png]
Figure 8
Figure 8. Figure 8: Comparison of transferred color palettes (500 random pixels) for the second [PITH_FULL_IMAGE:figures/full_fig_p018_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 39 canonical work pages

  1. [17]

    Gigli, N.: On H¨ older continuity-in-time of the optimal transport map towards measures along a curve. Proc. Edinb. Math. Soc. 54(2), 401–409 (2011)

    work page 2011

  2. [1]

    Ambrosio, L., Pratelli, A.: Existence and stability results in the L1 theory of optimal transportation, pp. 123–160. Springer, Berlin, Heidelberg (2003)

    work page 2003

  3. [2]

    Arjovsky, M., Chintala, S., Bottou, L.: Wasserstein generative adversarial net- works. In: Int. Conf. Mach. Learn., pp. 214–223. PMLR (2017)

    work page 2017

  4. [3]

    Azagra, D., Mudarra, C.: An extension theorem for convex functions of class C1,1 on Hilbert spaces. J. Math. Anal. Appl. 446(2), 1167–1182 (2017)

    work page 2017

  5. [4]

    Benamou, J., Froese, B.D., Oberman, A.M.: Numerical solution of the optimal transportation problem using the Monge–Amp` ere equation. J. Comput. Phys. 260, 107–126 (2014)

    work page 2014

  6. [5]

    John Wiley & Sons, New York (1999)

    Billingsley, P.: Convergence of Probability Measures, 2 edn. John Wiley & Sons, New York (1999)

    work page 1999

  7. [6]

    Bortoli, V.D., Thornton, J., Heng, J., Doucet, A.: Diffusion schr¨ odinger bridge with applications to score-based generative modeling. In: Adv. Neural Inform. Process. Syst., vol. 34, pp. 17695–17709 (2021)

    work page 2021

  8. [7]

    In: Proc

    Bouallegue, G., Djemal, R.: EEG data augmentation using wasserstein GAN. In: Proc. Int. Conf. Sci. Tech. Autom. Control Comput. Eng., pp. 40–45. IEEE (2020)

    work page 2020

Show all 44 references
  1. [8]

    In: Proc

    Chai, W., Jiang, Z., Hwang, J.N., Wang, G.: Global adaptation meets local gen- eralization: Unsupervised domain adaptation for 3D human pose estimation. In: Proc. IEEE/CVF Int. Conf. Comput. Vis., pp. 14655–14665. IEEE (2023)

    2023

  2. [9]

    arXiv preprint arXiv:2303.02554 (2023)

    Cui, T., Dolgov, S., Zahm, O.: Self-reinforced polynomial approximation methods for concentrated probability densities. arXiv preprint arXiv:2303.02554 (2023)

    2023 arXiv

  3. [10]

    Cuturi, M.: Sinkhorn distances: Lightspeed computation of optimal transport. In: Adv. Neural Inform. Process. Syst., vol. 26 (2013) 19

    2013

  4. [11]

    Daniels, M., Maunu, T., Hand, P.: Score-based generative neural networks for large-scale optimal transport. In: Adv. Neural Inform. Process. Syst., vol. 34, pp. 12955–12965 (2021)

    2021

  5. [12]

    E, W., Yu, B.: The Deep Ritz method: a deep learning-based numerical algorithm for solving variational problems. Commun. Math. Stat. 6(1), 1–12 (2018)

    2018

  6. [13]

    arXiv preprint arXiv:2106.03812 (2021)

    Fan, J., Liu, S., Ma, S., Chen, Y., Zhou, H.: Scalable computation of Monge maps with general costs. arXiv preprint arXiv:2106.03812 (2021)

    2021 arXiv

  7. [14]

    Fan, J., Liu, S., Ma, S., Zhou, H.M., Chen, Y.: Neural Monge map estimation and its applications. Trans. Mach. Learn. Res. (2023)

    2023

  8. [15]

    Flamary, R., Courty, N., Gramfort, A., et al.: POT: Python optimal transport. J. Mach. Learn. Res. 22(78), 1–8 (2021)

    2021

  9. [16]

    Fournier, N., Guillin, A.: On the rate of convergence in Wasserstein distance of the empirical measure. Probab. Theory Relat. Fields 162(3), 707–738 (2015)

    2015

  10. [18]

    Guminov, S., Dvurechensky, P., Tupitsa, N., Gasnikov, A.: On a combination of alternating minimization and Nesterov’s momentum. In: Int. Conf. Mach. Learn., pp. 3886–3898. PMLR (2021)

    2021

  11. [19]

    arXiv preprint arXiv:2405.14449 (2024)

    Gushchin, N., Selikhanovych, D., Kholkin, S., Burnaev, E., Korotin, A.: Adver- sarial Schr¨ odinger bridge matching. arXiv preprint arXiv:2405.14449 (2024)

    2024 arXiv

  12. [20]

    arXiv preprint arXiv:1812.02463 (2018)

    Haloui, I., Gupta, J.S., Feuillard, V.: Anomaly detection with Wasserstein GAN. arXiv preprint arXiv:1812.02463 (2018)

    2018 arXiv

  13. [21]

    In: IEEE Conf

    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: IEEE Conf. Comput. Vis. Pattern Recognit. (2016)

    2016

  14. [22]

    Huang, K.W., Chen, G.W., Huang, Z.H., Lee, S.H.: IWGAN: Anomaly detection in airport based on improved Wasserstein generative adversarial network. Appl. Sci. 13(3) (2023)

    2023

  15. [23]

    The Annals of Statistics 49(2) (2021)

    H¨ utter, J.C., Rigollet, P.: Minimax estimation of smooth optimal transport maps. The Annals of Statistics 49(2) (2021)

    2021

  16. [24]

    Kantorovich, L.K.: On a problem of Monge. J. Math. Sci. 133(4) (2006)

    2006

  17. [25]

    arXiv preprint arXiv:1909.13082 (2019)

    Korotin, A., Egiazarian, V., Asadulaev, A., Safin, A., Burnaev, E.: Wasserstein-2 generative networks. arXiv preprint arXiv:1909.13082 (2019)

    2019 arXiv

  18. [26]

    arXiv preprint arXiv:1902.02934 (2019)

    Lei, N., Guo, Y., An, D., Qi, X., Luo, Z., Yau, S.T., Gu, X.: Mode collapse and regularity of optimal transportation maps. arXiv preprint arXiv:1902.02934 (2019)

    2019 arXiv

  19. [27]

    In: International Conference in Communications, Signal Processing, and Systems, pp

    Li, B., Liu, D., Yang, J., Zhou, H., Lin, D.: An RF fingerprint data enhancement method based on WGAN. In: International Conference in Communications, Signal Processing, and Systems, pp. 539–547. Springer, Singapore (2024)

    2024

  20. [28]

    Lin, T., Ho, N., Jordan, M.I.: On the efficiency of entropic regularized algorithms for optimal transport. J. Mach. Learn. Res. 23(137), 1–42 (2022) 20

    2022

  21. [29]

    In: The Eleventh International Conference on Learning Representations (2023)

    Luo, Y., Jiang, Z., Cohen, S., Grefenstette, E., Deisenroth, M.P.: Optimal trans- port for offline imitation learning. In: The Eleventh International Conference on Learning Representations (2023)

    2023

  22. [30]

    In: Neural Information Processing, pp

    Luo, Y., Zhang, S.Y., Zheng, W.L., Lu, B.L.: WGAN domain adaptation for EEG-based emotion recognition. In: Neural Information Processing, pp. 275–286. Springer (2018)

    2018

  23. [31]

    Maggi, F.: Sets of Finite Perimeter and Geometric Variational Problems: An Introduction to Geometric Measure Theory, vol. 135. Cambridge University Press, Cambridge (2012)

    2012

  24. [32]

    Makkuva, A., Taghvaei, A., Oh, S., Lee, J.: Optimal transport mapping via input convex neural networks. In: Int. Conf. Mach. Learn., pp. 6672–6681. PMLR (2020)

    2020

  25. [33]

    Advances in mathe- matics 128(1), 153–179 (1997)

    McCann, R.J.: A convexity principle for interacting gases. Advances in mathe- matics 128(1), 153–179 (1997)

    1997

  26. [34]

    Analyst 148(16), 3860–3869 (2023)

    McHardy, R.G., Antoniou, G., Conn, J.J.A., Baker, M.J., Palmer, D.S.: Augmen- tation of FTIR spectral datasets using wasserstein generative adversarial networks for cancer liquid biopsies. Analyst 148(16), 3860–3869 (2023)

    2023

  27. [35]

    arXiv preprint arXiv:2304.06094 (2023)

    Mokrov, P., Korotin, A., Kolesov, A., Gushchin, N., Burnaev, E.: Energy-guided entropic neural optimal transport. arXiv preprint arXiv:2304.06094 (2023)

    2023 arXiv

  28. [36]

    Monge, G.: M´ emoire sur la th´ eorie des d´ eblais et des remblais. Hist. Acad. Roy. Sci. Paris pp. 666–704 (1781)

  29. [37]

    arXiv preprint arXiv:2110.02999 (2021)

    Rout, L., Korotin, A., Burnaev, E.: Generative modeling with optimal transport maps. arXiv preprint arXiv:2110.02999 (2021)

    2021 arXiv

  30. [38]

    arXiv preprint arXiv:1711.02283 (2017)

    Seguy, V., Damodaran, B.B., Flamary, R., Courty, N., Rolet, A., Blondel, M.: Large-scale optimal transport and mapping estimation. arXiv preprint arXiv:1711.02283 (2017)

    2017 arXiv

  31. [39]

    In: In- ternational Conference on Computer Research and Development, pp

    Shao, J., Chen, L., Wu, Y.: SR WGANTV: Image super-resolution through Wasser- stein generative adversarial networks with total variational regularization. In: In- ternational Conference on Computer Research and Development, pp. 21–26 (2021)

    2021

  32. [40]

    Villani, C.: Optimal Transport: Old and New, vol. 338. Springer Science & Business Media, Berlin (2009)

    2009

  33. [41]

    Villani, C.: Topics in Optimal Transportation, Graduate Studies in Mathematics , vol. 58. American Mathematical Society, Providence (2021)

    2021

  34. [42]

    Wang, Z., Xin, J., Zhang, Z.: Deepparticle: Learning invariant measure by a deep neural network minimizing Wasserstein distance on data generated from an interacting particle method. J. Comput. Phys. 464, 111309 (2022)

    2022

  35. [43]

    Phys- ica D: Nonlinear Phenomena 460, 134082 (2024)

    Wang, Z., Xin, J., Zhang, Z.: A deepparticle method for learning and generating aggregation patterns in multi-dimensional keller–segel chemotaxis systems. Phys- ica D: Nonlinear Phenomena 460, 134082 (2024)

    2024

  36. [44]

    Xie, Y., Wang, Z., Zhang, Z.: Randomized methods for computing optimal trans- port without regularization and their convergence analysis. J. Sci. Comput.100(2), 37 (2024) 21 A Supplementary results and proofs We first recall some crucial definitions to characterize convexity a...

    2024

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.