- Rating
- 2(+1)
- Confidence
- 2
- Informativeness
- 1
- Presentation
- 2
- Soundness
- 3
Read review
Paper summary
This paper proposes an extension of an existing continuous entropy-regularized optimal transport solver to certain set of Riemannian manifolds by choosing a parametrization of the corresponding Gibbs kernel in such a way that the loss introduced by [1] remains tractable. Specifically the authors consider parametrizations for the following group manifolds (Proposition 1 and following): Sphere, Torus, SO(3) and SE(3) and show that their parametrization implies tractable loss also for arbitrary products of these manifolds (Proposition 2). Further the authors provide numerical experiments on synthetic as well as two real world datasets.
Critical strengths
The paper notes that for some particular choices of Gibbs kernel and manifolds the loss proposed in [1] remains tractable.
Critical weaknesses
The reviewer leans towards that the paper in its current form should be rejected, because of the following core issues: (1) Limitation in choice and lack of motivation in OT cost functions that can be used with their method, (2) lack of adequate experimental metrics that measure the quality of the proposed solution. In the following we describe them in more detail:
- Main problem: In usual entropy-regularized OT the kernel stems from a single cost function via . In this case the parameter has a clear interpretation as a regularizer of the problem. However, the method proposed here takes as its design object of choice the Gibbs kernel itself, which then induces the cost . In particular, for the proposed kernels of the authors only the kernel for the sphere and torus admit such a representation with a fixed cost (independent of ). For the case of sphere the induced cost is and for the torus we have . However, it is not clear why these cost functions would be interesting for these manifolds in the first place. For instance, the cost on the sphere is up to constants equivalent to , the ambient space cost, which seems to defeat the purpose of using Riemannian OT at all. Furthermore, for the manifolds and the induced cost even depends on , which makes the interpretation of as a regularization parameter to some fixed OT-problem inapplicable. In this case it is not clear what problem the authors are solving.
- Lack of metrics: While claiming to address the entropy-regularized OT problem, the authors seem only to evaluate their method by using a metric on the transported marginal of their computed transport plan and (see e.g. synthetic experiments). The reviewer found no reported expected transport costs, EOT objective values relative to e.g. a reference optimum or errors of their computed plan against a known optimal coupling. In this sense their reported metrics would be optimal already for the trivial coupling , which defeats the purpose of solving a EOT problem.
[1] Alexander Korotin, Nikita Gushchin, and Evgeny Burnaev. Light Schrodinger bridge. In The Twelfth International Conference on Learning Representations, 2024.
Questions for authors
The authors should spend more effort in reasoning why these choices of costs matter in applications and what problem they are solving: generative modeling (transporting to ) or EOT (transporting to while minimizing the objective) and add corresponding metrics and adequate comparisons with other methods solving the problem with the same cost and entropy parameters.
5 comments
Thank you for your reply and clarifying the geometric properties and the dependence of on for the considered costs.
My main concern is why these particular costs define useful transport problems for the applications considered. It seems that they were chosen mainly because of their computational tractability and convenience in this work and its not clear why they are relevant (note that some geometric justification for these costs does not imply they are relevant).
As mentioned before, the application metrics (e.g. in Continental drift and Crystallographic texture) only assess the matching of the (marginal) target distribution (so do not assess the cost relevance). While the new synthetic benchmark in Section 3.4 does measure (some) quality of the conditional distribution (and thus that of the coupling), it validates recovery under the selected costs themselves (so it again can't serve as a justification for the choice of these costs).
I am happy to increase my score if the authors provide a concrete modeling justification for these costs or test whether their coupling preserves or predicts a relationship that matters in the concrete application.
We thank the reviewer for the follow-up response and the specific suggestions. To address this concern, we constructed an alternative Continental Drift experiment. In this setup, we retain the same source distribution at 150 million years ago (150 Ma), but obtain the target distribution by reconstructing the source locations to the present day (0 Ma) using the plate tectonic model of Müller et al. (2019). This construction provides paired geographic locations at 150 Ma and 0 Ma, allowing us to evaluate the learned transport coupling against the reconstructed correspondences. Importantly, the methods are trained using only unpaired source and target samples, while the correspondences are used only for evaluation on unseen testing data. We trained ManifoldLightOT, ERNOT with geodesic and kernel-infered cost.
The computed metrics are the following:
- Spherical MMD²: Evaluates marginal distribution matching.
- Expected Geodesic Error (°): Measures correspondence accuracy by averaging the geodesic distances between samples from the learned conditional distributions and their corresponding reconstructed destinations.
- Energy Score: Evaluates conditional prediction accuracy and dispersion as , averaged over source points, where are independent samples from .
For each source point, we generate 128 samples from the learned conditional distribution to estimate the expected geodesic error and energy score. Using multiple conditional samples allows us to evaluate the stochastic coupling rather than relying on a single prediction, providing more reliable estimates of correspondence accuracy and predictive dispersion. Results can be found in the table below:
| Method | MMD² (×10⁻²) ↓ | Expected Geodesic Error (°) ↓ | Energy Score ↓ |
|---|---|---|---|
| ManifoldLightOT | 0.070 ± 0.005 | 14.854 ± 0.032 | 0.162 ± 0.001 |
| ERNOT (geodesic) | 0.174 ± 0.047 | 21.585 ± 0.235 | 0.166 ± 0.002 |
| ERNOT (kernel-induced) | 0.271 ± 0.013 | 28.999 ± 0.094 | 0.190 ± 0.002 |
| Identity | 6.564 ± 0.000 | 30.446 ± 0.000 | 0.521 ± 0.000 |
These results demonstrate that ManifoldLightOT and its kernel-induced costs can learn meaningful couplings. A similar result is expected for applications on different manifolds, as each of our kernel-induced costs is closely related to the corresponding geodesic distance (as we explained in our previous answer and in the updated manuscript). We hope our answer addresses the reviewer's concerns in this regard.
Müller et al. (2019), A Global Plate Model Including Lithospheric Deformation Along Major Rifts and Orogens Since the Triassic, published in Tectonics.
Answer [1/1]
We thank the reviewer for the comments and observations about our work, below we provide detailed answers for the raised concerns.
[W1] About geometric interpretations and interpretation of ε.
We thank the reviewer for the observations, below we provide detailed answers for them. For clarity, we split the first Weakness into two.
[W1.1] Geometric interpretation of the induced costs.
We would like to clarify that the proposed kernel-induced costs are closely related to geodesic distances on the corresponding manifolds, although they are not generally identical to them.
For the sphere, c(x,y)=−⟨x,y⟩=−cosdM(x,y) is an increasing function of the geodesic distance and satisfies c(x,y)=−1+21dM(x,y)2+O(dM(x,y)4). Similarly, on the torus, c(x,y)=−∑j=1dcos(xj−yj) is periodic and, away from the cut locus, satisfies c(x,y)=−d+21dM(x,y)2+O(dM(x,y)4).
For SO(3), ∣⟨qx,qy⟩∣=cos(dSO(3)(x,y)/2). Hence, the antipodally symmetric kernel-induced cost is increasing with the rotational geodesic distance and is locally a positive multiple of dSO(3)(x,y)2. On SE(3), the cost adds a squared Euclidean translation term to this rotational cost and is therefore locally equivalent to a weighted sum of squared translational and rotational geodesic distances.
Thus, these costs respect the local geometry of their respective manifolds while enabling closed-form normalization and direct conditional sampling, which are the main computational advantages of our construction. The updated manuscript details these equivalences in Appendix A.1 at the end of each manifold-specific parameterization.
[W1.1] Dependence on ε.
Regarding the dependence on ε, we clarify that each proposed kernel Kε defines a well-posed EOT problem with induced cost cε(x,y)=−εlogKε(x,y), under the usual assumptions. For spheres and tori, these costs are independent of ε, so it acts as the standard entropy regularization parameter. For SO(3), the induced cost depends on ε, but this dependence is controlled: cε(x,y)=−εlogcosh(λR⟨x,y⟩/ε) converges uniformly to the fixed cost c0(x,y)=−λR∣⟨x,y⟩∣, with 0≤cε(x,y)−c0(x,y)≤εlog2. For SE(3), our construction uses σε2=εst2, making the translational cost ∥xt−yt∥2/(2st2) independent of ε, while the rotational component satisfies the same bound as on SO(3). Consequently, although ε affects both the cost and entropy regularization in the rotational constructions, the induced costs converge to fixed, geometrically meaningful costs as ε→0. Moreover, the corresponding EOT objective values differ from those of the fixed-limit-cost formulations by at most εlog2, providing a precise connection to conventional fixed-cost EOT.
[W2] Lack of metrics
In order to address the lack of metrics, we have introduced a new benchmark construction in Section 3.4, together with conditional and unconditional tangent-space Bures-Wasserstein unexplained variance percentages (cTBW-UVP and TBW-UVP), alongside MMD2. These metrics allow us to evaluate the learned conditional distributions against known optimal EOT conditionals, rather than relying exclusively on marginal metrics. Further details about the experimental construction and results are provided in our response to Reviewer K4Eg, [W1]. We also include these experiments in the updated manuscript.
On the other hand, it is important to clarify that evaluating continuous EOT couplings on manifolds remains challenging. Existing evaluation protocols are not always suitable for our method. For instance, RNOT uses metrics tailored to deterministic transport maps, which are not directly applicable to our stochastic couplings. Meanwhile, ERNOT relies on discrete reference solutions with small support sizes. Although this approach can be reasonable in low dimensions, such limited supports may poorly approximate continuous couplings in higher dimensions, making these metrics less reliable. Therefore, our contribution extends beyond the proposed solvers by introducing a new benchmark setup for evaluating continuous Riemannian EOT solvers that can accommodate general cost functions.
Concluding remarks. Kindly let us know if the clarifications provided address your concerns about our work. We're happy to discuss any remaining questions during the discussion phase. If you find our responses satisfactory, we would appreciate it if you could consider raising your score.