- Rating
- 3(+1)
- Confidence
- 4
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
Paper proposes MoLa-PDE, latent-space PDE surrogate that jointly evolves latent features and their spatial coordinates. It combines geometry-aware interaction operator with potential-based motion, allowing a compact representation to follow moving and deforming structures.
Critical strengths
-
Moving latent coordinates is well-motivated idea for transport problems, where fixed representations may struggle to follow evolving structures.
-
Extensive experiments.
Critical weaknesses
-
Experiments use only 8 latent points in 1D and 16 in 2D. It is unclear whether the gains remain substantial when the static model receives more latent points under similar computational budget. How does it scale??
-
Results do not report variability across training seeds making the smaller improvements difficult to assess.
-
The stability of the moving support needs more discussion. Point clustering or nearly degenerate neighborhoods could affect the local least-squares derivative estimates.
-
Moving supports add noticeable computational overhead, while the model remains less accurate than some baselines on Navier–Stokes and cylinder flow. The accuracy–cost tradeoff needs stronger evaluation.
-
Missing Baselines. Can author compare against Transolver and Latent Mamba operator as it seems to be relevant recent baselines.
Questions for authors
-
Does moving the support still help when the static model uses more latent points at a matched inference cost? What's the tradeoff??
-
How does method prevent latent points from collapsing together or leaving poorly covered regions? Is the local least-squares system regularized when geometry becomes ill-conditioned?
-
Could you separately ablate global translation, scalar-potential motion and rotational motion to show which parts are needed and when they are useful??
-
Would learning an input-dependent initial support recover much of the gain without moving the points at every step? it seems particularly relevant to the cylinder case.
-
Could you report results across multiple seeds and evaluate beyond the training horizon?
-
Can you quantify how the method differs from existing adaptive-mesh and moving-point approaches, beyond applying movement in latent space [1, 2] and compare against them and with SSM based operator [3]??
[1] Wen, Shizheng, et al. "Geometry aware operator transformer as an efficient and accurate neural surrogate for pdes on arbitrary domains." Advances in Neural Information Processing Systems 38 (2026): 155423-155501.
[2] Mousavi, Sepehr, et al. "RIGNO: A Graph-based framework for robust and accurate operator learning for PDEs on arbitrary domains." Advances in Neural Information Processing Systems 38 (2026): 150039-150101.
[3] Tiwari, Karn, Niladri Dutta, and N. M. Krishnan. "Latent mamba operator for partial differential equations." arXiv preprint arXiv:2505.19105 (2025).
7 comments
Response to W1 / W4 / Q1
Response to W1 / W4 / Q1
We thank the reviewers for highlighting the tradeoffs between prediction accuracy, latent capacity, and computational cost. To investigate these questions, we double the number of latent points from 8 to 16 in 1D and from 16 to 32 in 2D, comparing static and moving supports at both latent sizes and measuring their full-rollout inference latency.
First, the benefits of latent motion persist as the latent size increases. At the doubled latent size, MoLa-PDE continues to outperform its static counterpart across all six PDE benchmarks, suggesting that the advantage of adaptive latent geometry is not limited to highly compressed representations.
Second, increasing the number of fixed latent points does not necessarily compensate for the absence of motion. On four of the six benchmarks (Burgers, Diffusion–Reaction, Advection, and Cylinder), the original moving model achieves lower prediction errors than the static model with twice as many latent points. This suggests that adapting where latent information is represented can be more effective than simply increasing the number of fixed spatial supports.
| PDE | Static (1x) | Motion (1x) | Static (2x) | Motion (2x) |
|---|---|---|---|---|
| Burgers | 2.16 | 1.74 | 1.90 | 1.62 |
| Diffusion–Reaction | 0.178 | 0.163 | 0.171 | 0.155 |
| Advection | 1.43 | 0.434 | 1.36 | 0.223 |
| Navier–Stokes | 9.92 | 9.04 | 8.75 | 8.06 |
| Shallow Water | 0.225 | 0.198 | 0.182 | 0.173 |
| Cylinder | 15.0 | 12.9 | 13.1 | 12.2 |
Comparison of relative error () for static and moving supports using the original and doubled latent sizes.
Third, the accuracy–cost comparison reveals different tradeoffs across spatial dimensions and PDE dynamics. In 1D, Motion-8 outperforms Static-16 on all three benchmarks, with only 26% higher full-rollout latency (30.83 vs. 24.41 ms). In 2D, Motion-16 and Static-32 have comparable latency (27.62 vs. 28.59 ms), with Motion-16 achieving lower error on Cylinder. Increasing the moving support to 32 points further improves accuracy over Static-32 on all three 2D benchmarks, at approximately 34% higher latency. These results indicate that latent motion offers an alternative to increasing static resolution, although its accuracy–cost advantage depends on the underlying dynamics.
| Configuration | Static (ms) | Motion (ms) |
|---|---|---|
| 1D, 8 points | 17.32 | 30.83 |
| 1D, 16 points | 24.41 | 37.19 |
| 2D, 16 points | 19.46 | 27.62 |
| 2D, 32 points | 28.59 | 38.21 |
Full-rollout inference latency (ms) of static and moving models at original and doubled latent sizes.
Overall, these experiments show that adaptive latent geometry provides benefits beyond increasing static latent capacity, while introducing additional computational cost. Rather than claiming universal accuracy–efficiency cost optimality, our results demonstrate that evolving the latent support offers a complementary way to improve compact PDE representations. We have incorporated these scaling and efficiency comparisons into the revised manuscript.
Response to W2 / Q5
Response to W2 / Q5
We thank the reviewer for this suggestion. To assess the robustness of our results across random seeds, we repeat the experiments for both MoLa-PDE and its static counterpart using three random seeds on all six benchmarks.
| PDE | Static | Motion | Rel. Diff. |
|---|---|---|---|
| Burgers | 24.06% | ||
| Diffusion–Reaction | 10.69% | ||
| Advection | 217.4% | ||
| Navier–Stokes | 9.860% | ||
| Shallow Water | 11.28% | ||
| Cylinder | 8.390% |
Relative error () over three random seeds for moving and static supports, reported as mean standard deviation. Relative difference is normalized by the moving mean.
For the beyond-horizon evaluation, we retrained Motion and Static with shortened supervision horizons and evaluated their free-running predictions over the remaining unseen timesteps. As shown in Table 9, Motion achieves lower extrapolation errors on five of the six PDE benchmarks, suggesting that the benefits of adaptive latent geometry generally extend beyond the supervised horizon. The exception is Navier–Stokes (NS), where Static achieves a lower extrapolation error than Motion (28.63 vs. 45.32). To better understand this difference, we further examine the prediction behavior and underlying flow dynamics.
On NS, Motion maintains lower errors within the supervised interval and comparable accuracy at the first extrapolation step, with the performance gap emerging over subsequent unseen steps. Interestingly, we find that these later steps also involve a shift in the underlying flow regime. The mean velocity RMS reconstructed from the vorticity fields increases by approximately 25% from frame 14 to frame 19, and by the final frame, all test trajectories exceed the maximum velocity RMS observed during training. Thus, this evaluation involves not only temporal extrapolation but also a shift toward higher flow velocities beyond the training distribution.
This distribution shift provides a plausible explanation for the observed behavior. In MoLa-PDE, latent features and support coordinates evolve jointly, allowing the representation to adapt to changing spatial structures. However, when extrapolating to unseen dynamical regimes, errors in the predicted motion may affect subsequent feature interactions through the evolving geometry, leading to additional error accumulation. The static model, in contrast, does not involve this coordinate–feature feedback. These observations suggest that the NS performance gap is associated with the additional challenge of extrapolating coupled latent dynamics under a shifted flow regime, rather than indicating a general limitation of moving supports in modeling transport and deformation. These observations are consistent with the benefits of adaptive latent geometry demonstrated across our standard benchmarks and controlled transport experiments, while also highlighting the challenges of generalizing learned support motion to dynamical regimes beyond the training distribution. We have incorporated these additional experimental results and analyses, together with a discussion of the corresponding limitations, into the revised manuscript.
| Dataset | Training Horizon | Extrapolation Horizon | Static | Motion |
|---|---|---|---|---|
| Burgers | 1–30 | 31–40 | 2.647 | 2.274 |
| Diffusion-reaction | 1–15 | 16–20 | 0.231 | 0.208 |
| Advection | 1–30 | 31–40 | 4.017 | 0.539 |
| Navier–Stokes | 1–14 | 15–19 | 28.63 | 45.32 |
| Shallow water | 1–15 | 16–20 | 1.137 | 0.707 |
| Cylinder | 1–10 | 11–14 | 15.71 | 14.64 |
Relative error () beyond the training horizon.
General Response and Summary of Revisions
We thank the reviewer for the detailed and constructive feedback. To address the questions, we conducted additional experiments and analyses to further evaluate the accuracy, efficiency, robustness, and stability of MoLa-PDE along six dimensions:
The detailed experimental results and methodological clarifications are provided in our responses below. We have incorporated these additional analyses into the revised manuscript and expanded the discussion of the corresponding tradeoffs and limitations.