- Rating
- 3(+1)
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
The paper introduces Double-Stitch, a simulation-free method for learning population dynamics from unpaired temporal snapshots. Building on Stitching, it jointly optimizes a KDE population trajectory, potential energy, and friction using a data-fitting loss and a second-order mechanical residual. Training avoids differentiating through sequential simulations, while the learned dynamics can be integrated for prediction. Experiments cover synthetic oscillators, embryoid-body single-cell data, and ocean vortices, with substantial training speedups over simulation-based methods.
Critical strengths
-
The collocation-vs-shooting framing is well-motivated and practical. Bringing simulation-free residual training to second-order population dynamics is a natural extension, and the speedup mechanism is well explained, namely the fact that the residual parallelizes over time nodes, while a rollout is sequential. Extending simulation-free residual training to second-order dynamics is well motivated. Moreover, the mathematical derivations are correct and rigorous.
-
The training speedups in the reported settings are substantial (Table 4) over WLM in the evaluated settings.
Critical weaknesses
-
Limited methodological novelty. Double-Stitch closely follows Stitching’s framework of jointly fitting a population path and an energy model, retaining the KDE representation, data-fitting objective, and energy form. Its principal algorithmic change is the second-order residual, supplemented by the non-gradient-velocity derivation and initial-velocity constraint. These are useful extensions, but the existing ablations do not establish whether learning the force improves held-out interpolation beyond a simpler acceleration-smoothing penalty. Such a comparison would strengthen the evidence for the dynamical modeling component.
-
Limited evidence that the learned law is accurate, especially for extrapolation. Extrapolation is the paper’s stated advantage over flow-based methods. On large-vortex forecasting, static Double-Stitch obtains under pushforward but under rollout, raising concerns about the accuracy of its learned dynamics over longer trajectories. Pushforward also uses observed velocities at its restart time, so this gap does not isolate extrapolation error. Nevertheless, it raises a substantive concern about the learned dynamics over longer uninterrupted trajectories. The time-dependent variant substantially improves rollout error to , close to WLM’s , but forecasting is evaluated at only one future snapshot. Although the paper acknowledges these limitations, the reported results leave unclear when the static model’s error accumulates and provide limited evidence of forecasting reliability beyond this setting.
Questions for authors
-
Residual and self-consistency. Could you report the empirical residual in Equation (12) at convergence, together with the between the fitted path and a rollout of the learned law from the fitted initial state, at the training times. Evaluating this at the training times would clarify how closely the learned law reproduces the optimized path.
-
Static ocean rollout. What explains the large static rollout error? Could you provide a per-snapshot rollout for the static and time-dependent models at ? This would show whether the discrepancy builds up within the observed interval, only beyond it, or both.
-
Ablations of the dynamics. Could you compare held-out interpolation accuracy against the same trajectory model with an acceleration-only penalty, obtained by setting , while retaining the residual weight? These would separate the contributions of path fitting, the smoothness prior and the learned force.
4 comments
W2. Limited evidence that the learned law is accurate, especially for extrapolation. Extrapolation is the paper’s stated advantage over flow-based methods. On large-vortex forecasting, static Double-Stitch obtains under pushforward but under rollout, raising concerns about the accuracy of its learned dynamics over longer trajectories. Pushforward also uses observed velocities at its restart time, so this gap does not isolate extrapolation error. Nevertheless, it raises a substantive concern about the learned dynamics over longer uninterrupted trajectories. The time-dependent variant substantially improves rollout error to , close to WLM’s , but forecasting is evaluated at only one future snapshot. Although the paper acknowledges these limitations, the reported results leave unclear when the static model’s error accumulates and provide limited evidence of forecasting reliability beyond this setting.
Q2. Static ocean rollout. What explains the large static rollout error? Could you provide a per-snapshot rollout for the static and time-dependent models at ? This would show whether the discrepancy builds up within the observed interval, only beyond it, or both.
Direct prediction scores the optimized particles, while rollout simulates the learned law, so a poor rollout does not imply a poor direct prediction. We report both in Table 14.
To demonstrate extrapolation across multiple future time steps, we add an undamped-orbit experiment to Appendix E.2. We generate 300 unpaired samples per snapshot from a Gaussian population moving around a circle. We train on the first rotation (), learning a static quadratic potential and damping, and simulate the next rotation from the final fitted positions and velocities. The learned stiffness is (true ) and the damping decreases from 0.1 to (true ). The table compares the fitted particles at three training snapshots with simulated predictions at three future snapshots (, mean std over three seeds). We see that the extrapolation error remains close to the fitting error.
| Snapshot | 4 | 8 | 12 | 16 | 20 | 24 |
|---|---|---|---|---|---|---|
| Double-Stitch | 0.0621 0.0018 | 0.0606 0.0023 | 0.0556 0.0014 | 0.0695 0.0020 | 0.0782 0.0003 | 0.0826 0.0021 |
| Type | Training | Training | Training | Forecast | Forecast | Forecast |
Note that in general for any method different laws can match the same snapshots and diverge after the last one (see our non-identifiability discussion in Section 2.1 and Appendix E.1).
We also provide the requested per-snapshot rollout errors on the large ocean vortex (Appendix E.4, Figure 12). We simulate from the observed initial positions and velocities without restarting, using the existing fits. The table reports over three seeds (training snapshots , forecast at ). We see that the static model's error already grows within the observed interval, while the time-dependent model stays closer to the data.
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | |
|---|---|---|---|---|---|---|---|---|---|---|
| 0.040 0.004 | 0.120 0.066 | 0.073 0.040 | 0.056 0.022 | 0.130 0.097 | 0.254 0.208 | 0.447 0.361 | 0.750 0.589 | 1.306 0.797 | 1.824 0.562 | |
| 0.101 0.003 | 0.163 0.025 | 0.084 0.024 | 0.107 0.009 | 0.131 0.057 | 0.164 0.067 | 0.207 0.052 | 0.183 0.019 | 0.144 0.007 | 0.223 0.008 |
From the static model's error grows by a factor of about two with every snapshot. The same happens when we start from the fitted initial state and compare with the fitted path (Figure 12, left), so the initial condition is not the cause. Both models have a similar particle residual (see Appendix C.5.3: and RMS), so the static model does not fit the law worse locally. Instead its small local error accumulates along the trajectory, which the residual does not control (Appendix C.5). The time-dependent model does not show this growth.
We agree that our introduction stated extrapolation too categorically, so we rewrote that sentence (Section 1) and added the rollout drift to the Limitations.
Q1. Residual and self-consistency. Could you report the empirical residual in Equation (12) at convergence, together with the between the fitted path and a rollout of the learned law from the fitted initial state, at the training times. Evaluating this at the training times would clarify how closely the learned law reproduces the optimized path.
To illustrate how residual and prediction quality change during fitting, we show residual RMS and prediction at initialization and after training (see Table 6 in Appendix C.5.4; same evaluation modes as in the paper; mean std over 3 seeds).
| Synthetic | EB LPO | Oceans interpolation | ||||
|---|---|---|---|---|---|---|
| Metric | Residual RMS | Prediction | Residual RMS | Prediction | Residual RMS | Prediction |
| Double-Stitch initialization | 20.749 | 0.410 | 3.988 0.051 | 1.970 0.005 | 1.214 0.002 | 0.231 0.000 |
| Double-Stitch fitted | 0.755 | 0.389 | 0.163 0.001 | 0.618 0.001 | 0.093 0.014 | 0.065 0.004 |
We see that training reduces both residual RMS and prediction error in these settings (training lowers the residual RMS by more than ten times). Note, however, that in general, low residual value does not necessarily mean that the dynamical law is satisfied (see the discussion in Appendix C.5).
We also provide the requested fit-to-rollout comparison. Specifically, we simulate from the fitted initial positions and velocities without restarting and compare with the fitted particles at the training snapshots (see Table 7 in Appendix C.5.4; raw Eq. (12) loss at convergence and the fit-to-rollout ).
| Dataset (training snapshots after ) | Eq. (12) loss | 1st | 2nd | 3rd | 4th | Mean |
|---|---|---|---|---|---|---|
| EB LPO () | 0.107 0.001 | 0.026 0.001 | 0.169 0.003 | 0.434 0.004 | 0.683 0.010 | 0.328 0.004 |
| Oceans interpolation () | 0.071 0.019 | 0.185 0.084 | 0.141 0.070 | 0.383 0.080 | 0.484 0.152 | 0.298 0.022 |
We see that the rollout follows the fitted path over a short horizon and then drifts. On EB the gap is after one snapshot interval and after four, which is as large as the prediction error (). On oceans it grows from to . So the learned law reproduces the fitted path locally, but not over a long uninterrupted simulation: small local errors accumulate, and the residual does not control this (see Appendix C.5). This is why direct and pushforward predictions, which do not need a long simulation, are more accurate than rollouts (Table 14). We added this to the Limitations.
In summary, we clarified how learned-particle predictions and rollouts are evaluated, reported empirical residual values, reported static ocean rollout error, and added an ablation of the dynamics. If we managed to clarify the points you raised, we would appreciate it if you could consider updating the score. In any case, thank you once again for the detailed review.
Thank you for the thorough and candid response. The acceleration-only ablation shows that the mechanical terms contribute beyond smoothing on EB LPO, L1O and both ocean tasks, and the requested residual and rollout diagnostics make clear what the residual does and does not guarantee. Together with the softened extrapolation claim and the added limitation, the paper now states its trade-off openly: a simulation-free method that trains 4–14 times faster than WLM at some cost in accuracy. On that basis I am raising my score to 3.
Thank you for the detailed review! We appreciate you noting that the derivations are rigorous and the training speedup is useful.
We numbered the weaknesses and questions and grouped them as {W1/Q3, W2/Q2, and Q1} and addressed them below. Also please see the updated manuscript: the edits addressing and hopefully resolving your concerns are highlighted in green (some edits that answer several reviewers may have another color, we give their specific sections below).
We agree that Double-Stitch builds on Stitching and keeps its KDE path and data loss. Our contribution is the extension to second-order mechanics. First, we derive the dynamical law for general (non-gradient) velocities (see Appendix B.4) (when fitting paths directly, velocities are not guaranteed to be gradient, so WLM’s derivation can’t be used). Second, we derive the corresponding residual and show what a small residual does and does not guarantee (see Appendices C.3 and C.5). Third, we show theoretically and empirically that the simulation-free method is generally faster to train because the residual is evaluated at all time nodes at once (see Section 4.4 and Proposition 1).
We add the requested acceleration-only ablation on EB and oceans datasets (see Table 21 in Appendix E.6). We fix all potential, entropy and interaction force coefficients and damping to zero, keeping the other training settings unchanged. Both models use direct prediction on the same held-out data and the same three seeds (mean ± std below; see Table 14 for Direct evaluation mode for oceans). For L1O, we average the three holdouts within each seed.
We find lower error with the full residual on LPO, L1O and both ocean tasks in every seed. On L2O, the difference is small and mixed across seeds. This shows that the mechanical terms contribute more than just smoothing, although the benefit depends on the task.