- Rating
- 1
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 3
Read review
Paper summary
This paper use graph constraint to regularize the flow matching based dynamics inference models. In classical dynamics inference algorithms such as OT-CFM, the velocity net is parameterized by a MLP. To take varaible correlations such as gene-gene regulations into account, the authors use a GNN, whose graph structure is obtained by some prior knowledge such as GRN, to replace the MLP. This technique advances the model performance on several datasets.
Critical strengths
The paper tackles with an important problem. Its method is also clear and easy to follow. I'd like to say this paper is well written.
Critical weaknesses
The new technique is quite natural and easy. (But turns out consistently useful. So that is not a big problem.)
My main concerns are 1.The baselines are quite easy, some critical baselines are missing.
2.The results in table1 and table2 are unmatched, which makes me wonder if there is any artifact.
Questions for authors
1.I am so surprised that only knowing the graph structure without the explicit form of the dynamical system can consistently improve the inferenced dynamics. Can author give some intuitive interpretation?
Main concerns
2.What is the definition of graph corruption? There are only experimental results, but no experiment details. Since this is a central ablation study, this definition is necessary.
3.Baselines are OT-MFM, OT-MMFM, OT-ALI-FM. First, the most basic baseline OT-CFM is missing. Second, these baselines and graph corruption ablations are not enough convincing for the claim that the improvment is gained through using the graph information. I advice the authors to add OT-CFM as a baseline. Also, I advice the authors to add naive GraphFM baselines where the graph are chosen to be (1) random graph (2) full connected graph (3) no graph (which is repalcing the GNN with MLP). Since GraphFM introduces several losses other than just graph constraint, I believe these baselines can further support the necessity of the graph constraints.
4.Why the last row of table1 is not consistent to the first row of table2? Aren't they the same experiment? I give rate 1 mainly due to this question. If there are any misunderstanding, I will raise my score.
2 comments
Question 3:
We thank the reviewer for the valuable suggestions. Following these suggestions, we have conducted additional experiments to further evaluate the effectiveness of GraphFM and the contribution of graph information.
- Additional Baseline Comparisons
We have added OT-CFM as an additional baseline, as suggested by the reviewer. We have also included OTP-FM, following the recommendation of another reviewer. The comparison with OT-CFM is presented below.
| Dataset / Setting | Metric ↓ | GraphFM | OT-CFM | OTP-FM(W2Inf) |
|---|---|---|---|---|
| Synthetic (Table 1) | MMD | 0.04570 ± 0.00726 | 0.08392 ± 0.00341 | 0.12461 ± 0.00170 |
| Synthetic (Table 1) | SWD | 0.08199 ± 0.01109 | 0.12809 ± 0.00516 | 0.24327 ± 0.00316 |
| EB (Table 3) | EMD | 1.3711 ± 0.0532 | 1.3839 ± 0.0042 | 1.4971 ± 0.0312 |
| Multi (Table 3) | EMD | 7.9472 ± 0.0604 | 8.4602 ± 0.0221 | 9.5370 ± 0.2852 |
| HT206B1 (Table 4) | EMD | 7.5191 ± 0.0892 | 7.8663 ± 0.0030 | 7.5491 ± 0.0239 |
| Traffic (Leave-One-Out, Table 5) | EMD | 34.930 ± 0.137 | 35.882 ± 0.143 | 37.497 ± 0.112 |
| Traffic (Leave-One-Out, Table 5) | MAE | 2.837 ± 0.010 | 3.055 ± 0.015 | 3.028 ± 0.006 |
| Traffic (Leave-One-Out, Table 5) | MSE | 34.714 ± 0.237 | 35.959 ± 0.212 | 39.983 ± 0.271 |
| Traffic (Joint Missing, Table 5) | EMD | 34.992 ± 0.034 | 35.994 ± 0.120 | 37.439 ± 0.186 |
| Traffic (Joint Missing, Table 5) | MAE | 2.844 ± 0.009 | 3.036 ± 0.009 | 3.014 ± 0.018 |
| Traffic (Joint Missing, Table 5) | MSE | 34.884 ± 0.116 | 36.373 ± 0.184 | 39.968 ± 0.395 |
As shown above, GraphFM consistently outperforms OT-CFM across all evaluated settings, further demonstrating the effectiveness of our framework.
- Additional Graph Structure Ablations
To further isolate the contribution of graph information, we have conducted additional ablation experiments in both single-cell trajectory inference and traffic data imputation, considering the three settings suggested by the reviewer: (1) random graph, (2) fully connected graphs, and (3) no graph, where the GNN is replaced with an MLP. For each dataset, we generated an independent directed random graph. For each realization, we sampled without replacement from all ordered pairs of distinct nodes, keeping the number of edges exactly equal to that of the original graph. Edges present in the original graph were eligible for sampling. We did not constrain node degrees or connectivity. Each sampled graph was fixed across all held-out time points and five training seeds.
| Setting (EMD ↓) | Original Graph (GraphFM) | Random Graph | Fully Connected GNN | No Graph (MLP) |
|---|---|---|---|---|
| EB single-cell | 1.371 ± 0.053 | 1.345 ± 0.033 | 1.379 ± 0.008 | 1.349 ± 0.008 |
| Multi single-cell | 7.947 ± 0.060 | 8.135 ± 0.258 | 8.020 ± 0.017 | 8.210 ± 0.032 |
| Traffic (LOO) | 34.930 ± 0.137 | 35.043 ± 0.098 | 36.311 ± 0.088 | 35.008 ± 0.031 |
| Traffic (Joint) | 34.992 ± 0.034 | 35.152 ± 0.074 | 36.273 ± 0.130 | 35.128 ± 0.090 |
The results show that using the original graph achieves the best performance in three out of four settings, outperforming random graph, fully connected graphs, and the MLP-based variant. The exception is the EB dataset, where the random graph yield slightly better results.
Overall, these additional ablations provide further empirical evidence that incorporating meaningful graph structure can improve the performance of GraphFM, beyond the benefits of the other components in our framework.
Question 1:
We thank the reviewer for this insightful question. Although the graph does not specify the functional form of the underlying dynamical system, it provides valuable information about dependencies among variables. By restricting irrelevant dependencies, such structural information can reduce the hypothesis space and introduce a useful inductive bias, potentially improving dynamics inference, particularly when observations are limited or noisy.
Our motivation is partly inspired by RegVelo [1], which incorporates gene regulatory network (GRN) priors into cellular dynamics inference. Biological processes are inherently complex, and their exact governing equations are often unknown. Although RegVelo assumes a specific form of RNA kinetics and parameterizes regulatory interactions using a nonlinear function, such formulations are still approximations of the underlying biological processes, for which alternative models, such as Hill-type regulation, also exist [2]. Importantly, RegVelo demonstrates that incorporating GRN priors can improve the inference of cellular dynamics within its assumed kinetic framework. This provides supporting evidence that structural information can offer valuable inductive biases even when the exact underlying regulatory mechanisms are not fully characterized.
Inspired by this observation, GraphFM explores the role of graph priors in a more general setting without assuming a specific mechanistic form of the governing dynamics. Rather than specifying how variables interact, we use the graph to constrain the dependencies of the interpolation model, and learning the corresponding dynamical behavior from data. Specifically, our message-passing GNN imposes an (L)-hop receptive field, such that each variable's interpolated state depends only on the endpoint variables within its graph-defined neighborhood.
[1] Wang, W., et al. RegVelo: Gene-regulatory-informed dynamics of single cells. Cell, 189(12), 3773–3800.e44, 2026.
[2] Tyson J J, Laomettachit T, Kraikivski P. Modeling the dynamic behavior of biochemical regulatory networks[J]. Journal of theoretical biology, 2019, 462: 514-527.
Question 2:
Thanks for your comments. We have made this point clearer. In Table 2, graph corruption is defined as replacing a fraction r of the directed edges in the original graph. Specifically, for a graph with E edges, we randomly remove \mathrm(rE) existing edges and add the same number of previously absent directed edges, uniformly sampled without replacement from all previously absent directed node pairs (excluding self-loops). Thus, the number of edges remains fixed. We use r\in{0,0.3,0.6,0.9,1.0}, where r=0 denotes the original graph and r=1 has no edges in common with it. The sampled graphs have no self-loops and are required to remain strongly connected. For each corruption level, the same graph is used across three training seeds. All conditions in this ablation use the SWD-only training setting.
Question 4:
Thanks for your comment. We have made our experimental setting clearer in our revised manuscript.. The first row of Table 2 reports the results obtained using SWD as the sole loss function, rather than the full loss configuration used in Table 1 (SWD + linear regularization + second-order smoothing). This explains the discrepancy between the two tables. Importantly, all experiments in Table 2 were conducted under the same loss configuration (SWD only), ensuring a fair comparison that isolates the effect of graph structure without interference from other regularization terms.
To further ensure transparency and reproducibility, we have made our implementation publicly available. The reviewer can verify the experimental settings and reproduce the results through our code repository: https://anonymous.4open.science/r/GraphFM-80E4A/