- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 1
- Soundness
- 1
Read review
Paper summary
The paper tries to connect diffusion and flow matching using fractional Brownian motion, whose Hurst index controls noise smoothness and temporal correlation. Its key trick is adding the noise process’s terminal value to a finite-dimensional approximation, helping preserve the straight-line limit as the Hurst index approaches one. The main contributions are this improved approximation, closed-form approximation coefficients, and a Hurst-conditioned generative model that lets one choose the noise regime at sampling time without retraining. Verifications include mathematical analysis, approximation-error plots, toy trajectories, and image-generation experiments on MNIST, CIFAR-10 and latent-space CelebA-HQ. Some configurations slightly improve FD-DINOv2 or FID
Critical strengths
The terminal-value augmentation addresses a concrete weakness in earlier noise approximations: losing Type I fractional Brownian motion’s straight-line behavior as the Hurst index approaches one. The idea has a clear mathematical motivation, and the approximation-error plots support it. This gives the paper an interpretable contribution that could be useful beyond its experiments.
Critical weaknesses
-
The overall evaluation is weak. The main results benchmarking on CIFAR-10 and CelebA-HQ are far from convincing because the improvements are mostly minor. The recalls of many models on CelebA-HQ are too low (around 0.3), not sure if there is an evaluation issue.
-
Proposition 3 claims convergence to straight-line paths over the full interval [0,1], but Appendix Proposition 12 rules out that convergence for the canonical construction. The main text also uses a regularized drift that differs from the proof’s setup. Since this limit is central to the paper’s diffusion-to-flow story, the authors need a corrected theorem that actually matches their model.
-
Missing comparisons of training and evaluation efficiency with the baselines.
Questions for authors
none
3 comments
Rebuttal (ii)
The overall evaluation is weak. The main results benchmarking on CIFAR-10 and CelebA-HQ are far from convincing because the improvements are mostly minor.
We respectfully disagree with the reviewer that the reported improvements are mostly minor. Comparing GFI against the strongest diffusion and flow-matching baselines in our experiments, we observe a 17.76% relative improvement in FID on CIFAR-10 (5.80 -> 4.77, Table 1) and an 8.96% improvement on CelebA-HQ (7.92 -> 7.21, Table 2). To put these improvements into perspective, influential conceptual contributions to generative modeling have reported relative FID improvements of comparable magnitude, such as 13.3% (DDPM++ -> DDPM++cont.) and 10.2% (NCSN++ -> NCSN++cont.) for score-based generative models [1, Table 3] and 15.1% for flow matching [2, Table 1]. We believe these comparisons demonstrate that the improvements achieved by GFI are substantial, particularly when considered alongside our theoretical and conceptual contributions. Furthermore, we highlight again the low additional computational cost of GFI and the broad evaluation of baseline configurations in the appendix K.
The recalls of many models on CelebA-HQ are too low (around 0.3), not sure if there is an evaluation issue.
In the reported results, precision and recall were evaluated using the same feature extractor and checkpoint selected for FD-DINOv2, following [4]. To address the reviewer's concern and allow a better comparison with related work, we additionally evaluate precision and recall using Inception-V3 features at the checkpoints selected by FID. As shown in the table below, recall improves across all models, reaching 0.37–0.40 compared with the previously reported 0.31–0.34. Compared with LDM [3, Table 1], which reports a recall of 0.49, our models still achieve lower recall, but higher precision. Together with the precision–recall trade-off, we believe these are sufficient explanation for the reported recall values. Together with the precision–recall trade-off, we believe that this is a a reasonable explanation for the reported recall values.
| Method | Recall (inceptionv3) | Precision (inceptionv3) | Recall (dino, Table 2) | Precision (dino, Table 2) |
|---|---|---|---|---|
| LDM [3, Table] | 0.49* | 0.72 | - | - |
| Diffusion FM | 0.38 | 0.76 | 0.33 | 0.78 |
| VP | 0.37 | 0.77 | 0.32 | 0.80 |
| FM | 0.39 | 0.73 | 0.34 | 0.75 |
| GFI$_{FM(K=7)}$ | 0.39 | 0.76 | 0.34 | 0.80 |
| GFI$_{FM(K=10)}$ | 0.40 | 0.76 | 0.31 | 0.80 |
*calculated between the 28k training images and 50k generated images, while we use only 28k generated images.
[1] Song et al. Score-Based Generative Modeling through Stochastic Differential Equations. ICLR 2021.
[2] Lipman et al. Flow Matching for Generative Modeling. ICLR 2023.
[3] Rombach et al. High-Resolution Image Synthesis with Latent Diffusion Models. CVPR 2022.
[4] Stein et al. Exposing flaws of generative model evaluation metrics and their unfair treatment of diffusion models. NeurIPS 2023.
[5] Gao et al. Diffusion Meets Flow Matching: Two Sides of the Same Coin, ICLR Blogpost Track 2025.
[6] Nobis et al. Generative Fractional Diffusion Models. NeurIPS 2024.
Presentation: 1: Difficult to follow despite careful reading.
We have revised our manuscript, particularly Section 4 and the appendix, to improve the clarity and readability. We hope these revisions address the reviewer's concerns regarding the presentation of our work.
Lastly, we would like to point the reviewers attention to the qualitative results included now in Appendix B.4, the new abaltion regarding few step samplings from GFI and additional baselines in our rebuttal to Reviewer Lz1L.
Rebuttal (iii)
We thank the reviewer again for identifying a mistake in our theoretical contribution. We hope that our revisions have addressed their concerns and would appreciate a reconsideration of their assessment. We would be happy to clarify any remaining questions.
Rebuttal (i)
We thank the reviewer for their feedback and address the three concerns raised by the reviewer in detail below.
We thank the reviewer for noticing this mistake. We have updated our claims and proofs for arbitrary T>0 and now explicitly restrict convergence in Proposition 3 to [0,S] with S<T. The revised results are summarized below, with Lp convergence holding for every finite p≥1 under the common-terminal coupling (Assumption 1):
We further note that the extension to arbitrary time horizons was already discussed in our original submission (lines 953–957), where we outlined the proof strategy using the self-similarity of fBM to rescale the time horizon. However, this was not correctly reflected in Proposition 3. The discrepancy happened because we initially formulated the construction with T=1 and simulated up to T−δ. To evaluate all models on the same interval [0,1], we subsequently changed the implementation to T=1+δ for GFI(FM), without explictly updating the theoretical statement accordingly. The revised manuscript now explicitly uses S=T=1 for GFI(VP) and S=1, T=1+10−2 for GFI(FM) (Secs. 4, 6, and E.2), with the driving noise defined on [0,T].
We apologize for the incorrect statement and thank the reviewer again for pointing it out. We hope that our revisions resolve the reviewer's concerns regarding the soundness of our theoretical contribution.
We thank the reviewer for raising this point. The additional computations of our method contribute only a small fraction of the overall computational cost compared to neural network evaluations for models of size DiT-B/2 or larger. We have now included a runtime comparison with the baselines in Sec. K.1 and provide the results below, reporting computational cost (KFLOPs) and wall-clock time (ms) for both training and sampling. All measurements were performed on an NVIDIA RTX 5090 with 32 GB VRAM and we report the median of 30 timings of either a full forward and backward pass or a sampling step. We observe in the table below that GFI incurs only a small additional computational cost during training (+1.1% ms) and sampling (+3.43 ms), when using DiT-B/2, that amortizes over the computations of the score network.
[1] Song et al. Score-Based Generative Modeling through Stochastic Differential Equations. ICLR 2021.