- Rating
- 2
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
This paper studies diffusion-based inverse solvers that use Tweedie plug-in likelihood guidance, such as DPS. The authors observe that, for the same image and task, runs with larger likelihood-score norms in the low-noise stage tend to give worse LPIPS. Based on this signal, they propose NCLG. A theoretical analysis on a Gaussian mixture model relates the likelihood-score norm to the true conditional-score error.
Critical strengths
- The method is plug-and-play and needs no training. It improves four different host solvers, and the task coverage is broad.
- The comparisons under similar compute are convincing. NCLG performs better than hosts with more sampling steps, and also better than Best-of-8, CAS, ReGuidance, random restarts and data-consistency selection.
Critical weaknesses
- The writing is hard to read. Hedging phrases such as "in the evaluated settings" and "across the reported rows" are repeated many times. The paper is 38 pages long, and the appendix is also very long.
- Some experimental details are missing. I could not find the number of evaluation images in the main paper or the appendix. All experiments use pixel-space diffusion models, and no latent diffusion based inverse solver is tested. The paper also does not compare with newer and stronger solvers.
- The extra cost is large. On DPS, full NCLG takes about 5.5 times longer than the host, and peak GPU memory rises from 2.81 GB to 10.24 GB.
Questions for authors
- How many images are used for evaluation?
- Can the method be applied to inverse solvers based on latent diffusion models?
6 comments
Thank you for the additional experiments.
Could you clarify whether the 100 evaluation instances correspond to 100 distinct images, and how many sampling runs are used per image?
Please also report the runtime of the newly added comparisons and the peak GPU memory of the lower-cost RS+LNP configuration. The runtime comparisons are helpful, but the memory overhead remains a concern.
Reply to reviewer zbDW part. 1
We thank the reviewer for these follow-up questions.
1. Evaluation images and sampling runs
We are sorry for the confusion. We clarify that, in our implementation, we use 100 distinct images with 10 independent sampling runs per image.
Here we report the mean LPIPS and LPIPS standard deviation for DPS and DPS+RS+LNP on FFHQ center inpainting with , as in Table 4. We compute the mean and sample standard deviation of LPIPS over the 10 runs for each image, then average both statistics across the 100 images. For normalized std, we divide the standard deviation by the mean LPIPS for each image, then average these ratios across the 100 images.
| Method | Mean LPIPS ↓ | Average within-image LPIPS standard deviation | Normalized std ↓ |
|---|---|---|---|
| DPS | 0.316 | 0.0418 | 0.1365 |
| DPS + RS + LNP | 0.138 | 0.0031 | 0.0217 |
We find that NCLG is more stable across sampling runs. This may be attributed to RS, which obtains a data-informed initialization through deterministic inversion of the first-pass reconstruction. Similar first-pass reconstructions of the same observation can lead to similar restart states, supporting consistent refinement across runs. We will clarify the number of images and runs per image and include these stability statistics in the revised manuscript.
2. Runtime of the newly added comparisons
We report the runtimes of the additional comparisons on FFHQ center inpainting with , using 100 distinct images and 10 sampling runs per image. We also compare the augmented baselines to compare reconstruction quality and runtime at a larger sampling budget.
Recent solvers and their NCLG variants.
| Method | LPIPS ↓ | Low-noise norm aggregate ↓ | Time (s/image) ↓ |
|---|---|---|---|
| DPS | 0.316 | 938.87 | 5.25 |
| DPS (1,200 steps) | 0.263 | 915.53 | 30.42 |
| DPS + RS + LNP | 0.138 | 106.74 | 13.48 |
| DiffStateGrad-DPS | 0.264 | 892.63 | 6.32 |
| DiffStateGrad-DPS (1,000 steps) | 0.248 | 838.43 | 29.20 |
| DiffStateGrad-DPS + RS + LNP | 0.132 | 105.81 | 14.79 |
| SPGD | 0.247 | 517.92 | 25.74 |
| SPGD (1,000 outer steps) | 0.219 | 454.82 | 115.94 |
| SPGD + RS + LNP | 0.125 | 96.14 | 53.27 |
SPGD step counts refer to outer diffusion steps, with five warm-up iterations per step.
Latent diffusion (LDM-VQ-4).
| Method | LPIPS ↓ | Low-noise latent likelihood-score norm aggregate ↓ | Time (s/image) ↓ |
|---|---|---|---|
| Latent DPS (LDM-VQ-4) | 0.324 | 74.35 | 48.13 |
| Latent DPS (LDM-VQ-4, 1,000 steps) | 0.296 | 72.07 | 234.85 |
| Latent DPS (LDM-VQ-4) + RS + LNP | 0.143 | 36.51 | 93.57 |
Compared with increasing host sampling steps, RS+LNP achieves 42.9–51.7% lower LPIPS with 49.3–60.2% less runtime across all four hosts, demonstrating a better computation–performance trade-off. We will add these runtimes and extended-step comparisons to the corresponding experimental tables in the revision.
We thank the reviewer for the helpful feedback on our presentation and evaluation. We respond to the concerns in turn below.
W1. Presentation and length
We thank the reviewer for this feedback. Regarding presentation, we have removed redundant qualifiers such as "in the evaluated settings" and "across the reported rows" from the introduction, experimental discussion, and appendix to make the writing more direct. The experimental scope is specified in Section 5.1.
Regarding length, the paper covers theoretical analysis, three sampling modules, and evaluations across multiple solvers and tasks. We have made this material easier to navigate: Section 5.4 summarizes the additional experiments in four focused bullet points, each with a direct reference to the corresponding appendix section.
W2 and Q1. Number of evaluation images
We thank the reviewer for requesting this clarification. As stated in Appendix H.3 (Evaluation Protocol), mean LPIPS is computed over 100 evaluation instances per configuration. Each instance uses an independent sampling run. We have stated this explicitly in Section 5.1.
W2. Comparison with more recent solvers
The paper already includes comparisons with DCPS [2024], MGPS [2025], DAPS [2025], CAS [2024], adapted misalignment-map selection [2023], and PnP-DM [2024]. The NCLG variants improve reconstruction quality over their respective host solvers. To supplement these results, we additionally provide comparisons with DiffStateGrad-DPS [1, 2025] and SPGD [2, 2025] on FFHQ center inpainting with σy=0.3, using the same evaluation instances, observation operator, and pretrained prior. The NCLG variants retain each host's gradient-management mechanism and add RS+LNP, with NSS disabled.
W2 and Q2. Applicability to latent diffusion models
Yes, the norm-control mechanism can be adapted to latent diffusion solvers directly. Let D be the decoder and z^0(zt,t) the clean-latent estimate. For Gaussian observation noise, the corresponding likelihood guidance is gtz=−∇zt2σy2∣y−A(D(z^0(zt,t)))∣22. LNP can use this latent-space guidance and its norm; RS can restart through the latent model's inversion procedure; and NSS can select among stochastic latent proposals using low-noise norm probes. To evaluate this extension, we use the pretrained unconditional FFHQ LDM-VQ-4 model [3] on FFHQ 256×256 center inpainting with σy=0.3. The setting uses a 200-step host schedule and the same 100 evaluation instances, masks, and noisy observations as our pixel-space evaluation.
W3. The extra cost is large
We thank the reviewer for raising this concern. We clarify that our existing experiments already show that our method outperforms the baselines under comparable computational budgets. For a direct comparison, we reproduce below the results for ImageNet256 center inpainting with σy=0.3 reported in Appendix H.8.
Lower-cost configuration: RS+LNP versus 1,000-step baselines.
Higher-cost configuration: full NCLG versus 2,000-step baselines.
Full NCLG applies NSS at the start of the second pass, tinv=⌈0.7T⌉, as clarified in Section 4.1 and Appendix H.5.
These results are reported in Appendix H.8, Table 19. We have highlighted both budget regimes in Section 5.2.
References
[1] Diffusion State-Guided Projected Gradient for Inverse Problems. [2] Enhancing Diffusion Model Stability for Image Restoration via Gradient Management. [3] High-Resolution Image Synthesis with Latent Diffusion Models.