Research map / Image generation / Fast diffusion sampling
Noise-Adaptive Diffusion Sampling for Inverse Problems Without Task-Specific Tuning
Yingzhi Xia, Setthakorn Tanomkiattikun, Liangli Zhen, ZAIWANG GU
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use NA-NHMC instead of best baseline on nonlinear tasks (FFHQ, sigma 0.05) | NA-NHMC · FFHQ phase retrieval | PSNR: 18.52 -> 19.30 | Improved |
| use NA-NHMC instead of DPS | NA-NHMC · FFHQ HDR reconstruction | LPIPS: 0.168 vs 0.217 | Hurt |
| use NA-NHMC instead of SITCOM on linear sigma 0.05 | NA-NHMC · FFHQ SR x4 | PSNR: 27.35 vs 27.29 | No clear effect |
| use NA-NHMC at high noise sigma_y=0.20 (vs MAP-like DAPS) | NA-NHMC · FFHQ Gaussian deblurring | PSNR: 8.12 -> 25.57 | Improved |
| use NA-NHMC under impulse noise without retuning | NA-NHMC · FFHQ nonlinear deblurring | PSNR: 23.79 -> 24.16 | Improved |
| use NA-NHMC under speckle noise | NA-NHMC · FFHQ nonlinear deblurring | PSNR: 26.49 -> 27.97 | Improved |
| increase leapfrog steps L from 10 to 25 (ablation) | NA-NHMC · FFHQ SR x4 | PSNR: 26.86 -> 27.36 | Improved |
| increase initial step size from 0.02 to 0.10 (ablation) | NA-NHMC · FFHQ SR x4 | PSNR: 27.12 -> 27.31 | Improved |
| vary step size decay gamma from 0.91 to 0.99 (ablation) | NA-NHMC · FFHQ SR x4 | PSNR: 27.29 -> 27.30 | No clear effect |
| use NA-NHMC instead of SITCOM at SR x16, sigma 0.05 | NA-NHMC · FFHQ SR x16 | SSIM: 0.574 vs 0.531 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- SITCOM — SITCOM: Step-wise Triple-Consistent Diffusion Sampling For Inverse Problems (ICML 2025) · better on 4 of 5 shared results
- DMPlug — DMPlug: A Plug-in Method for Solving Inverse Problems with Diffusion Models (NeurIPS 2024) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Improving Diffusion Inverse Problem Solving with Decoupled Noise Annealing (CVPR 2025 · similarity 0.92)
- Divide-and-Conquer Posterior Sampling for Denoising Diffusion priors (NeurIPS 2024 · similarity 0.91)
- Improving Diffusion Models for Inverse Problems Using Optimal Posterior Covariance (ICML 2024 · similarity 0.91)
- A Variational Perspective on Solving Inverse Problems with Diffusion Models (ICLR 2024 · similarity 0.91)
- Unleashing the Denoising Capability of Diffusion Prior for Solving Inverse Problems (NeurIPS 2024 · similarity 0.91)
- Provable Posterior Sampling with Denoising Oracles via Tilted Transport (NeurIPS 2024 · similarity 0.90)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
