Research map / Image generation / Image restoration
Cold Diffusion: Inverting Arbitrary Image Transforms Without Noise
Arpit Bansal, Eitan Borgnia, Hong-Min Chu, Jie S. Li, Hamid Kazemi, Furong Huang, Micah Goldblum, Jonas Geiping, Tom Goldstein
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 iterative cold sampling instead of direct reconstruction (ablation) | deblurring model · CelebA | FID: 36.37 -> 26.14 | Improved |
| use iterative cold sampling instead of direct reconstruction (ablation) | deblurring model · CelebA | RMSE: 0.083 -> 0.093 | Hurt |
| use iterative cold sampling instead of direct reconstruction (ablation) | inpainting model · CelebA | FID: 7.74 -> 5.73 | Improved |
| use iterative cold sampling instead of direct reconstruction (ablation) | super-resolution model · MNIST | FID: 4.05 -> 4.33 | Hurt |
| use iterative cold sampling instead of direct reconstruction (ablation) | super-resolution model · CelebA | FID: 112.84 -> 96.92 | Improved |
| add small Gaussian noise to break pixel symmetry in x_T (ablation) | blur cold diffusion generation · CelebA | FID: 97.00 -> 49.45 | Improved |
| add small Gaussian noise to break pixel symmetry in x_T (ablation) | blur cold diffusion generation · AFHQ | FID: 93.05 -> 54.68 | Improved |
| replace Gaussian noise degradation with blur (cold diffusion) | diffusion generation (estimated noise) · CelebA | FID: 23.11 -> 49.45 | Hurt |
Closest papers by meaning
- Warm Diffusion: Recipe for Blur-Noise Mixture Diffusion Models (ICLR 2025 · similarity 0.92)
- Neural Diffusion Models (ICML 2024 · similarity 0.91)
- A Variational Perspective on Solving Inverse Problems with Diffusion Models (ICLR 2024 · similarity 0.90)
- Simplified and Generalized Masked Diffusion for Discrete Data (NeurIPS 2024 · similarity 0.90)
- Denoising Diffusion Bridge Models (ICLR 2024 · similarity 0.90)
- Faster Diffusion: Rethinking the Role of the Encoder for Diffusion Model Inference (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.
