Research map / Image generation / Fast diffusion sampling
Come-Closer-Diffuse-Faster: Accelerating Conditional Diffusion Models for Inverse Problems Through Stochastic Contraction
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 CCDF instead of the strongest listed baseline | Conditional diffusion for FFHQ super-resolution · FFHQ test, ×4 super-resolution | FID: 63.14 -> 60.90 | Helped |
| Use CCDF instead of the strongest listed baseline | Conditional diffusion for FFHQ super-resolution · FFHQ test, ×8 super-resolution | FID: 80.27 -> 75.76 | Helped |
| Use CCDF instead of the strongest listed baseline | Conditional diffusion for FFHQ super-resolution · FFHQ test, ×16 super-resolution | FID: 92.32 -> 88.39 | Helped |
| Use CCDF instead of the strongest listed baseline | Conditional diffusion for AFHQ super-resolution · AFHQ test, ×4 super-resolution | FID: 18.70 -> 15.53 | Helped |
| Use CCDF instead of the strongest listed baseline | Conditional diffusion for AFHQ super-resolution · AFHQ test, ×8 super-resolution | FID: 34.85 -> 32.30 | Helped |
| Use CCDF instead of the strongest listed baseline | Conditional diffusion for AFHQ super-resolution · AFHQ test, ×16 super-resolution | FID: 47.28 -> 48.77 | Hurt |
| Use CCDF with 200 iterations rather than the strongest listed baseline | Conditional diffusion for FFHQ inpainting · FFHQ test, box 96 inpainting | FID: 46.42 -> 45.99 | Helped |
| Use CCDF with 200 iterations rather than the strongest listed baseline | Conditional diffusion for FFHQ inpainting · FFHQ test, box 128 inpainting | FID: 52.63 -> 49.77 | Helped |
| Use CCDF with 200 iterations rather than the strongest listed baseline | Conditional diffusion for FFHQ inpainting · FFHQ test, box 160 inpainting | FID: 66.25 -> 57.99 | Helped |
| Use CCDF instead of the strongest listed baseline | Diffusion MRI reconstruction · fastMRI, ×2 reconstruction | PSNR: 32.93 -> 33.41 | Helped |
| Use CCDF instead of the strongest listed baseline | Diffusion MRI reconstruction · fastMRI, ×4 reconstruction | PSNR: 31.45 -> 32.51 | Helped |
| Use CCDF instead of the strongest listed baseline | Diffusion MRI reconstruction · fastMRI, ×6 reconstruction | PSNR: 31.15 -> 31.30 | Helped |
Closest papers by meaning
- A Variational Perspective on Solving Inverse Problems with Diffusion Models (ICLR 2024 · similarity 0.92)
- Fast samplers for Inverse Problems in Iterative Refinement models (NeurIPS 2024 · similarity 0.91)
- Ambient Diffusion Posterior Sampling: Solving Inverse Problems with Diffusion Models Trained on Corrupted Data (ICLR 2025 · similarity 0.91)
- DEFT: Efficient Fine-tuning of Diffusion Models by Learning the Generalised $h$-transform (NeurIPS 2024 · similarity 0.91)
- Adapt and Diffuse: Sample-adaptive Reconstruction via Latent Diffusion Models (ICML 2024 · similarity 0.91)
- SITCOM: Step-wise Triple-Consistent Diffusion Sampling For Inverse Problems (ICML 2025 · similarity 0.91)
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.