AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

A Noise is Worth Diffusion Guidance

Donghoon Ahn, Jiwon Kang, Sanghyun Lee, Jaewon Min, Minjae Kim, Wooseok Jang, Hyoungwon Cho, Sayak Paul, SeonHwa Kim, Eunju Cha, Kyong Hwan Jin, Seungryong Kim

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Guidance for conditional diffusion · 46 citations (OpenAlex)

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 changedWhereMeasuredEffect
Refine initial noise with NoiseRefine instead of Gaussian noise (no guidance in both)Stable Diffusion 2.1 · MS-COCO 2014 val (30K prompts)FID: 42.71 -> 11.39Improved
Refined noise without guidance versus Gaussian noise with CFG+PAGStable Diffusion 2.1 · MS-COCO 2014 val (30K prompts)FID: 13.38 -> 11.39Improved
Refined noise versus guidance on Inception ScoreStable Diffusion 2.1 · MS-COCO 2014 val (30K prompts)Inception Score: 37.64 -> 35.73Hurt
Replace guidance by refined noise to halve inference costStable Diffusion 2.1 · MS-COCO 2014 val (30K prompts)inference time (s): 2.589 -> 1.504Improved
Refine noise instead of Gaussian noise on DrawBenchStable Diffusion 2.1 · DrawBenchPickScore: 19.67 -> 20.90Improved
Increase training denoising steps from 5 to 10 (ablation)NoiseRefine · MS-COCO FID (ablation)FID: 13.74 -> 13.36Improved
Add aesthetic-score filtering of training data (ablation)NoiseRefine · MS-COCO FID (ablation)FID: 15.27 -> 13.55Improved
Initialize noise refiner from pretrained diffusion UNet with LoRA instead of from scratch (ablation)NoiseRefine · Filtered MS-COCO (25K steps)FID: 37.87 -> 13.74Improved
Add multistep score distillation (detach gradients through denoiser) (ablation)NoiseRefine training · Optimization convergenceconvergence speed and image qualityImproved
Train noise refiner instead of LoRA-fine-tuning the denoising network (ablation)Guidance-free T2I pipeline · Filtered MS-COCO trainingconvergence and qualityImproved
Refined noise versus Gaussian noise plus CFG in user studyStable Diffusion 2.1 · User study (45 participants)image quality preference (%): 46.04% -> 53.96%No clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy