AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

Diverse Text-to-Image Generation via Contrastive Noise Optimization

Byungjun Kim, Soobin Um, Jong Chul Ye

ICLR 2026 · Image generation › Text-to-image diffusion · approach: Inference-time T2I improvement · 11 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
Optimize initial noise with contrastive loss (CNO) instead of plain DDIMSD1.5 · MS-COCO val promptsVendi Score: 4.6949 -> 4.7855Improved
Optimize initial noise with CNOSDXL · MS-COCO val promptsMSS: 0.2169 -> 0.1623Improved
Optimize initial noise with CNOSD1.5 · MS-COCO val promptsPickScore / Image-Reward: PickScore 21.5081 -> 21.3779; IR 0.2222 -> 0.1284Hurt
Use gamma=0.01 (strong attraction) (ablation)CNO · MS-COCO val promptsVendi Score: 4.7945 -> 4.7412Hurt
Use gamma=100.0 (weak attraction) instead of 1.0 (ablation)CNO · MS-COCO val promptsImage-Reward: 0.0360 -> 0.0185Hurt
Lower gamma from 1.0 toward 0.7 (ablation)CNO · SD1.5 metric variancesample variance of metrics: Vendi variance 0.076 -> 0.050 (x1e-4)Improved
Pool Tweedie latents to w=4 (ablation)CNO · PickScore-Vendi Pareto frontierPareto frontierHurt
Use moderate pooling w=8 or 16 instead of full resolution w=64 (ablation)CNO · PickScore-Vendi Pareto frontierPareto frontierNo clear effect
Keep full spatial resolution (w=64) instead of w=1 (ablation)CNO · T2I-CompBenchImage-Reward: -0.2335 -> -0.1697Improved
Keep full resolution (w=64) instead of w=1 (ablation)CNO · GenEvalVendi Score: 4.8374 -> 4.8376No clear effect
Scale initial noise std by 1.01-1.05 instead of optimizing (ablation)DDIM · SD1.5Vendi Score: 4.7029 -> 4.7058No clear effect
Scale initial noise std to 1.1 or more (ablation)DDIM · SD1.5PickScore / Vendi: Vendi 4.7029 -> 4.5561 at 1.15Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most 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

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