AI Paper Atlas

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

MIRO: MultI-Reward cOnditioned pretraining improves T2I quality and efficiency

Nicolas Dufour, Lucas Degeorge, Arijit Ghosh, Vicky Kalogeiton, David Picard

ICML 2026 · Image generation › Text-to-image diffusion · approach: T2I evaluation and controllable adapters

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
condition on seven rewards during pretraining instead of no reward conditioning0.36B flow matching T2I model · GenEval (real captions)overall: 52 -> 57Improved
condition on multiple rewards instead of aesthetic reward only (ablation)MIRO · GenEval (real captions)overall: 33 -> 57Improved
condition on multiple rewards instead of one alignment-related reward (CLIP, VQA) (ablation)MIRO · GenEval (real captions)overall: 57 -> 57No clear effect
condition on multiple rewards instead of single reward (ablation)MIRO · CC12M+LA6 reward metricsAestheticScore, PickScore, ImageReward, HPSv2Improved
add reward conditioning to pretraining0.36B flow matching T2I model · CC12M+LA6 training convergencesteps to reach baseline performance: 19x faster AestheticScore, 3.3x ImageRewardImproved
add synthetic captions to MIROMIRO · GenEvaloverall: 57 -> 68Improved
set aesthetic positive target to 0.625 at inference (ablation)MIRO (synthetic captions) · GenEvaloverall: 68 -> 75Improved
set aesthetic positive target to 0.625 at inference (ablation)MIRO (synthetic captions) · PartiPromptsAesthetic score: 6.28 -> 5.24Hurt
add best-of-N test-time scaling (128 samples) with ImageReward selectionMIRO (synthetic captions) · GenEvaloverall: 68 -> 75Improved
select samples by aesthetic score with 128-sample scalingMIRO (synthetic captions) · GenEvaloverall: 68 -> 63Hurt
use MIRO instead of baseline under best-of-N samplingMIRO · test-time scaling on ImageReward and PickScoresamples needed to match baseline at 128: 8 samples (ImageReward), 4 samples (PickScore)Improved
condition on multiple rewards instead of aesthetic reward only (ablation)MIRO · PartiPromptsImageReward: 0.00 -> 1.06Improved

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