AI Paper Atlas

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

Consis-GCPO: Consistency-Preserving Group Causal Preference Optimization for Vision Customization

Qiaoqiao Jin, Dong She, Siming Fu, Hualiang Wang, Mu Liu, Jidong Jiang

ICLR 2026 · Image generation › Text-to-image diffusion · approach: Preference optimization for diffusion

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
add full decoupled causal interventions to Flow-GRPO (ablation)Flow-GRPO · Multi-subject R2IDINO-I: 0.551 -> 0.572Improved
add full decoupled causal interventions to Flow-GRPO (ablation)Flow-GRPO · Multi-subject R2VCLIP-I: 0.645 -> 0.674Improved
add prompt-only interventions to Flow-GRPO (ablation)Flow-GRPO · Multi-subject R2ICLIP-T: 0.325 -> 0.338Improved
add prompt-only interventions to Flow-GRPO (ablation)Flow-GRPO · Multi-subject R2IDINO-I: 0.551 -> 0.544Hurt
add reference-only interventions to Flow-GRPO (ablation)Flow-GRPO · Multi-subject R2ICLIP-I: 0.742 -> 0.780Improved
add reference-only interventions to Flow-GRPO (ablation)Flow-GRPO · Multi-subject R2VCLIP-T: 0.265 -> 0.255Hurt
use joint instead of alternating optimization of text and image rewards (ablation)Consis-GCPO · DreamBench (single-subject)DINO-I: 0.762 -> 0.781Improved
use joint instead of sequential optimization (ablation)Consis-GCPO · DreamBench (single-subject)CLIP-T: 0.308 -> 0.325Improved
joint optimization shares backward pass (ablation)Consis-GCPO · DreamBench (single-subject)training cost: 1.8x slower (alternating) -> 1.0xImproved
raise lambda_Ir from 1.0 to 2.0 (ablation)Consis-GCPO · DreamBench (single-subject)CLIP-T: 0.325 -> 0.305Hurt
use softmax temperature 1.0 instead of 0.8 (ablation)Consis-GCPO · DreamBench (single-subject)DINO-I: 0.765 -> 0.781Improved
use softmax temperature 1.0 instead of 1.2 (ablation)Consis-GCPO · DreamBench (single-subject)DINO-I: 0.772 -> 0.781Improved

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