AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

ADIEE: Automatic Dataset Creation and Scorer for Instruction-Guided Image Editing Evaluation

Sherry X. Chen, Yi Wei, Luowei Zhou, Suren Kumar

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Multimodal LLM benchmarks · 7 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
add score token and MLP decoder (ablation)ADIEE scorer · ImagenHubSpearman correlation with human: 0.2931 -> 0.3450Improved
add score token and MLP decoder (ablation)ADIEE scorer · AURORA-Bench (point-wise)Spearman correlation with human: 0.3558 -> 0.4744Improved
add score token and MLP decoder (ablation)ADIEE scorer · GenAI-Benchpairwise accuracy: 58.32 -> 59.41Improved
add score token and MLP decoder (ablation)ADIEE scorer · AURORA-Bench (pair-wise)pairwise accuracy: 20.50 -> 52.38Improved
add reward conditioning prompts when fine-tuning MagicBrush on scored data (ablation)MagicBrush · ImagenHub (MagicBrush editing)avg score by ADIEE scorer: 4.27 -> 6.67Improved
train reward-conditioned model on all samples instead of only high-score samples (ablation)MagicBrush · ImagenHub (MagicBrush editing)avg score by ADIEE scorer: 5.09 -> 6.67Improved
fine-tune MagicBrush with ADIEE reward conditioningMagicBrush · ImagenHubavg score by ADIEE scorer: 5.90 -> 6.43Improved
fine-tune MagicBrush with reward feedback learningMagicBrush · ImagenHubavg score by ADIEE scorer: 5.90 -> 6.30Improved
reward conditioning on MagicBrush, CLIP-IMagicBrush · MagicBrush editingCLIP-I: 95.231 -> 95.215No clear effect
reward conditioning on MagicBrush, CLIP-DMagicBrush · MagicBrush editingCLIP-D: 0.13604 -> 0.13279Hurt
reward conditioning on MagicBrush, DINO-IMagicBrush · MagicBrush editingDINO-I: 90.643 -> 89.935Hurt
reward conditioning on MagicBrush, GPT-4o scoreMagicBrush · MagicBrush editingGPT-4o score: 4.02 -> 4.36Improved

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