AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Steering Diffusion Models via Class-Contrastive Influence for Few-Shot Classification

Jeeyung Kim, Erfan Esmaeili, Qiang Qiu

ECCV 2026 · Image generation › Fast diffusion sampling · approach: Diffusion models as perception tools

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 C2I-rewarded RL fine-tuned diffusion augmentationViT-B/16 trained on original data · 3 MedMNIST (avg)AUC: 0.873 -> 0.894Improved
add C2I-rewarded RL fine-tuning on top of DataDreamDataDream · 3 MedMNIST (avg, ViT)AUC: 0.866 -> 0.894Improved
add C2I augmentation to ResNet18 (not used in RL)ResNet18 · 3 MedMNIST (avg)AUC: 0.842 -> 0.882Improved
add C2I augmentationViT-B/16 · PneumoniaMNISTAUC: 0.941 -> 0.945No clear effect
add C2I augmentation under noisy test inputsViT-B/16 · BreastMNIST noisyAUC: 0.761 -> 0.819Improved
add C2I augmentation under noisy test inputsViT-B/16 · DermaMNIST-binary noisyAUC: 0.800 -> 0.834Improved
add C2I augmentation under noisy test inputsViT-B/16 · PneumoniaMNIST noisyAUC: 0.907 -> 0.893Hurt
add C2I multi-class rewardViT-B/16 · DermaMNIST-allaccuracy: 0.648 -> 0.683Improved
lengthen RL fine-tuning from 10 to 20 epochs (ablation)C2I · BreastMNISTAUC: 0.83 -> 0.89Improved
use C2I reward instead of fine-tuning SD on misclassified samples (ablation)Simple hard-example baseline · BreastMNISTAUC: 0.778 -> 0.885Improved
shrink RL validation set from 135 to 16 per class (ablation)C2I · PneumoniaMNISTAUC: 0.944 -> 0.943No clear effect
increase number of synthesized images (ablation)DataDream and Dataset Expansion · BreastMNISTAUCHurt

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