AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

Step-Wise Distribution-Aligned Style Prompt Tuning for Source-Free Cross-Domain Few-Shot Learning

Huali Xu, Li Liu, Tianpeng Liu, Shuaifeng Zhi, Shuzhou Sun, Ming‐Ming Cheng

TPAMI 2025 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Cross-domain few-shot with transformers · 14 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 style prompt to step-wise alignment (ablation)StepSPT (ConvNeXt) · 5-way 1-shot avgaccuracy: 63.98 -> 65.31Improved
add style prompt to step-wise alignment (ablation)StepSPT (ConvNeXt) · 5-way 5-shot avgaccuracy: 73.10 -> 73.70Improved
add step-wise distribution alignment to style prompt (ablation)StepSPT (ConvNeXt) · 5-way 1-shot avgaccuracy: 64.11 -> 65.31Improved
add step-wise distribution alignment to style prompt (ablation)StepSPT (ConvNeXt) · 5-way 5-shot avgaccuracy: 73.50 -> 73.70Improved
add style prompt only (no step-wise alignment) to baseline (ablation)ConvNeXt linear baseline · 5-way 1-shot avgaccuracy: 60.95 -> 64.11Improved
add step-wise distribution alignment only (no style prompt) to baseline (ablation)ConvNeXt linear baseline · 5-way 1-shot avgaccuracy: 60.95 -> 63.98Improved
add step-wise distribution alignment only to baseline (ablation)ConvNeXt linear baseline · 5-way 5-shot avgaccuracy: 72.17 -> 73.10Improved
add step-wise distribution alignment only to baseline, ISIC 5-shot (ablation)ConvNeXt linear baseline · ISIC 5-way 5-shotaccuracy: 52.51 -> 52.24No clear effect
StepSPT style prompt vs CoOp/CoCoOp text prompts in same pipeline (ablation)StepSPT · 5-way 1-shot and 5-shotaverage accuracyImproved
step size E=100 vs E=2 in step-wise alignment (ablation)StepSPT (ConvNeXt) · 5-way 1-shot avgaccuracy: 64.36 -> 65.31Improved
use StepSPT instead of LoRA on frozen CLIPCLIP · 5-way 1-shot avgaccuracy: 61.80 -> 61.16Hurt
use StepSPT instead of LoRA on frozen ConvNeXtConvNeXt · 5-way 1-shot avgaccuracy: 64.31 -> 65.31Improved

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