Shift-Aware Calibration for Fine-Tuned CLIP: Leveraging Image-Text Alignment 文章

ArXiv CS.CV2026-07-28PAPERen作者: Song-Lin Lv, Yu-Yang Chen, Zhi Zhou, Lan-Zhe Guo

详细信息

来源站点
ArXiv CS.CV
作者
Song-Lin Lv, Yu-Yang Chen, Zhi Zhou, Lan-Zhe Guo
文章类型
PAPER
语言
en
发布日期
2026-07-28

摘要

arXiv:2501.19060v4 Announce Type: replace Abstract: Vision-language models (VLMs), such as CLIP, adapt effectively to downstream tasks through prompt tuning, but fine-tuning can misalign predictive confidence and accuracy, particularly on unseen classes. Existing VLM-specific calibration methods mainly rely on textual features of train classes, limiting their applicability across class and distribution shifts. We propose \textbf{Shift-Aware Calibration (SAC)}, a training-free, sample-wise method that uses the discrepancy between the output logits of the original and fine-tuned CLIP as a calibration signal. We explicitly call this quantity \emph{logit shift}, since it is measured in prediction space rather than between hidden representations. SAC maps this logit shift to a positive scaling factor that adjusts confidence while preserving the predicted class.

相关事件

暂无数据

相关公司查看全部 (4)

A
ATHCOMPANY
A
AdjustCOMPANY
A
Advanced Photon SourceRESEARCH_INSTITUTE

相关人物

暂无数据