Handling Imbalanced Pseudolabels for Vision-Language Models with Concept Alignment and Confusion-Aware Calibrated Margin

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Yuchen, Bai, Xuefeng, Li, Xiucheng, Guan, Weili, Nie, Liqiang, Chen, Xinyang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913819672969216
author Wang, Yuchen
Bai, Xuefeng
Li, Xiucheng
Guan, Weili
Nie, Liqiang
Chen, Xinyang
author_facet Wang, Yuchen
Bai, Xuefeng
Li, Xiucheng
Guan, Weili
Nie, Liqiang
Chen, Xinyang
contents Adapting vision-language models (VLMs) to downstream tasks with pseudolabels has gained increasing attention. A major obstacle is that the pseudolabels generated by VLMs tend to be imbalanced, leading to inferior performance. While existing methods have explored various strategies to address this, the underlying causes of imbalance remain insufficiently investigated. To fill this gap, we delve into imbalanced pseudolabels and identify two primary contributing factors: concept mismatch and concept confusion. To mitigate these two issues, we propose a novel framework incorporating concept alignment and confusion-aware calibrated margin mechanisms. The core of our approach lies in enhancing underperforming classes and promoting balanced predictions across categories, thus mitigating imbalance. Extensive experiments on six benchmark datasets with three learning paradigms demonstrate that the proposed method effectively enhances the accuracy and balance of pseudolabels, achieving a relative improvement of 6.29% over the SoTA method. Our code is avaliable at https://anonymous.4open.science/r/CAP-C642/
format Preprint
id arxiv_https___arxiv_org_abs_2505_02056
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Handling Imbalanced Pseudolabels for Vision-Language Models with Concept Alignment and Confusion-Aware Calibrated Margin
Wang, Yuchen
Bai, Xuefeng
Li, Xiucheng
Guan, Weili
Nie, Liqiang
Chen, Xinyang
Computer Vision and Pattern Recognition
Machine Learning
Adapting vision-language models (VLMs) to downstream tasks with pseudolabels has gained increasing attention. A major obstacle is that the pseudolabels generated by VLMs tend to be imbalanced, leading to inferior performance. While existing methods have explored various strategies to address this, the underlying causes of imbalance remain insufficiently investigated. To fill this gap, we delve into imbalanced pseudolabels and identify two primary contributing factors: concept mismatch and concept confusion. To mitigate these two issues, we propose a novel framework incorporating concept alignment and confusion-aware calibrated margin mechanisms. The core of our approach lies in enhancing underperforming classes and promoting balanced predictions across categories, thus mitigating imbalance. Extensive experiments on six benchmark datasets with three learning paradigms demonstrate that the proposed method effectively enhances the accuracy and balance of pseudolabels, achieving a relative improvement of 6.29% over the SoTA method. Our code is avaliable at https://anonymous.4open.science/r/CAP-C642/
title Handling Imbalanced Pseudolabels for Vision-Language Models with Concept Alignment and Confusion-Aware Calibrated Margin
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2505.02056