分类器(UML)
计算机科学
机器学习
人工智能
半监督学习
水准点(测量)
特征学习
学习分类器系统
测距
模式识别(心理学)
强化学习
大地测量学
电信
地理
作者
Yue Fan,Dengxin Dai,Anna Kukleva,Bernt Schiele
出处
期刊:Proceedings
[Institute of Electrical and Electronics Engineers]
日期:2022-06-01
卷期号:: 14554-14564
被引量:100
标识
DOI:10.1109/cvpr52688.2022.01417
摘要
Standard semi-supervised learning (SSL) using classbalanced datasets has shown great progress to leverage unlabeled data effectively. However, the more realistic setting of class-imbalanced data - called imbalanced SSL - is largely underexplored and standard SSL tends to under-perform. In this paper, we propose a novel co-learning framework (CoSSL), which decouples representation and classifier learning while coupling them closely. To handle the data imbalance, we devise Tail-class Feature Enhancement (TFE) for classifier learning. Furthermore, the current evaluation protocol for imbalanced SSL focuses only on balanced test sets, which has limited practicality in real-world scenarios. Therefore, we further conduct a comprehensive evaluation under various shifted test distributions. In experiments, we show that our approach outperforms other methods over a large range of shifted distributions, achieving state-of-the-art performance on benchmark datasets ranging from CIFAR-10, CIFAR-100, ImageNet, to Food-101. Our code will be made publicly available.
科研通智能强力驱动
Strongly Powered by AbleSci AI