S2Match: Self-paced sampling for data-limited semi-supervised learning
Guan, Dayan Xing, Yun Huang, Jiaxing Xiao, Aoran El Saddik, Abdulmotaleb
Guan, Dayan
Xing, Yun
Huang, Jiaxing
Xiao, Aoran
El Saddik, Abdulmotaleb
Supervisor
Department
Computer Vision
Embargo End Date
Type
Journal article
Date
2025
License
Language
English
Collections
Research Projects
Organizational Units
Journal Issue
Abstract
Data-limited semi-supervised learning tends to be severely degraded by miscalibration (i.e., misalignment between confidence and correctness of predicted pseudo labels) and stuck at poor local minima while learning from the same set of over-confident yet incorrect pseudo labels repeatedly. We design a simple and effective self-paced sampling technique that can greatly alleviate the impact of miscalibration and learn more accurate semi-supervised models from limited training data. Instead of employing static or dynamic confidence thresholds which is sensitive to miscalibration, the proposed self-paced sampling follows a simple linear policy to select pseudo labels which eases repeated learning from the same set of falsely predicted pseudo labels at the early training stage and lowers the chance of being stuck at local minima effectively. Despite its simplicity, extensive evaluations over multiple data-limited semi-supervised tasks show the proposed self-paced sampling outperforms the state-of-the-art consistently by large margins.
Citation
D. Guan, Y. Xing, J. Huang, A. Xiao, A. El Saddik, and S. Lu, “S2Match: Self-paced sampling for data-limited semi-supervised learning,” Pattern Recognit, vol. 159, p. 111121, Mar. 2025, doi: 10.1016/J.PATCOG.2024.111121.
Source
Pattern Recognition
Conference
Keywords
Semi-supervised learning, Self-paced learning, Miscalibration, Image classification, Limited data
Subjects
Source
Publisher
Elsevier