S2match: self-paced sampling for data-limited semi-supervised learning

• Main Authors: Guan, Dayan, Xing, Yun, Huang, Jiaxing, Xiao, Aoran, El Saddik, Abdulmotaleb, Lu, Shijian
• Other Authors: College of Computing and Data Science
• Format: Article
• Language: English
• Published: 2025
• Subjects: Computer and Information Science; Semi-supervised learning; Self-paced learning
• Online Access: https://hdl.handle.net/10356/182563
• Institution: Nanyang Technological University

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...