Learning from Uncertain Similarity and Unlabeled Data
This work addresses privacy concerns in similarity-based learning for applications where sensitive label information must be protected, representing an incremental improvement over existing methods.
The paper tackles the problem of privacy risks from precise similarity annotations in weakly supervised learning by proposing USimUL, a framework that embeds uncertainty into similarity pairs to reduce label leakage, and demonstrates superior classification performance on benchmark and real-world datasets.
Existing similarity-based weakly supervised learning approaches often rely on precise similarity annotations between data pairs, which may inadvertently expose sensitive label information and raise privacy risks. To mitigate this issue, we propose Uncertain Similarity and Unlabeled Learning (USimUL), a novel framework where each similarity pair is embedded with an uncertainty component to reduce label leakage. In this paper, we propose an unbiased risk estimator that learns from uncertain similarity and unlabeled data. Additionally, we theoretically prove that the estimator achieves statistically optimal parametric convergence rates. Extensive experiments on both benchmark and real-world datasets show that our method achieves superior classification performance compared to conventional similarity-based approaches.