Diversify and Conquer: Open-Set Disagreement for Robust Semi-Supervised Learning With Outliers

Abstract

Conventional semi-supervised learning (SSL) ideally assumes that labeled and unlabeled data share an identical class distribution; however, in practice, this assumption is easily violated, as unlabeled data often includes unknown class data, i.e., outliers. The outliers are treated as noise, considerably degrading the performance of SSL models. To address this drawback, we propose a novel framework, diversify and conquer (DAC), to enhance SSL robustness in the context of open-set SSL (OSSL). In particular, we note that existing OSSL methods rely on prediction discrepancies between inliers and outliers from a single model trained on labeled data. This approach can be easily failed when the labeled data are insufficient, leading to performance degradation that is worse than naive SSL that do not account for outliers. In contrast, our approach exploits prediction disagreements among multiple models that are differently biased toward the unlabeled distribution. By leveraging the discrepancies arising from training on unlabeled data, our method enables robust outlier detection, even when the labeled data are underspecified. Our key contribution is constructing a collection of differently biased models through a single training process. By encouraging divergent heads to be differently biased toward outliers while making consistent predictions for inliers, we exploit the disagreement among these heads as a measure to identify unknown concepts. Extensive experiments demonstrate that our method significantly surpasses state-of-the-art OSSL methods across various protocols.