AI summaryⓘ
The authors address the problem of continually learning to recognize new categories from unlabeled data without forgetting old ones, a challenge known as Continual Generalized Category Discovery (C-GCD). They propose a method that temporarily groups uncertain samples into 'virtual categories' to avoid errors from incorrect labeling, allowing safer learning from new data. To improve how well the model distinguishes between old and new classes, the authors add a technique that uses relationships between data points to create clearer category separations. Their approach performs better than previous methods on several image datasets, offering a practical way to learn new visual categories over time.
Continual LearningGeneralized Category DiscoveryPseudo-labelingVirtual Category LearningContrastive LearningRepresentation LearningUnlabeled DataClass SeparationImage ClassificationOpen-world Learning
Authors
Jiahui Xiong, Qiuxia Lai, Hongsong Wang
Abstract
Continual Generalized Category Discovery (C-GCD) aims to incrementally identify novel categories from sequential unlabeled data while preserving recognition of known classes, which is an essential capability for open-world visual learning. A major bottleneck lies in ambiguous unlabeled samples that cannot be confidently assigned to known classes nor reliably grouped as novel ones, making pseudo-labeling brittle and often biasing learning toward familiar categories. In this work, we introduce Virtual Category-Guided Continual Generalized Category Discovery by adapting Virtual Category Learning (VCL) to the continual setting. Our method identifies uncertain samples and assigns them to temporary virtual categories, enabling safe and informative learning from unlabeled streams without injecting noisy labels, while improving unlabeled data utilization and mitigating prediction bias. To further stabilize discovery across sessions and enhance class separation, we augment VCL with Expanded Neighborhood Contrastive Learning (ENCL), which exploits extended neighborhood relations and an adaptive margin to learn more discriminative and well-separated representations for both old and emerging classes. Extensive experiments on CIFAR-100, Tiny ImageNet, and ImageNet-100 demonstrate that our approach consistently outperforms state-of-the-art methods, establishing a scalable and effective solution for C-GCD.