提出固定几何结构的分类器，实现持续发现新类别时不遗忘旧类别。
GOAL: Geometrically Optimal Alignment for Continual Generalized Category Discovery
- 用固定等角紧框架分类器保持特征对齐一致性。
- 在4个基准上减少16.1%遗忘率，提升3.2%新类识别率。
- 适合长期持续学习中需稳定发现新类别的场景。
持续广义类别发现（C-GCD）要求从无标签数据中识别新类别，同时保留已知类别的知识。现有方法通常动态更新分类器权重，导致遗忘和特征对齐不一致。我们提出GOAL，一种统一框架，引入固定的等角紧框架（ETF）分类器，在整个学习过程中施加一致的几何结构。GOAL对有标签样本进行监督对齐，对新样本采用置信度引导对齐，实现新类别的稳定融合而不破坏旧类知识。在四个基准上的实验表明，相比先前方法Happy，GOAL将遗忘率降低16.1%，新类别发现率提升3.2%，为长时程持续发现提供了强解决方案。
原文摘要 · Abstract (English)
Continual Generalized Category Discovery (C-GCD) requires identifying novel classes from unlabeled data while retaining knowledge of known classes over time. Existing methods typically update classifier weights dynamically, resulting in forgetting and inconsistent feature alignment. We propose GOAL, a unified framework that introduces a fixed Equiangular Tight Frame (ETF) classifier to impose a consistent geometric structure throughout learning. GOAL conducts supervised alignment for labeled samples and confidence-guided alignment for novel samples, enabling stable integration of new classes without disrupting old ones. Experiments on four benchmarks show that GOAL outperforms the prior method Happy, reducing forgetting by 16.1% and boosting novel class discovery by 3.2%, establishing a strong solution for long-horizon continual discovery.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。