通过统一优化目标提升未知类别发现准确率
Consistent Supervised-Unsupervised Alignment for Generalized Category Discovery
- 用固定等角紧框架原型构建统一几何结构
- 在多个基准上显著提升未知类别的识别准确率
- 适合需要稳定发现新类别的场景
广义类别发现（GCD）旨在对已知类别进行分类的同时，从无标签数据中发现新类别。然而，现有方法因优化目标不一致和类别混淆，导致特征重叠，影响新类别的性能。为此，我们提出受神经坍缩启发的广义类别发现框架（NC-GCD）。通过预先设定并固定等角紧框架（ETF）原型，确保已知与未知类别具有最优几何结构和一致的优化目标。引入一致ETF对齐损失，统一监督与无监督的ETF对齐，增强类别可分性。同时设计语义一致性匹配器（SCM），保持聚类迭代中标签分配的稳定一致。该方法在多个GCD基准上表现优异，显著提升新类别准确率，验证了有效性。
原文摘要 · Abstract (English)
Generalized Category Discovery (GCD) focuses on classifying known categories while simultaneously discovering novel categories from unlabeled data. However, previous GCD methods face challenges due to inconsistent optimization objectives and category confusion. This leads to feature overlap and ultimately hinders performance on novel categories. To address these issues, we propose the Neural Collapse-inspired Generalized Category Discovery (NC-GCD) framework. By pre-assigning and fixing Equiangular Tight Frame (ETF) prototypes, our method ensures an optimal geometric structure and a consistent optimization objective for both known and novel categories. We introduce a Consistent ETF Alignment Loss that unifies supervised and unsupervised ETF alignment and enhances category separability. Additionally, a Semantic Consistency Matcher (SCM) is designed to maintain stable and consistent label assignments across clustering iterations. Our method achieves strong performance on multiple GCD benchmarks, significantly enhancing novel category accuracy and demonstrating its effectiveness.
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。