Consistent Supervised-Unsupervised Alignment for Generalized Category Discovery

Jizhou Han, Shaokun Wang, Yuhang He, Chenhao Ding, Qiang Wang, Xinyuan Gao, SongLin Dong, Yihong Gong

Keywords: Generalized Category Discovery, Neural Collapse

TL;DR: We propose NC-GCD, a Neural Collapse–inspired framework for generalized category discovery that leverages ETF prototypes and a semantic consistency matcher for consistent alignment, enabling structured features and SOTA novel category accuracy.

Abstract: 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 significantly enhancing novel category accuracy and demonstrating its effectiveness.

Primary Area: Applications (e.g., vision, language, speech and audio, Creative AI)

Submission Number: 19904