Prior-Constrained Association Learning for Fine-Grained Generalized Category Discovery

2025-02-13

AI Summary

This paper addresses Generalized Category Discovery (GCD), a task requiring fine-grained clustering of mixed data where known classes are fully labeled while unknown classes are entirely unlabeled. To tackle challenges including uncontrolled semantic distribution and ambiguous class boundaries in unlabeled data, we propose a Prior-Embedded Association Learning framework. First, category priors are explicitly incorporated throughout graph construction and optimization to enforce semantic grouping constraints. Second, a prior-guided non-parametric prototype contrastive mechanism is designed to jointly optimize the parametric classifier and non-parametric prototype representations. Third, a self-distillation-enhanced dual-path classification architecture is introduced to improve discriminative robustness. Extensive experiments on multiple GCD benchmarks demonstrate significant improvements over state-of-the-art methods, particularly in fine-grained separation of unknown categories and cross-category discrimination capability.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Search and Optimization: Mixed Discrete/Continuous Search
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

This paper addresses generalized category discovery (GCD), the task of clustering unlabeled data from potentially known or unknown categories with the help of labeled instances from each known category. Compared to traditional semi-supervised learning, GCD is more challenging because unlabeled data could be from novel categories not appearing in labeled data. Current state-of-the-art methods typically learn a parametric classifier assisted by self-distillation. While being effective, these methods do not make use of cross-instance similarity to discover class-specific semantics which are essential for representation learning and category discovery. In this paper, we revisit the association-based paradigm and propose a Prior-constrained Association Learning method to capture and learn the semantic relations within data. In particular, the labeled data from known categories provides a unique prior for the association of unlabeled data. Unlike previous methods that only adopts the prior as a pre or post-clustering refinement, we fully incorporate the prior into the association process, and let it constrain the association towards a reliable grouping outcome. The estimated semantic groups are utilized through non-parametric prototypical contrast to enhance the representation learning. A further combination of both parametric and non-parametric classification complements each other and leads to a model that outperforms existing methods by a significant margin. On multiple GCD benchmarks, we perform extensive experiments and validate the effectiveness of our proposed method.

Problem

• Clustering unlabeled data with known categories
• Utilizing cross-instance similarity for semantics discovery
• Incorporating prior knowledge into association learning process

Innovation

• Prior-constrained Association Learning
• Non-parametric prototypical contrast
• Combination of parametric, non-parametric classification