GLEAN: Generalized Category Discovery with Diverse and Quality-Enhanced LLM Feedback

2025-02-25

AI Summary

Generalized Category Discovery (GCD) aims to jointly identify both known and unknown categories in unlabeled data using only a few labeled examples from known classes; however, existing methods suffer from weak semantic mining, difficulty in error correction, and reliance on additional human annotations. This paper proposes the first active learning framework integrating diversity-aware sampling with high-quality large language model (LLM) feedback. It employs three synergistic mechanisms: instance-level contrastive enhancement, class-level descriptive generation, and uncertainty-driven semantic alignment—thereby optimizing cluster structure while enabling interpretable semantic understanding. The method unifies contrastive learning, multi-granularity LLM feedback (discriminative, generative, and retrieval-based), and semantics-guided adaptive clustering. Extensive experiments across multiple benchmarks and low-supervision settings demonstrate significant improvements over state-of-the-art methods, achieving high-accuracy category discovery and intrinsic semantic interpretability—without any human annotation.

Technology Category

• Machine Learning: Active Learning
• Search and Optimization: Learning to Search
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Generalized Category Discovery (GCD) is a practical and challenging open-world task that aims to recognize both known and novel categories in unlabeled data using limited labeled data from known categories. Due to the lack of supervision, previous GCD methods face significant challenges, such as difficulty in rectifying errors for confusing instances, and inability to effectively uncover and leverage the semantic meanings of discovered clusters. Therefore, additional annotations are usually required for real-world applicability. However, human annotation is extremely costly and inefficient. To address these issues, we propose GLEAN, a unified framework for generalized category discovery that actively learns from diverse and quality-enhanced LLM feedback. Our approach leverages three different types of LLM feedback to: (1) improve instance-level contrastive features, (2) generate category descriptions, and (3) align uncertain instances with LLM-selected category descriptions. Extensive experiments demonstrate the superior performance of MethodName over state-of-the-art models across diverse datasets, metrics, and supervision settings. Our code is available at https://github.com/amazon-science/Glean.

Problem

• Recognize known and novel categories in unlabeled data
• Reduce reliance on costly human annotations
• Enhance semantic understanding of discovered clusters

Innovation

• Leverages diverse LLM feedback
• Enhances instance-level features
• Aligns instances with descriptions