Generalized Category Discovery in Event-Centric Contexts: Latent Pattern Mining with LLMs

2025/05/29 by Yi Luo, Luo, Yi, Qiwen Wang +13
Computer Science · #Computation and Language (cs.CL) #FOS: Computer and information sciences #Imbalanced Data Classification Techniques #Text and Document Classification Technologies #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2505.23304

openalex publication_date 2025/05/29 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Generalized Category Discovery (GCD) aims to classify both known and novel categories using partially labeled data that contains only known classes. Despite achieving strong performance on existing benchmarks, current textual GCD methods lack sufficient validation in realistic settings. We introduce Event-Centric GCD (EC-GCD), characterized by long, complex narratives and highly imbalanced class distributions, posing two main challenges: (1) divergent clustering versus classification groupings caused by subjective criteria, and (2) Unfair alignment for minority classes. To tackle these, we propose PaMA, a framework leveraging LLMs to extract and refine event patterns for improved cluster-class alignment. Additionally, a ranking-filtering-mining pipeline ensures balanced representation of prototypes across imbalanced categories. Evaluations on two EC-GCD benchmarks, including a newly constructed Scam Report dataset, demonstrate that PaMA outperforms prior methods with up to 12.58% H-score gains, while maintaining strong generalization on base GCD datasets.

Citations

• GLEAN: Active Generalized Category Discovery with Diverse LLM Feedback
• Qwen2.5 Technical Report
• Happy: A Debiased Learning Framework for Continual Generalized Category Discovery
• Transfer and Alignment Network for Generalized Category Discovery
• Generalized Category Discovery with Large Language Models in the Loop
• DeCrisisMB: Debiased Semi-Supervised Learning for Crisis Tweet Classification via Memory Bank
• Towards Distribution-Agnostic Generalized Category Discovery
• Dynamic Conceptional Contrastive Learning for Generalized Category Discovery
• Generalized Category Discovery with Decoupled Prototypical Network
• Parametric Classification for Generalized Category Discovery: A Baseline Study
• Watch the Neighbors: A Unified K-Nearest Neighbor Contrastive Learning Framework for OOD Intent Discovery
• New Intent Discovery with Pre-training and Contrastive Learning
• A Comparative Survey of Deep Active Learning
• Generalized Category Discovery
• Batch Active Learning at Scale
• Mind Your Outliers! Investigating the Negative Impact of Outliers on Active Learning for Visual Question Answering
• Neighborhood Contrastive Learning for Novel Class Discovery
• On the Importance of Effectively Adapting Pretrained Language Models for Active Learning
• OVANet: One-vs-All Network for Universal Domain Adaptation
• Active Sentence Learning by Adversarial Uncertainty Sampling in Discrete Space
• Efficient Intent Detection with Dual Sentence Encoders
• An Evaluation Dataset for Intent Classification and Out-of-Scope Prediction
• Representation Learning with Contrastive Predictive Coding
• Prototypical Networks for Few-shot Learning
• Deep Bayesian Active Learning with Image Data
• Unsupervised Deep Embedding for Clustering Analysis
• BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

Related

• GLEAN: Active Generalized Category Discovery with Diverse LLM Feedback
• Parametric Classification for Generalized Category Discovery: A Baseline Study
• Generalized Categories Discovery for Long-tailed Recognition
• ImbaGCD: Imbalanced Generalized Category Discovery
• Generalized Category Discovery via Reciprocal Learning and Class-Wise Distribution Regularization