LLM-MemCluster: Empowering Large Language Models with Dynamic Memory for Text Clustering

2025-11-19

AI Summary

Existing LLM-based text clustering approaches lack stateful memory and granularity control, rely heavily on complex external modules, and thus fail to achieve end-to-end semantic clustering. This work reframes clustering as a native LLM task: we propose a dynamic memory mechanism to preserve iterative optimization states and design a dual-prompt strategy that autonomously infers the optimal number of clusters. To our knowledge, this is the first unsupervised, zero-shot, end-to-end clustering method for LLMs that requires no fine-tuning and no external components. The approach supports adaptive granularity inference and offers strong interpretability. Extensive experiments across multiple benchmark datasets demonstrate consistent and significant improvements over strong baselines, with superior clustering quality and robustness.

Technology Category

• Machine Learning: Clustering
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Search and Optimization: Learning to Search

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Large Language Models (LLMs) are reshaping unsupervised learning by offering an unprecedented ability to perform text clustering based on their deep semantic understanding. However, their direct application is fundamentally limited by a lack of stateful memory for iterative refinement and the difficulty of managing cluster granularity. As a result, existing methods often rely on complex pipelines with external modules, sacrificing a truly end-to-end approach. We introduce LLM-MemCluster, a novel framework that reconceptualizes clustering as a fully LLM-native task. It leverages a Dynamic Memory to instill state awareness and a Dual-Prompt Strategy to enable the model to reason about and determine the number of clusters. Evaluated on several benchmark datasets, our tuning-free framework significantly and consistently outperforms strong baselines. LLM-MemCluster presents an effective, interpretable, and truly end-to-end paradigm for LLM-based text clustering.

Problem

• Enables stateful memory for iterative text clustering refinement
• Solves cluster granularity management through dual-prompt reasoning
• Eliminates external modules for truly end-to-end clustering

Innovation

• Dynamic Memory enables state awareness for clustering
• Dual-Prompt Strategy determines optimal cluster count
• Tuning-free framework provides end-to-end clustering solution