Application of Large Language Model-Driven Pairwise Constrained Clustering Fusion Algorithm in New Intent Discovery
Authors:
- Feiyang ZhangFZFeiyang Zhang
- Qimeng YangQYQimeng Yang
- Xingge RenXRXingge Ren
- Jinmiao SongJSJinmiao Song
- Xinjun PeiXPXinjun Pei
Preprint
- 11 March 2026
- preprint
- Published by Elsevier in SSRN Electronic Journal
Abstract
In open-world dialogue systems, new intent discovery (NID) aims to identify previously unseen user intents from raw utterances and expand the supported intent set. This process is traditionally time-consuming and often requires domain experts to explore data and define intent labels, which substantially increases the cost. To address this challenge, we propose SCNID, a semi-supervised framework that integrates contrastive representation learning with LLM-driven pairwise-constraint-guided clustering fusion. Specifically, SCNID learns discriminative intent representations via semi-supervised contrastive learning, and then performs clustering fusion by combining a Gaussian Mixture Model (GMM) with K-means++ to improve robustness and stability. Moreover, we leverage a large language model (LLM) to generate Must-Link and Cannot-Link constraints, providing structured semantic guidance for clustering. Extensive experiments on four benchmarks---three intent discovery datasets (Clinc150, Banking77, and StackOverflow20) and an additional cross-domain topic dataset (DBPedia14)---demonstrate that SCNID achieves strong and stable performance under diverse data distributions, validating the effectiveness of the proposed approach.Keywords
This publication has 14 references indexed in Scilit:
- Uncertainty-Aware Contrastive Learning for semi-supervised named entity recognitionKnowledge-Based Systems, 2024
- A Clustering Framework for Unsupervised and Semi-Supervised New Intent DiscoveryIEEE Transactions on Knowledge and Data Engineering, 2023
- Discovering New Intents with Deep Aligned ClusteringProceedings of the AAAI Conference on Artificial Intelligence, 2021
- Discovering New Intents via Constrained Deep Adaptive Clustering with Cluster RefinementProceedings of the AAAI Conference on Artificial Intelligence, 2020
- Efficient Intent Detection with Dual Sentence Encoders2020
- Semi-supervised deep embedded clusteringNeurocomputing, 2019
- Determinants of FDI in developed and developing countries: a quantitative analysis using GMMJournal of Economic Studies, 2018
- Spectrum of Restrictive and Infiltrative CardiomyopathiesJACC, 2018
- Estimation of KL Divergence: Optimal Minimax RateIEEE Transactions on Information Theory, 2018
- A feature reduction and unsupervised classification algorithm for multispectral dataPattern Recognition, 1984