IOCC: Aligning Semantic and Cluster Centers for Few-shot Short Text Clustering

2025-08-08

AI Summary

Short text clustering suffers from misalignment between semantic centers and cluster centers, leading to biased representation learning. To address this, we propose IOCC: first, Interaction-Enhanced Optimal Transport (IEOT) dynamically generates high-confidence pseudo-labels and constructs pseudo-centers to align semantic and cluster structures; second, Center-Aware Contrastive Learning (CACL) refines the representation space under pseudo-center guidance. IOCC is the first framework to integrate optimal transport with semantic interaction modeling, establishing a dual-module cooperative architecture that significantly improves clustering quality and stability in few-shot settings. Extensive experiments on eight benchmark datasets demonstrate consistent superiority over state-of-the-art methods, with relative improvements of up to 7.34% on biomedical datasets. The approach further exhibits high efficiency and robustness.

Technology Category

• Search and Optimization: Learning to Search
• Machine Learning: Clustering
• Computer Vision: Representation Learning for Vision

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

In clustering tasks, it is essential to structure the feature space into clear, well-separated distributions. However, because short text representations have limited expressiveness, conventional methods struggle to identify cluster centers that truly capture each category's underlying semantics, causing the representations to be optimized in suboptimal directions. To address this issue, we propose IOCC, a novel few-shot contrastive learning method that achieves alignment between the cluster centers and the semantic centers. IOCC consists of two key modules: Interaction-enhanced Optimal Transport (IEOT) and Center-aware Contrastive Learning (CACL). Specifically, IEOT incorporates semantic interactions between individual samples into the conventional optimal transport problem, and generate pseudo-labels. Based on these pseudo-labels, we aggregate high-confidence samples to construct pseudo-centers that approximate the semantic centers. Next, CACL optimizes text representations toward their corresponding pseudo-centers. As training progresses, the collaboration between the two modules gradually reduces the gap between cluster centers and semantic centers. Therefore, the model will learn a high-quality distribution, improving clustering performance. Extensive experiments on eight benchmark datasets show that IOCC outperforms previous methods, achieving up to 7.34% improvement on challenging Biomedical dataset and also excelling in clustering stability and efficiency. The code is available at: https://anonymous.4open.science/r/IOCC-C438.

Problem

• Aligning cluster centers with semantic centers for few-shot short text clustering
• Addressing limited expressiveness in short text representations for clustering
• Reducing gap between cluster distributions and underlying semantic categories

Innovation

• Interaction-enhanced Optimal Transport for pseudo-labels
• Center-aware Contrastive Learning optimizes representations
• Aligns cluster centers with semantic centers gradually