A Generic Method for Fine-grained Category Discovery in Natural Language Texts

2024-06-18

Conference on Empirical Methods in Natural Language Processing

AI Summary

For fine-grained text category discovery under coarse-grained supervision only, this paper proposes a semantic similarity-guided unsupervised clustering method. It models fine-grained semantic similarity in log-space and introduces a learnable dynamic centroid inference mechanism, jointly optimizing intra-cluster compactness and inter-cluster separability—thereby overcoming key limitations of conventional contrastive learning, such as neglecting semantic structure and reliance on pre-collected test samples. The method enables real-time inference without requiring a pre-specified number of clusters. Evaluated on three benchmarks, it achieves significant improvements over state-of-the-art methods in Accuracy, Adjusted Rand Index (ARI), and Normalized Mutual Information (NMI). Theoretical guarantees are provided for the proposed framework. Code and datasets are publicly available.

Technology Category

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

Application Category

• Web Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text
• 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

Fine-grained category discovery using only coarse-grained supervision is a cost-effective yet challenging task. Previous training methods focus on aligning query samples with positive samples and distancing them from negatives. They often neglect intra-category and inter-category semantic similarities of fine-grained categories when navigating sample distributions in the embedding space. Furthermore, some evaluation techniques that rely on pre-collected test samples are inadequate for real-time applications. To address these shortcomings, we introduce a method that successfully detects fine-grained clusters of semantically similar texts guided by a novel objective function. The method uses semantic similarities in a logarithmic space to guide sample distributions in the Euclidean space and to form distinct clusters that represent fine-grained categories. We also propose a centroid inference mechanism to support real-time applications. The efficacy of the method is both theoretically justified and empirically confirmed on three benchmark tasks. The proposed objective function is integrated in multiple contrastive learning based neural models. Its results surpass existing state-of-the-art approaches in terms of Accuracy, Adjusted Rand Index and Normalized Mutual Information of the detected fine-grained categories. Code and data are publicly available at https://github.com/changtianluckyforever/F-grained-STAR.

Problem

• Fine-grained category discovery
• Semantic similarity optimization
• Real-time centroid inference

Innovation

• Logarithmic semantic similarity guidance
• Centroid inference for real-time
• Integrated contrastive learning models