arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2511.04248cs.CL
收藏

基于图标注的高效主题提取:深度模型的轻量级替代方案

Efficient Topic Extraction via Graph-Based Labeling: A Lightweight Alternative to Deep Models

University of Limerick(利默里克大学);Euromed University Of Fez(弗斯欧罗姆大学);University Sidi Mohamed Ben Abdellah(西迪莫哈梅德·本·阿卜杜勒拉赫曼大学)

Salma Mekaoui, Hiba Sofyan, Imane Amaaz, Imane Benchrif, Arsalane Zarghili, Ilham Chaker, Nikola S. Nikolov

AI总结:

本文提出一种基于图的轻量级主题标注方法,通过语义关系扩展主题词并推导标签,在低计算成本下取得优于传统基准、可比 ChatGPT-3.5 的效果。

摘要:

从文本中提取主题已成为一项基本任务,尤其是在非结构化文本数据快速增长的背景下。现有的大多数工作依赖高计算成本的方法来应对这一挑战。在本文中,我们认为概率与统计方法(例如主题建模(TM))可以提供需要更少计算资源的有效替代方案。TM 是一种统计方法,可在大规模无标注文本集合中自动发现主题;然而,它生成的主题是代表性词语的分布,往往缺乏清晰的可解释性。我们的目标是通过为这些词语集合分配有意义的标签来执行主题标注。为了在不依赖计算成本高昂的模型的情况下实现这一目标,我们提出了一种基于图的方法,该方法不仅利用语义相关的术语来丰富主题词,还探索它们之间的关系。通过分析图中的这些连接,我们推导出能够准确捕捉每个主题含义的合适标签。我们在两个不同的数据集上,将所提出的方法与包括 ChatGPT-3.5 在内的多个基准进行了比较研究。在 BERTScore 和余弦相似度方面,我们的方法始终优于传统基准,并且取得了与 ChatGPT-3.5 相当的结果,同时保持了计算效率。最后,我们讨论了主题标注的未来方向,并强调了增强可解释性和自动化程度的潜在研究途径。

英文摘要:

Extracting topics from text has become an essential task, especially with the rapid growth of unstructured textual data. Most existing works rely on highly computational methods to address this challenge. In this paper, we argue that probabilistic and statistical approaches, such as topic modeling (TM), can offer effective alternatives that require fewer computational resources. TM is a statistical method that automatically discovers topics in large collections of unlabeled text; however, it produces topics as distributions of representative words, which often lack clear interpretability. Our objective is to perform topic labeling by assigning meaningful labels to these sets of words. To achieve this without relying on computationally expensive models, we propose a graph-based approach that not only enriches topic words with semantically related terms but also explores the relationships among them. By analyzing these connections within the graph, we derive suitable labels that accurately capture each topic's meaning. We present a comparative study between our proposed method and several benchmarks, including ChatGPT-3.5, across two different datasets. Our method achieved consistently better results than traditional benchmarks in terms of BERTScore and cosine similarity and produced results comparable to ChatGPT-3.5, while remaining computationally efficient. Finally, we discuss future directions for topic labeling and highlight potential research avenues for enhancing interpretability and automation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑