Keyword Generation for Russian-Language Scientific Texts Using the mT5 Model

IF 0.6 Q4 AUTOMATION & CONTROL SYSTEMS
A. V. Glazkova, D. A. Morozov, M. S. Vorobeva, A. A. Stupnikov
{"title":"Keyword Generation for Russian-Language Scientific Texts Using the mT5 Model","authors":"A. V. Glazkova,&nbsp;D. A. Morozov,&nbsp;M. S. Vorobeva,&nbsp;A. A. Stupnikov","doi":"10.3103/S014641162470041X","DOIUrl":null,"url":null,"abstract":"<p>The authors propose an approach to generate keywords for Russian-language scientific texts using the mT5 (multilingual text-to-text transformer) model, fine-tuned on the Keyphrases CS&amp;Math Russian text corpus. Automatic keyword selection is an urgent task in natural language processing, since keywords help readers search for articles and facilitate the systematization of scientific texts. In this paper, the task of selecting keywords is considered as a task of automatic text abstracting. Additional training of mT5 is carried out on the texts of abstracts of Russian-language scientific articles. The input and output data are abstracts and comma-separated lists of keywords, respectively. The results obtained using mT5 are compared with the results of several basic methods: TopicRank, YAKE!, RuTermExtract, and KeyBERT. The following metrics are used to present the results: F‑measure, ROUGE-1, and BERTScore. The best results on the test sample are obtained using mT5 and RuTermExtract. The highest F-measure is demonstrated by the mT5 model (11.24%), surpassing RuTermExtract by 0.22%. RuTermExtract shows the best result according to the ROUGE-1 metric (15.12%). The best results for BERTScore are also achieved by these two methods: mT5, 76.89% (BERTScore using the mBERT model); RuTermExtract, 75.8% (BERTScore based on ruSciBERT). The authors also assess the ability of mT5 to generate keywords that are not in the source text. The limitations of the proposed approach include the need to form a training sample for additional model training and probably the limited applicability of the additional trained model for texts in other subject areas. The advantages of keyword generation using mT5 are the absence of the need to set fixed values for the length and number of keywords, the need for normalization, which is especially important for inflected languages, and the ability to generate keywords that are not explicitly present in the text.</p>","PeriodicalId":46238,"journal":{"name":"AUTOMATIC CONTROL AND COMPUTER SCIENCES","volume":"58 7","pages":"995 - 1002"},"PeriodicalIF":0.6000,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"AUTOMATIC CONTROL AND COMPUTER SCIENCES","FirstCategoryId":"1085","ListUrlMain":"https://link.springer.com/article/10.3103/S014641162470041X","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"AUTOMATION & CONTROL SYSTEMS","Score":null,"Total":0}
引用次数: 0

Abstract

The authors propose an approach to generate keywords for Russian-language scientific texts using the mT5 (multilingual text-to-text transformer) model, fine-tuned on the Keyphrases CS&Math Russian text corpus. Automatic keyword selection is an urgent task in natural language processing, since keywords help readers search for articles and facilitate the systematization of scientific texts. In this paper, the task of selecting keywords is considered as a task of automatic text abstracting. Additional training of mT5 is carried out on the texts of abstracts of Russian-language scientific articles. The input and output data are abstracts and comma-separated lists of keywords, respectively. The results obtained using mT5 are compared with the results of several basic methods: TopicRank, YAKE!, RuTermExtract, and KeyBERT. The following metrics are used to present the results: F‑measure, ROUGE-1, and BERTScore. The best results on the test sample are obtained using mT5 and RuTermExtract. The highest F-measure is demonstrated by the mT5 model (11.24%), surpassing RuTermExtract by 0.22%. RuTermExtract shows the best result according to the ROUGE-1 metric (15.12%). The best results for BERTScore are also achieved by these two methods: mT5, 76.89% (BERTScore using the mBERT model); RuTermExtract, 75.8% (BERTScore based on ruSciBERT). The authors also assess the ability of mT5 to generate keywords that are not in the source text. The limitations of the proposed approach include the need to form a training sample for additional model training and probably the limited applicability of the additional trained model for texts in other subject areas. The advantages of keyword generation using mT5 are the absence of the need to set fixed values for the length and number of keywords, the need for normalization, which is especially important for inflected languages, and the ability to generate keywords that are not explicitly present in the text.

Abstract Image

使用 mT5 模型为俄语科学文本生成关键词
作者提出了一种使用mT5(多语言文本到文本转换器)模型为俄语科学文本生成关键字的方法,该模型在Keyphrases CS&;Math Russian文本语料库上进行了微调。关键词自动选择是自然语言处理中的一项紧迫任务,因为关键词可以帮助读者搜索文章,促进科学文本的系统化。本文将关键词选择任务视为文本自动文摘的一个任务。mT5对俄语科学文章摘要文本进行了额外的训练。输入和输出数据分别是摘要和以逗号分隔的关键字列表。使用mT5得到的结果与几种基本方法的结果进行了比较:TopicRank, YAKE!、RuTermExtract和KeyBERT。以下指标用于显示结果:F - measure、ROUGE-1和BERTScore。使用mT5和RuTermExtract在测试样品上获得了最好的结果。mT5模型的f值最高(11.24%),比RuTermExtract高出0.22%。根据ROUGE-1度量,RuTermExtract显示出最佳结果(15.12%)。这两种方法对BERTScore也取得了最好的结果:mT5, 76.89%(使用mBERT模型的BERTScore);RuTermExtract, 75.8%(基于ruSciBERT的BERTScore)。作者还评估了mT5生成不在源文本中的关键字的能力。所建议的方法的局限性包括需要为额外的模型训练形成一个训练样本,并且可能对其他主题领域的文本的额外训练模型的有限适用性。使用mT5生成关键字的优点是不需要为关键字的长度和数量设置固定的值,不需要规范化,这对于屈折变化的语言尤其重要,并且能够生成文本中没有显式出现的关键字。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
AUTOMATIC CONTROL AND COMPUTER SCIENCES
AUTOMATIC CONTROL AND COMPUTER SCIENCES AUTOMATION & CONTROL SYSTEMS-
CiteScore
1.70
自引率
22.20%
发文量
47
期刊介绍: Automatic Control and Computer Sciences is a peer reviewed journal that publishes articles on• Control systems, cyber-physical system, real-time systems, robotics, smart sensors, embedded intelligence • Network information technologies, information security, statistical methods of data processing, distributed artificial intelligence, complex systems modeling, knowledge representation, processing and management • Signal and image processing, machine learning, machine perception, computer vision
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信