Using Semantic Search to Select and Rank Geological Publications
Main Article Content
Abstract
The aggregation of scientific information is essential for a comprehensive analysis of geological objects. This paper explores the potential and possibilities of semantic search to select thematically similar publications in the geological domain. Various language models are examined in the context of identifying similarities and differences in texts describing mineral deposits. After additional training of language models, a significant improvement in search results is demonstrated. Two web services are presented, based on a method for calculating the semantic similarity between texts and providing a quantitative assessment of their similarity.
Article Details
References
2. Ваш путеводитель по миру NLP (обработке естественного языка). URL: https://habr.com/ru/companies/otus/articles/705482/ (дата обращения 10.09.2024)
3. Semantic similarity. URL: https://en.wikipedia.org/wiki/Semantic_similarity (дата обращения 10.09.2024)
4. Devlin J., Chang M.W., Lee K., Toutanova K. Bert: pre-training of deep bidirectional transformers for language understanding. URL: arxiv.org/abs/1810.04805 (2018)
5. Объясняем простым языком, что такое трансформеры. URL: https://habr.com/ru/companies/mws/articles/770202/ (дата обращения 10.09.2024)
6. BERT (языковая модель). URL: https://neerc.ifmo.ru/wiki/index.php?title= BERT_(языковая_модель) (дата обращения 10.09.2024)
7. BERT community. URL: https://huggingface.co/google-bert (дата обращения 10.09.2024)
8. Akila D., Jayakumar C. Semantic Similarity – A Review of Approaches and Metrics // International Journal of Applied Engineering Research. 2014. Vol. 9, No. 24. P. 27581–27600.
9. TF-IDF. URL: https://ru.wikipedia.org/wiki/TF-IDF (дата обращения 10.09.2024)
10. Okapi BM25. URL: https://ru.wikipedia.org/wiki/Okapi_BM25 (дата обращения 10.11.2024)
11. Word2vec. URL: https://ru.wikipedia.org/wiki/Word2vec (дата обращения 10.09.2024)
12. Краснов Ф.В., Смазневич И.С., Баскакова Е.Н. Проблема потери решений в задаче поиска схожих документов: Применение терминологии при построении векторной модели корпуса // Бизнес-информатика. 2021. Т. 15. № 2. С. 60–74. https://doi.org/10.17323/2587-814X.2021.2.60.74
13. Атаева О.М., Серебряков В.А., Тучкова Н.П. Модель поиска схожих документов в семантической библиотеке // Научный сервис в сети Интернет: труды XXIII Всероссийской научной конференции (20–23 сентября 2021 г., онлайн). М.: ИПМ им. М.В. Келдыша, 2021. С. 54–64. https://doi.org/10.20948/abrau-2021-24
14. Kanerva J., Kitti H., Chang L.-H., Vahtola T., Creutz M., Ginter F. Semantic search as extractive paraphrase span detection // Lang Resources and Evaluation. 2024. https://doi.org/10.1007/s10579-023-09715-7
15. Denghui Yang, Dengyun Zhu, Hailong Gai, Fucheng Wan. Semantic Similarity Calculating based on BERT // Journal of Electrical Systems. 2024. Vol. 20, No. 2. P. 73–79.
16. Kuang M. et al. Multi-task Learning Based Keywords Weighted Siamese Model for Semantic Retrieval // In: Kashima H., Ide T., Peng W/C. (Eds.) Advances in Knowledge Discovery and Data Mining. PAKDD 2023. Lecture Notes in Computer Science. 2023. Vol. 13937. Springer, Cham. https://doi.org/10.1007/978-3-031-33380-4_7
17. Elibrary. URL: https://www.elibrary.ru/defaultx.asp (дата обращения 10.11.2024)
18. Sci-rus-tiny. URL: https://huggingface.co/mlsa-iai-msu-lab/sci-rus-tiny (дата обращения 10.09.2024)
19. ruSciBench — бенчмарк для оценки эмбеддингов научных текстов. URL: https://habr.com/ru/articles/781032/ (дата обращения 10.09.2024).
20. Научная электронная библиотека «КиберЛенинка». URL: https://cyberleninka.ru/ (дата обращения 10.09.2024)
21. Fine-tuning BERT for Semantic Textual Similarity with Transformers in Python. URL: https://thepythoncode.com/article/finetune-bert-for-semantic-textual-similarity-in-python (дата обращения 10.09.2024)
22. Lawley C.J.M., Raimondo S., Chen T., Brin L., Zakharov A., Kur D., Hui J., Newton G., Burgoyne S.L., Marquis G. Geoscience language models and their intrinsic evaluation // Applied Computing and Geosciences. 2022. Vol. 14, 100084. P. 1–10. https://doi.org/10.1016/j.acags.2022.100084
23. Cosine similarity. URL: https://en.wikipedia.org/wiki/Cosine_similarity (дата обращения 10.09.2024)
24. Rubert-tiny2. URL: https://huggingface.co/cointegrated/rubert-tiny2 (дата обращения 10.09.2024)
25. Multilingual-e5-base. URL: https://huggingface.co/intfloat/multilingual-e5-base (дата обращения 10.09.2024)
26. E5-large-en-ru. URL: https://huggingface.co/d0rj/e5-large-en-ru (дата обращения 10.09.2024)
27. Multilingual-e5-large. URL: https://huggingface.co/intfloat/multilingual-e5-large (дата обращения 10.09.2024)
28. Тест производительности. URL: https://ru.wikipedia.org/wiki/Тест_производительности (дата обращения 10.09.2024)
29. Патук М.И., Наумова В.В., Ерёменко В.С. Цифровой репозиторий "geologyscience.ru": открытый доступ к научным публикациям по геологии России // Электронные библиотеки. 2020. Т. 23, № 6. С. 1324–1338. https://doi.org/10.26907/1562-5419-2020-23-6-1324-1338
30. Geologyscience.ru – Виртуальный ассистент – Сервисы с использованием ИИ – Сервисы нейросетевого анализа геологических текстов. URL: https://service.geologyscience.ru/ (дата обращения 10.09.2024)
This work is licensed under a Creative Commons Attribution 4.0 International License.
Presenting an article for publication in the Russian Digital Libraries Journal (RDLJ), the authors automatically give consent to grant a limited license to use the materials of the Kazan (Volga) Federal University (KFU) (of course, only if the article is accepted for publication). This means that KFU has the right to publish an article in the next issue of the journal (on the website or in printed form), as well as to reprint this article in the archives of RDLJ CDs or to include in a particular information system or database, produced by KFU.
All copyrighted materials are placed in RDLJ with the consent of the authors. In the event that any of the authors have objected to its publication of materials on this site, the material can be removed, subject to notification to the Editor in writing.
Documents published in RDLJ are protected by copyright and all rights are reserved by the authors. Authors independently monitor compliance with their rights to reproduce or translate their papers published in the journal. If the material is published in RDLJ, reprinted with permission by another publisher or translated into another language, a reference to the original publication.
By submitting an article for publication in RDLJ, authors should take into account that the publication on the Internet, on the one hand, provide unique opportunities for access to their content, but on the other hand, are a new form of information exchange in the global information society where authors and publishers is not always provided with protection against unauthorized copying or other use of materials protected by copyright.
RDLJ is copyrighted. When using materials from the log must indicate the URL: index.phtml page = elbib / rus / journal?. Any change, addition or editing of the author's text are not allowed. Copying individual fragments of articles from the journal is allowed for distribute, remix, adapt, and build upon article, even commercially, as long as they credit that article for the original creation.
Request for the right to reproduce or use any of the materials published in RDLJ should be addressed to the Editor-in-Chief A.M. Elizarov at the following address: amelizarov@gmail.com.
The publishers of RDLJ is not responsible for the view, set out in the published opinion articles.
We suggest the authors of articles downloaded from this page, sign it and send it to the journal publisher's address by e-mail scan copyright agreements on the transfer of non-exclusive rights to use the work.