METHODOLOGY FOR FORMING A DATABASE OF CHARACTERISTICS OF A COMPLEX TECHNOLOGICAL OBJECT USING LARGE LANGUAGE MODELS
https://doi.org/10.26583/vestnik.2024.5.7
EDN: PJFXNC
Abstract
Nuclear energy plays an important role in ensuring the safety of many countries in the world.When designing and operating complex technological objects (CTO) such as nuclear power plants (NPP), it is critical to take into account their characteristics to ensure safe operation.The relevance of the research topic lies in the need to develop a methodology that can speed up the process of identifying target information contained in scientific publications for nuclear industry enterprises.The lack of scientific papers describing the use of language models for analyzing and extracting characteristics from complex technological objects emphasizes the need for research.In this paper, a NPP is chosen as an example of such an object.To conduct a series of experiments to identify the technical characteristics of the CTO, a list of parameters of the nuclear power plant profile (35 parameters) was compiled and a data set on nuclear power plants was formed (60 scientific publications containing information about the Ling Ao NPP).A program was developed that allows processing the data contained in scientific publications by loading articles into a language model, writing queries and receiving responses for subsequent compilation of a profile of a complex technological object.The results of the work showed that the proposed technique allows programmatic processing of scientific publications to compile a profile of a NPP.
About the Authors
A. R. Matveeva
National Research Nuclear University MEPhI (Moscow Engineering Physics Institute)
Russian Federation
Russian Federation
E. V. Antonov
National Research Nuclear University MEPhI (Moscow Engineering Physics Institute)
Russian Federation
Russian Federation
References
1. Polak M.P., Morgan D. Extracting accurate materials data from research papers with conversational language models and prompt engineering. Nature Communications, 2024. Vol. 15(1), 1569.
2. Yao Y., Duan J., Xu K., Cai Y., Sun Z., Zhang Y. A survey on large language model (LLM) security and privacy: The good, the bad, and the ugly. High-Confidence Computing, 2024. 100211.
3. Sui Y., Zhou M., Zhou M., Han S., Zhang D. Table meets LLM: Can large language models understand structured table data? A benchmark and empirical study. Proceedings of the 17th ACM International Conference on Web Search and Data Mining, 2024. Pp. 645–654.
4. Fedorova A.A. Neobhodimye pravila potrebleniya informacii dlya snizheniya negativnogo vliyaniya informacionnogo obshchestva [Necessary rules for information consumption to reduce the negative impact of the information society]. Skif. Voprosy studencheskoj nauki, 2020. No. 5–1. Pp. 157–162 (in Russian).
5. Jiang A.Q., Sablayrolles A., Mensch A., Bam-ford C., Chaplot D.S., etc. Mistral 7B. arXiv preprint arXiv:2310.06825.
6. Ali A.H., Alajanbi M., Yaseen M.G., Abed S.A. Chatgpt4, DALL· E, Bard, Claude, BERT: Open Possibilities. Babylonian Journal of Machine Learning, 2023. Рp. 17–18.
7. Dubois Y., Li C.X., Taori R., Zhang T., Gulrajani I., Ba J., etc. AlpacaFarm: A simulation framework for methods that learn from human feedback. Advances in Neural Information Processing Systems, 2024. V. 1. https://doi.org/10.48550/arXiv.2305.14387
8. Chaka C. Detecting AI content in responses generated by ChatGPT, YouChat, and Chatsonic: The case of five AI content detection tools. Journal of Applied Learning and Teaching, 2023. Vol. 6(2). Pp. 1–11.
9. Kumratova A.M., Morozova N.V., Vasilenko A.I., Kogaj I.E. Analiz vozmozhnostej nejronnoj seti na osnove yazykovoj modeli GPT-3 i sposoby ee primeneniya na proizvodstve [Analysis of the capabilities of a neural network based on the GPT-3 language model and methods of its application in production]. Vestnik Adygejskogo gosudarstvennogo universiteta. S.4: Estestvenno-matematicheskie i tekhnicheskie nauki, 2023. Iss. 1 (316). Pp. 80–85 (in Russian).
10. Zhan T., Shi C., Shi Y., Li H., Lin Y. Optimization Techniques for Sentiment Analysis Based on LLM (GPT-3). Applied and Computational Engineering, 2024. Vol. 67(1). Pp. 41–47.
11. Yenduri G., Ramalingam M., Selvi G.C., Sup-riya Y., Srivastava G., etc. GPT (generative pre-trained transformer) – a comprehensive review on enabling technologies, potential applications, emerging challenges, and future directions. IEEE Access, 2023. DOI:10.1109/ACCESS.2024.3389497
12. Wang B., Xie Q., Pei J., Chen Z., Tiwari P., Li Z., Fu J. Pre-trained language models in biomedical domain: A systematic survey. ACM Computing Surveys, 2023. Vol. 56(3). Pp. 1–52.
13. Hong Z., Ward, L., Chard, K., Blaiszik, B., Fos¬ter I. Challenges and advances in information extraction from scientific literature: a review. JOM, 2021. Vol. 73(11), Pp. 3383–3400.
14. Han, S., Zhang R. F., Shi L., Richie R., Li, H., Tseng A., etc. Classifying social determinants of health from unstructured electronic health records using deep learning-based natural language processing. Journal of biomedical informatics, 2022. Vol. 127. 103984.
15. Kåhrström F. Natural Language Processing for Swedish Nuclear Power Plants: A study of the challenges of applying Natural language processing in Operations and Maintenance and how BERT can be used in this industry. Visby. Uppsala Universitet, 2022. URL: http://uu.diva-portal.org/smash/get/diva2: 1678697/FULLTEXT01.pdf
Review
For citations:
Matveeva A.R., Antonov E.V. METHODOLOGY FOR FORMING A DATABASE OF CHARACTERISTICS OF A COMPLEX TECHNOLOGICAL OBJECT USING LARGE LANGUAGE MODELS. Vestnik natsional'nogo issledovatel'skogo yadernogo universiteta "MIFI". 2024;13(5):350-357. (In Russ.) https://doi.org/10.26583/vestnik.2024.5.7. EDN: PJFXNC
Views:
141
Email this article 