Article Details

Title · Authors · Abstract · Keywords · DOI · References

Intelligent Question-answering System for FPSO Production Process Based on Model Fine-tuning Integrated with RAG

WU Binfeng1, DONG Haijie2, SUN Kecheng2, LIU Xuesong2, LIU Donghui2

(1.Oil Production Service Shenzhen Branch of CNOOC Energy Technology and Service-Oil Production Services Company, Shenzhen 518054, China; 2.CNOOC Energy Technology and Service-Oil Production Services Company, Tianjin 300452, China)

Section：Intelligent Manufacturing · Vol. 2025, Issue 10 · Pages：170-175 · 2096-4706（2025）10-0170-06 · CLC number：TP181；TP311.1 · Document code：A · DOI：10.19850/j.cnki.2096-4706.2025.10.031

Abstract

In recent years, Large Language Models have gradually gained widespread application by virtue of their powerful text generation and comprehension capabilities. However, although Large Language Models possess strong general abilities, the phenomenon of “hallucination” still exists in the professional knowledge domain. Based on the professional knowledge base of offshore platforms, this paper utilizes Retrieval-Augmented Generation and efficient parameter fine-tuning methods to construct an intelligent question-answering system under process steady state based on a Large Language Model of vertical knowledge domain. Compared with General Large Language Models, Large Language Models of vertical knowledge domain focus more on downstream tasks in specific domains and possess higher professionalism and practicality.

Keywords

Retrieval-Augmented Generation; Machine Learning; fine-tuning; intelligent question-answering; Large Language Model

References

  1. [1] BROWN T B，MANN B，RYDER N，et al. Language Models are Few-Shot Learners [J/OL].arXiv:2005.14165 [cs.CL].（2020-07-22）.https://arxiv.org/abs/2005.14165?context=cs.
  2. [2] Xu Zhiwei, Li Hailong, Li Bo, etc Overview of AIGC Large Model Evaluation: Enabling Technologies, Security Hazards, and Responses [J]. Computer Science and Exploration, 2024, 18 (9): 2293-2325
  3. [3] Yue Zengying, Ye Xia, Liu Ruiheng A review of pre training techniques based on language models [J]. Chinese Journal of Information Science, 2021, 35 (9): 15
  4. [4] Ma Wuren, Gong Mengchun, Dai Hui, etc A review of the application of large language models represented by ChatGPT in clinical medicine [J]. Journal of Medical Informatics, 2023, 44 (7): 9-17
  5. [5] Kou Sijia, Yan Fengyun, Ma Jing Application of Domestic Large Language Models in Automatic Annotation of Subject Knowledge Graph: Taking Ethics and Rule of Law and Mathematics as Examples [J]. Journal of East China Normal University: Natural Science Edition, 2024 (5): 81-92
  6. [6] LEWIS P，PEREZ E，PIKTUS A，et al. Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks [J/OL].arXiv:2005.11401 [cs.CL].（2021-04-12）.https://arxiv.org/abs/2005.11401v4?file=2005.11401v4.
  7. [7] Tian Yonglin, Wang Xingxia, Wang Yutong, etc RAG-PHI: Retrieval Enhanced Generative Driven Parallel Humans and Parallel Intelligence [J]. Journal of Intelligent Science and Technology, 2024, 6 (1): 41-51
  8. [8] Zhang Tingting, Ma Mingdong, Wang Deyu Research on OCR Text Recognition Technology [J]. Computer Technology and Development, 2020, 30 (4): 85-88
  9. [9] GLM T，ZENG A，XU B，et al. ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools [J/OL].arXiv:2406.12793 [cs.CL].（2024-07-30）.https://arxiv.org/abs/2406.12793.
  10. [10] Liu Haiyang, Huang Dan, Wang Zhihai A Local Collaborative Filtering Ranking Recommendation Algorithm [J]. Journal of Shandong University: Engineering Edition, 2016, 46 (5): 29-36

About the authors

Wu Binfeng (1989-), male, Han ethnicity, from Xiantao, Hubei, is an intermediate engineer with a master's degree. His research interests include intelligent question answering using big language models, the application of artificial intelligence in enterprises, and engineering management.