基于LLM和RAG的水利知识问答系统研究  

Research on water conservancy knowledge question answering system based on LLM and RAG

作　　者：黄哲远 马帅森 程辉 王红霞[1] 俞佳辉 郑子龙 夏一凡 甘青青 袁潇雅 HUANG Zheyuan;MA Shuaisen;CHENG Hui;WANG Hongxia;YU Jiahui;ZHENG Zilong;XIA Yifan;GAN Qingqing;YUAN Xiaoya(School of Computer Science and Technology,Zhejiang University of Water Resources and Electric Power,Hangzhou 310000,China)

机构地区：[1]浙江水利水电学院计算机科学与技术学院,杭州310000

出　　处：《计算机应用文摘》2025年第8期137-140,共4页

基　　金：2024年浙江水利水电学院大学生创新创业训练计划项目:基于大语言模型的水利知识智能问答系统(202411481054)。

摘　　要：水利高质量发展对水利科技提出了新的要求,进一步推动了水利知识问答服务的创新需求。基于水利知识服务的导向,构建了一个基于前沿大语言模型技术的水利知识智能问答系统。通过整合大量水利行业数据,构建了全面的水利知识库,利用大语言模型理解用户复杂的自然语言输入,并结合信息检索和文本生成技术,提高了答案的准确性和实用性。该系统能够为用户提供更丰富、更灵活的回答。经测试验证,基于大语言模型和检索增强生成技术实现的水利知识智能问答系统在实际应用中可行,为水利行业的智能化管理提供了宝贵的参考。The high-quality development of water conservancy has put forward new requirements for water conservancy technology,further promoting the innovative demand for water conservancy knowledge Q&A services.Based on the guidance of water conservancy knowledge services,a water conservancy knowledge intelligent question answering system based on cutting-edge big language model technology has been constructed.By integrating a large amount of data from the water conservancy industry,a comprehensive knowledge base for water conservancy has been constructed.Large language models are used to understand users̓complex natural language inputs,and information retrieval and text generation technologies are combined to improve the accuracy and practicality of answers.This system can provide users with richer and more flexible answers.After testing and verification,the water conservancy knowledge intelligent question answering system based on big language model and retrieval enhanced generation technology is feasible in practical applications,providing valuable reference for the intelligent management of the water conservancy industry.

关 键 词：水利工程 大语言模型 知识问答 智能系统 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]