大模型知识管理系统

发布时间:2024-04-25 作者:周扬,蔡霈涵,董振江 阅读量:

 

摘要:企业知识管理至关重要,而传统企业知识管理系统存在构建成本高、知识利用率低问题。提出了基于大模型检索增强生成(RAG)技术构建企业知识管理系统的方案。首先介绍了整体方案架构、业务流程与4类知识构建技术,然后重点介绍了检索前处理、知识检索、检索后处理等全流程知识检索技术,并设计了全面的测评框架。经过实践检验,该方案具有知识构建效率高成本低、意图理解精确、知识检索准确等特点与优势。

关键词:RAG;知识管理系统;大模型;知识工程

 

Abstract: Enterprise knowledge management is very important, but traditional enterprise knowledge management systems suffer from high construction costs and low knowledge utilization rate. A scheme to build enterprise knowledge management system based on retrieval-augmented generation (RAG) technology is proposed. Firstly, the overall scheme architecture, business processes and four types of knowledge construction technologies are introduced, and then the whole process of knowledge retrieval technology are discussed, such as retrieval pre-processing, knowledge retrieval, retrieval post-processing, and subsequently a comprehensive evaluation framework is designed. The scheme has the characteristics and advantages of high efficiency and low cost of knowledge construction, accurate intention understanding, and accurate knowledge retrieval.

Keywords: RAG; knowledge management system; large model; knowledge engineering

在线PDF浏览: PDF