姜勇-RAG关键技术及未来趋势发展.pdf

1、姜勇 Dify首席架构师负责 Dify.AI 最佳实践探索及架构设计。一个 90%的 E 人，却喜欢前沿技术折腾，并认为 Code 是最纯粹的事情。在软件工程、服务高可用和数据处理领域有较为丰富的经验，曾独立搭建类 Notion 的笔记型知识库后端服务，超百万用户量；对 RAG 领域有着深刻的理解与实践，曾多次在向量数据库大会、A2M大会、人工智能峰会中进行过相关领域的知识分享。演讲主题：RAG关键技术及未来趋势发展RAG 关键技术及未来趋势发展姜勇Dify.AI 架构师建设 RAG 目前的困境RAG 发展史E En nt te er rp pr ri is se e R RA AG GRAG

2、 的展望RAG 难点from:https:/arxiv.org/html/2401.05856v1R RA AG G 难难点点FP1 Missing ContentFP2 Missed the Top Ranked DocumentsFP3 Not in Context-Consolidation strategy Limitations FP4 Not Extracted FP5 Wrong FormatFP6 Incorrect SpecificityFP7 IncompleteRAG 发展史第第一一阶阶段段：B Ba as si ic c R RA AG GRetrieveAnswerb

3、ased on vector searchBasic RAGRAG 发展史based on vector search第二阶段：Advanced RAGHybrid RetrieveAnswerResult ProcessA Ad dv va an nc ce ed d R RA AG GAdvanced RAGQuery typeKeywordNDCG3VectorNDCG3HybridNDCG3Hybrid+Semantic RankerNDCG3Concept seeking queries39.045.846.359.6Fact seeking queries37.849.049.16

4、3.4Exact snippet search51.141.551.060.8Web search-like queries41.846.350.058.9Keyword queries79.211.761.066.9Low query term overlap23.036.135.949.1Queries with misspellings28.839.140.654.6Long queries42.741.648.159.4Medium queries38.144.746.759.9Short queries53.138.853.063.9Advanced RAGAdvanced RAG“

5、G Ga ar rb ba ag ge e I In n G Ga ar rb ba ag ge e O Ou ut t.”Advanced RAG类类型型示示例例问题与语料不相关询问产品配置库关于货物运输的问题问题模糊“这篇文章的作者是谁？不是关于事实召回“总结一下这篇文章的主要内容”包含多个子问题今年的欧洲杯在哪里举办，什么时候开始?需要多跳逻辑“Who won the 2023 super bowl and where was their head coach from?”包含非语义组成（结构化）“What are movies about aliens in 1980”-should

6、filter by year=1980问题包含比较“江苏省房贷利率政策从2022年至2023年有哪些调整？”Advanced RAGbased on vector searchHybird RetrieveAnswerResult ProcessQuery TransformAdvanced RAGAdvanced RAG类类型型解解决决方方案案问题与语料不相关检索前增加问题分类或检查步骤，如查询路由（Query routing）问题模糊基于历史的问题重写（Rewrite）不是关于事实召回（总结）索引过程中实现摘要（Summary Index）长文本窗口