“RAG in a Box”：简化和加速数据搜索 [LRN3302].pdf

1、The following is intended to outline our general product direction.It is intended for information purposes only,and may not be incorporated into any contract.It is not a commitment to deliver any material,code,or functionality,and should not be relied upon in making purchasing decisions.The developm

2、ent,release,timing,and pricing of any features or functionality described for Oracles products may change and remains at the sole discretion of Oracle Corporation.Safe harbor statement1Copyright 2025,Oracle and/or its affiliatesRAG in a BoxSimplify and Accelerate Data SearchesProduct Manager for AI

3、Vector Search and Private AI Services ContainerDoug HoodVice President AI&Data Platform StrategyMassimo CastelliRetrieval Augmented Generation(RAG)Enables LLMs to use business data to produce better and more contextually relevant answers to user questions while keeping business data secure3Copyright

4、 2025,Oracle and/or its affiliatesTypical RAG PipelineOracle Copyright 2025,Oracle and/or its affiliatesUser InteractionLLM invocation and storage in a Vector DB,e.g.Pinecone,Milvus,etc.EmbeddingsChunked text+metadataSetup(frequency varies based on use case)LLM invocation to create embeddings from t

5、he user question and compare vector distance from Vector DB info.Eventual rerank.RetrievalUser QueryLLM invocation with local Context.AnswerLocal knowledge as contextSystem ResponseChunkingPython running OSS frameworks,e.g.,langchain,nlmatics,etc.DocumentsOracle Copyright 2025,Oracle and/or its affi

6、liates4Typical RAG PipelineOracle Copyright 2025,Oracle and/or its affiliatesUser InteractionLLM invocation and storage in a Vector DB,e.g.Pinecone,Milvus,etc.EmbeddingsChunked text+metadataSetup(frequency varies based on use case)LLM invocation to create embeddings from the user question and compar