Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Extended RAG – mit Langchains Multi-Retriever-Ansatz zu Fragen mit AI die beste Antwortquelle auswählen

Extended RAG – mit Langchains Multi-Retriever-Ansatz zu Fragen mit AI die beste Antwortquelle auswählen

Retrieval Augmented Generation (RAG) verwendet Daten aus Retrievern wie Vektor-Datenbanken, damit die relevanten Informationen gefunden werden und Benutzerfragen beantwortet werden können.

Wenn mehrere Retriever (z.B. Support-Tickets, Produkte, Kundenliste) verwendet werden sollen, kann die Auswahl der optimalen Datenquelle zur gestellten Frage eine Herausforderung sein. Marco Frodl wird das MultiRouteChain-Paradigma aus dem LangChain-Framework vorstellen, mit dem die AI dynamisch die Auswahl der besten Antwortquelle übernimmt. Die Live-Coding-Demonstration wird zeigen, wie MultiRouteChain die Leistung von RAG für das Beantworten von Benutzeranfragen verbessert.

Marco Frodl

May 30, 2024
Tweet

More Decks by Marco Frodl

Other Decks in Programming

Transcript

  1. Generative AI Infodays Bonn 27.5.-29.5.24 Extended RAG – mit LangChains

    Multi- Retriever-Ansatz zu Fragen mit AI die beste Antwortquelle auswählen Marco Frodl [email protected] Principal Consultant for Generative AI @marcofrodl
  2. Why is it important? Generative AI Extended RAG - AI-based

    Retriever Selection Generative AI AI understands and generates natural language AI can access knowledge from the training phase
  3. Generative AI Extended RAG - AI-based Retriever Selection What is

    RAG? https://aws.amazon.com/what-is/retrieval-augmented-generation/ RAG = Ingestion + Retrieval
  4. Generative AI Extended RAG - AI-based Retriever Selection About Me

    Marco Frodl Principal Consultant for Generative AI Thinktecture AG X: @marcofrodl E-Mail: [email protected] https://www.thinktecture.com/thinktects/marco-frodl/
  5. Ingestion Generative AI Extended RAG - AI-based Retriever Selection Simple

    RAG in a nutshell Splitted (smaller) parts Embedding- Model Embedding 𝑎 𝑏 𝑐 … Vector- Database Document Metadata: Reference to original document
  6. Similarity search in a Vector DB Generative AI Extended RAG

    - AI-based Retriever Selection Simple RAG in a nutshell
  7. Ingestion++ HyQE: Hypothetical Question Embedding Generative AI Extended RAG -

    AI-based Retriever Selection Simple Advanced RAG in a nutshell LLM, e.g. GPT-3.5-turbo Transformed document Write 3 questions, which are answered by the following document. Chunk of Document Embedding- Model Embedding 𝑎 𝑏 𝑐 … Vector- Database Metadata: content of original chunk
  8. Ask me anything Generative AI Extended RAG - AI-based Retriever

    Selection Simple RAG Question Prepare Search Search Results Question Answer LLM Vector DB Embedding Model Question as Vector Workflow Terms - Retriever - Chain Elements Embedding- Model Vector- DB Python LLM LangChain
  9. Similarity search in a Vector DB – Limits with K

    Generative AI Extended RAG - AI-based Retriever Selection Simple RAG in a nutshell
  10. Similarity search in a Vector DB – Threshold Generative AI

    Extended RAG - AI-based Retriever Selection Simple RAG in a nutshell
  11. Just one Vector DB/Retriever? • Multiple Generative AI-Apps • Scaling

    and Hosting • Query Parameter per Retriever • Prompts per Retriever • Fast Updates & Re-Indexing • Access Rights • Custom Retriever Generative AI Extended RAG - AI-based Retriever Selection What’s wrong with Simple RAG? ✅ ✅ ✅ On-Premise AI-Apps Cloud Docs Public Tickets Features Website Sales Docs Internal Tickets
  12. Best source determination before the search Generative AI Extended RAG

    - AI-based Retriever Selection Advanced RAG Question Retriever Selection 0-N Search Results Question Answer LLM Embedding Model Vector DB A Question as Vector Vector DB B LLM Prepare Search or
  13. Best source determination before the search Generative AI Extended RAG

    - AI-based Retriever Selection Advanced RAG Retriever Selection LLM Vector DB A Vector DB B or
  14. Best source determination before the search Generative AI Extended RAG

    - AI-based Retriever Selection Advanced RAG Question Retriever Selection 0-N Search Results Question Answer LLM Embedding Model Vector DB A Question as Vector Vector DB B LLM Prepare Search or Question Prepare Search Search Results Question Answer LLM Vector DB Embedding Model Question as Vector
  15. Dynamic Retriever Selection with AI Generative AI Extended RAG -

    AI-based Retriever Selection Advanced RAG
  16. Dynamic Retriever Selection with AI Generative AI Extended RAG -

    AI-based Retriever Selection Advanced RAG
  17. Dynamic Retriever Selection with AI Generative AI Extended RAG -

    AI-based Retriever Selection Advanced RAG
  18. Dynamic Retriever Selection with AI Generative AI Extended RAG -

    AI-based Retriever Selection Advanced RAG
  19. Dynamic Retriever Selection with AI Generative AI Extended RAG -

    AI-based Retriever Selection Advanced RAG