Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI Marco Frodl @marcofrodl Principal Consultant for Generative AI

Why is it important? Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI Generative AI User Input AI Processing Generated Output OpenAI GPTx DALL-E 3 GPT-4 Vision Whisper Text Text Text-to-Speech (TTS-1)

Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI What is RAG? https://aws.amazon.com/what-is/retrieval-augmented-generation/ RAG = Ingestion + Retrieval P

Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI Demo: RAG

Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI About Me Marco Frodl Principal Consultant for Generative AI Thinktecture AG X: @marcofrodl E-Mail: [email protected] https://www.thinktecture.com/thinktects/marco-frodl/

Ingestion Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI Simple RAG in a nutshell Splitted (smaller) parts Embedding- Model Embedding 𝑎 𝑏 𝑐 … Vector- Database Document Metadata: Reference to original document

Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI Demo: Ingestion

Ingestion++ HyQE: Hypothetical Question Embedding Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI 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

Watch the Webinar Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI Simple Advanced RAG in a nutshell https://www.thinktecture.com/webinare/moderne-semantic-search-mit-llms-vektor-datenbanken-und-langchain/

Ask me anything Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI 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 🦜🔗

Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI Demo: Simple RAG

Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI How to Debug/Trace Generative AI-Apps?

Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI Demo: Debugging

Just one Vector DB? Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI What’s wrong with Simple RAG?

Just one Vector DB/Retriever? • Multiple GenAI-Apps • Scaling and Load Balancing • Query Params per Retriever • Hosting (Environment, Product) • Fast Updates & Re-Indexing • Access Rights • Custom Retriever Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI What’s wrong with Simple RAG? On-Premise AI-App 🦜🔗 Cloud Docs Public Tickets Features Website Sales Docs Internal Tickets

Finding the best source before asking Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI 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 1

Finding the best source before asking Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI 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

Advanced RAG Dynamically Selecting the best Retrievers for Queries with AI Demo: Dynamic Retriever Selection with AI P