Definition
Retrieval-Augmented Generation (RAG) combines the strengths of retrieval-based and generation-based AI systems. Instead of relying solely on the parameters of a pre-trained language model, RAG first retrieves relevant documents or data snippets from an external database using semantic search. These retrieved pieces are then provided as context to the generative model, which uses them to produce accurate, up-to-date, and grounded responses, significantly reducing hallucinations and improving factual reliability.
Summary
RAG is an AI framework that enhances generative models by retrieving relevant information from external knowledge bases before generating responses.
Key Concepts
- Vector Database
- Embeddings
- Context Window
- Semantic Search
Use Cases
- Enterprise knowledge bases
- Customer support bots
- Research assistants