RAG
RAG (Retrieval-Augmented Generation) is a technique that combines retrieval-based methods with generative models to improve the quality and relevance of generated content. It leverages external knowledge sources to enhance the model's understanding and response generation capabilities.
Algorithm:
- Collect relevant documents or information from a knowledge base or external sources.
- Use a embedding model to convert the documents and the input query into numerical representations.
- Use a retrieval model to identify the most relevant documents
based on the input query or context.
- Use a nearest neighbor search algorithm (e.g., FAISS) to find the most similar documents in the embedding space.
- Feed the retrieved documents into a generative model (e.g., a language model) to generate a response that incorporates the retrieved information.