RAG with Hybrid Search: How Does Keyword Search Work?

Retrieval-Augmented Generation (RAG) systems using semantic search can struggle with retrieving exact keyword matches from a knowledge base. A hybrid search approach solves this by combining semantic search with keyword-based techniques like the BM25 ranking function. This method is built on the foundational concept of TF-IDF (Term Frequency-Inverse Document Frequency), which measures a word's importance in a document. BM25 improves upon TF-IDF by incorporating document length normalization and term frequency saturation to provide a more effective relevance score for information retrieval.