AI Termcirca 2020· Added May 26, 2026
RAG (Retrieval-Augmented Generation)
RAG combines retrieval techniques with generative models to produce contextually rich outputs.
Retrieval-Augmented Generation (RAG) is an NLP framework that integrates two separate tasks: retrieval and generation. In a RAG system, relevant data is first retrieved from an external source, such as a database or search engine, which is then used to provide contextual information for the text generation process. This approach enables the model to generate more accurate and informative responses because it utilizes up-to-date information rather than relying solely on pre-trained data.
Examples
- Using RAG, a chatbot retrieves current news articles to answer user queries about the latest events.
- A customer support system employs RAG to access its database and offer precise solutions based on past cases.
Common misconceptions
- People often think RAG generates without any external input, but it relies heavily on retrieved data.
- Some believe RAG can replace all sources of information; however, it enhances them by providing context.
Related terms
Want more like this?
Open the full library
Fresh AI mastery content every 2 hours.