From the course: Data Preparation, Feature Engineering, and Augmentation for AI Models

Unlock this course with a free trial

Join today to access over 25,200 courses taught by industry experts.

Retrieval-augmented generation (RAG)

Retrieval-augmented generation (RAG)

From the course: Data Preparation, Feature Engineering, and Augmentation for AI Models

Start my 1-month free trial Buy for my team

Retrieval-augmented generation (RAG)

“

- [Narrator] Now, I'd like to turn our attention to retrieval augmented generation, commonly referred to as RAG. Now, RAG is a technique that combines information retrieval with generative AI, and what it does is it enhances the outputs of large language models with relevant knowledge that's external to that model. And what we're doing is basically, bridging the gap between the static model knowledge that's encapsulated within an LLM with up-to-date information that we can get online. Now, one of the nice things about RAG is it can really help us with hallucinations, because with RAG we can ground responses in the retreat facts that we get from our information retrieval. Another nice thing about RAG is it allows us to augment what we can talk about or what the LLM can talk about, because it basically, enables specialized knowledge domains without fully retraining the model. All right, there are three major components of RAG. One, is the knowledge base. Now, this is a structured…

Contents