Harnessing the Power of RAG
By now, we know the FMs are trained using large datasets. However, the data used to train FMs might not be recent, and this can cause the models to hallucinate. In this chapter, we will harness the power of RAG by augmenting the model with external data sources to overcome the challenge of hallucination.
We will explore the importance of RAG in generative AI scenarios, how RAG works, and its components. We will then delve into the integration of RAG with Amazon Bedrock, including a fully managed RAG experience by Amazon Bedrock called Knowledge Bases. The chapter will then take a hands-on approach to the implementation of Knowledge Bases and using APIs.
We will explore some real-world scenarios of RAG and discuss a few solution architectures for implementing RAG. You will also be introduced to implementing a RAG framework using Amazon Bedrock, LangChain orchestration, and other generative AI systems. We will end by examining current limitations and...
