RAG (Retrieval-Augmented Generation)

Definition¶

RAG is a hybrid framework that combines retrieval-based and generation-based approaches for text generation. It enhances language model outputs by first retrieving relevant documents or passages from a knowledge base, then conditioning the generation process on both the input query and the retrieved information. This approach helps ground the model’s responses in specific, relevant information while maintaining the flexibility of generative models.

Tags¶

Natural Language Processing, Information Retrieval, Text Generation, Knowledge Base

References¶

• Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N., Küttler, H., Lewis, M., Yih, W., Rocktäschel, T., Riedel, S., & Kiela, D. (2020). Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks. Lewis et al. (2020)