RAG's roots traced to early 2000s IR research, not LLMs

A new paper argues that Retrieval-Augmented Generation (RAG), often seen as a novel LLM paradigm, has deep roots in earlier information retrieval and question answering research. The authors trace RAG's core concepts, such as integrating retrieval with generation and iterative query refinement, back to work from the early 2000s. They suggest that community fragmentation and shifting terminology have led to this historical continuity being overlooked, and propose viewing LLMs as an interface layer atop established QA architectures. This reframing could inform future RAG designs by leveraging underutilized prior work in areas like user modeling and answer validation. AI

IMPACT Reframes RAG's origins, potentially unlocking underutilized prior work for next-generation AI designs.

RANK_REASON The cluster contains an academic paper published on arXiv discussing the historical origins of a research concept. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Agentic RAG
• alphaXiv
• arXiv
• DagsHub
• Hugging Face
• information retrieval
• large language models
• Question Answering
• retrieval-augmented generation

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →