Improving the Domain Adaptation of Retrieval Augmented Generation (RAG) Models for Open Domain Question Answering
This work addresses the limitation of RAG models being optimized only for Wikipedia, enabling better performance in specialized domains like healthcare and news, though it is incremental as it builds directly on the existing RAG framework.
The paper tackles the problem of domain adaptation for Retrieval Augmented Generation (RAG) models in open-domain question answering by proposing RAG-end2end, which jointly trains retriever and generator components and uses an auxiliary training signal, achieving significant performance improvements on datasets from COVID-19, News, and Conversations domains compared to the original RAG model.
Retrieval Augment Generation (RAG) is a recent advancement in Open-Domain Question Answering (ODQA). RAG has only been trained and explored with a Wikipedia-based external knowledge base and is not optimized for use in other specialized domains such as healthcare and news. In this paper, we evaluate the impact of joint training of the retriever and generator components of RAG for the task of domain adaptation in ODQA. We propose \textit{RAG-end2end}, an extension to RAG, that can adapt to a domain-specific knowledge base by updating all components of the external knowledge base during training. In addition, we introduce an auxiliary training signal to inject more domain-specific knowledge. This auxiliary signal forces \textit{RAG-end2end} to reconstruct a given sentence by accessing the relevant information from the external knowledge base. Our novel contribution is unlike RAG, RAG-end2end does joint training of the retriever and generator for the end QA task and domain adaptation. We evaluate our approach with datasets from three domains: COVID-19, News, and Conversations, and achieve significant performance improvements compared to the original RAG model. Our work has been open-sourced through the Huggingface Transformers library, attesting to our work's credibility and technical consistency.