Recent state-of-the-art open-domain qa models are typically based on a two
stage retriever-reader approach in which the retriever first finds the relevant
knowledge/passages and the reader then leverages that to
Open-domain Question Answering research investigates the generalization performance of a retrieval-augmented QA model, proposing Corpus-Invariant Tuning as an effective training strategy to mitigate knowledge over-memorization and achieve better generalizability.