Open Domain Question Answering over Virtual Documents: A Unified Approach for Data and Text

Anonymous

Abstract: Due to its potential for a universal interface over both data and text, data-to-text generation is becoming increasingly popular.However, few prior work has focused on its application to downstream tasks, e.g. using the converted data for grounding or reasoning. In this work, we bridge this gap and use the data-to-text method as a means for encoding structured knowledge for knowledge-intensive applications, i.e. open-domain question answering (ODQA).Specifically, we propose a verbalizer-retriever-reader framework for ODQA over data and text where verbalized tables from Wikipedia and graphs from Wikidata are used as augmented knowledge sources.We show that our Unified Data and Text QA, UDT-QA.can effectively benefit from the expanded knowledge index, leading to large gains over text-only baselines.Notably, our approach sets the single-model state-of-the-art on Natural Questions.Furthermore, our analyses indicate that verbalized knowledge is preferred for answer reasoning for both adapted and hot-swap settings.