2000 character limit reached
Frustratingly Hard Evidence Retrieval for QA Over Books (2007.09878v1)
Published 20 Jul 2020 in cs.CL and cs.IR
Abstract: A lot of progress has been made to improve question answering (QA) in recent years, but the special problem of QA over narrative book stories has not been explored in-depth. We formulate BookQA as an open-domain QA task given its similar dependency on evidence retrieval. We further investigate how state-of-the-art open-domain QA approaches can help BookQA. Besides achieving state-of-the-art on the NarrativeQA benchmark, our study also reveals the difficulty of evidence retrieval in books with a wealth of experiments and analysis - which necessitates future effort on novel solutions for evidence retrieval in BookQA.
- Xiangyang Mou (8 papers)
- Mo Yu (117 papers)
- Bingsheng Yao (49 papers)
- Chenghao Yang (25 papers)
- Xiaoxiao Guo (38 papers)
- Saloni Potdar (20 papers)
- Hui Su (38 papers)