Can Machines Learn to Comprehend Scientific Literature?
- Authors
- Park, Donghyeon; Choi, Yonghwa; Kim, Daehan; Yu, Minhwan; Kim, Seongsoon; Kang, Jaewoo
- Issue Date
- 2019
- Publisher
- IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
- Keywords
- Artificial intelligence; crowdsourcing; data acquisition; data analysis; data collection; data mining; data preprocessing; knowledge discovery; machine intelligence; natural language processing; social computing; text analysis; text mining
- Citation
- IEEE ACCESS, v.7, pp.16246 - 16256
- Indexed
- SCIE
SCOPUS
- Journal Title
- IEEE ACCESS
- Volume
- 7
- Start Page
- 16246
- End Page
- 16256
- URI
- https://scholar.korea.ac.kr/handle/2021.sw.korea/68960
- DOI
- 10.1109/ACCESS.2019.2891666
- ISSN
- 2169-3536
- Abstract
- To measure the ability of a machine to understand professional-level scientific articles, we construct a scientific question answering task called PaperQA. The PaperQA task is based on more than 80 000 "fill-in-the-blank" type questions on articles from reputed scientific journals such as Nature and Science. We perform fine-grained linguistic analysis and evaluation to compare PaperQA and other conventional question and answering (QA) tasks on general literature (e.g., books, news articles, and Wikipedia texts). The results indicate that the PaperQA task is the most difficult QA task for both humans (lay people) and machines (deep-learning models). Moreover, humans generally outperform machines in conventional QA tasks, but we found that advanced deep-learning models outperform humans by 3%-13% on average in the PaperQA task. The PaperQA dataset used in this paper is publicly available at http://dmis.korea.ac.kr/downloads?id=PaperQA.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - Graduate School > Department of Computer Science and Engineering > 1. Journal Articles
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.