Constructing Datasets for Multi-hop Reading Comprehension Across Documents
- Submitting institution
-
University College London
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 14459
- Type
- D - Journal article
- DOI
-
10.1162/tacl_a_00021
- Title of journal
- Transactions of the Association for Computational Linguistics
- Article number
- -
- First page
- 287
- Volume
- 6
- Issue
- -
- ISSN
- 2307-387X
- Open access status
- Compliant
- Month of publication
- May
- Year of publication
- 2018
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
2
- Research group(s)
-
-
- Citation count
- -
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Introduced the first multi-step machine reading comprehension dataset using a novel method drawing upon existing knowledge base resources and publicly available text. It has since inspired numerous other works in this direction and become a standard dataset with 29 academic and industrial groups making submissions to the leader board during since its release. The work was accepted at the most prestigious journal in the field and is currently its most cited paper published in 2018 and its 12th most cited paper of all time.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -