TY - GEN
T1 - Swanson linking revisited
T2 - 55th Annual Meeting of the Association for Computational Linguistics, ACL 2017
AU - Hahn-Powell, Gus
AU - Valenzuela-Escárcega, Marco
AU - Surdeanu, Mihai
N1 - Funding Information:
This work was funded by the DARPA Big Mechanism program under ARO contract W911NF-14-1-0395 and by the Bill and Melinda Gates Foundation HBGDki Initiative. The authors declare a financial interest in lum.ai, which licenses the intellectual property involved in this research. This interest has been properly disclosed to the University of Arizona Institutional Review Committee and is managed in accordance with its conflict of interest policies.
Publisher Copyright:
© 2017 Association for Computational Linguistics
PY - 2017
Y1 - 2017
N2 - We introduce a modular approach for literature-based discovery consisting of a machine reading and knowledge assembly component that together produce a graph of influence relations (e.g., “A promotes B”) from a collection of publications. A search engine is used to explore direct and indirect influence chains. Query results are substantiated with textual evidence, ranked according to their relevance, and presented in both a table-based view, as well as a network graph visualization. Our approach operates in both domain-specific settings, where there are knowledge bases and ontologies available to guide reading, and in multi-domain settings where such resources are absent. We demonstrate that this deep reading and search system reduces the effort needed to uncover “undiscovered public knowledge”, and that with the aid of this tool a domain expert was able to drastically reduce her model building time from months to two days.
AB - We introduce a modular approach for literature-based discovery consisting of a machine reading and knowledge assembly component that together produce a graph of influence relations (e.g., “A promotes B”) from a collection of publications. A search engine is used to explore direct and indirect influence chains. Query results are substantiated with textual evidence, ranked according to their relevance, and presented in both a table-based view, as well as a network graph visualization. Our approach operates in both domain-specific settings, where there are knowledge bases and ontologies available to guide reading, and in multi-domain settings where such resources are absent. We demonstrate that this deep reading and search system reduces the effort needed to uncover “undiscovered public knowledge”, and that with the aid of this tool a domain expert was able to drastically reduce her model building time from months to two days.
UR - http://www.scopus.com/inward/record.url?scp=85054291342&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85054291342&partnerID=8YFLogxK
U2 - 10.18653/v1/P17-4018
DO - 10.18653/v1/P17-4018
M3 - Conference contribution
AN - SCOPUS:85054291342
SN - 9781945626715
T3 - ACL 2017 - 55th Annual Meeting of the Association for Computational Linguistics, Proceedings of System Demonstrations
SP - 103
EP - 108
BT - ACL 2017 - 55th Annual Meeting of the Association for Computational Linguistics, Proceedings of System Demonstrations
PB - Association for Computational Linguistics (ACL)
Y2 - 30 July 2017 through 4 August 2017
ER -