PGR-crowd: A hybrid approach toward biomedical relation extraction training corpora: combining distant supervision with crowdsourcing
This project encompasses the following contributions:
-
Pipeline for RE crowdsourcing (some auxiliary functions for MTurk submission).
-
New release of the PGR dataset (PGR-crowd).
-
Assessment of the quality of results obtained with the MTurk platform (through statistical analysis and direct comparison with on-site rater and domain expert).
Our academic paper which describes the work in detail is available here.
- Diana Sousa, Andre Lamurias, and Francisco M. Couto. 2020. A hybrid approach toward biomedical relation extraction training corpora: combining distant supervision with crowdsourcing. Database, 2020.