Toward Gender-Inclusive Coreference Resolution

Coreference resolution is a task of determining which textual references resolve tot he same real-world entity—requires making inferences about those entities, especially when those entities are people. In order to analyze what sort of gender biases exist in coreference resolution systems and crowd-worker judgments and move towards gender-inclusive coreference systems, we created twodatasets: MAP and GICoref.

More details are available in our paper. If you use these datasets, please acknowledge it by citing the original paper:

@inproceedings{cao2019toward,
  title={Toward Gender-Inclusive Coreference Resolution},
  author={Yang Trista Cao and Hal Daum{\'e}},
  booktitle={Proceedings of the Conference of
             the Association for Computational Linguistics (ACL)},
  year={2020},
  note={abs/1910.13913}
}