TY - GEN
T1 - Joint End-to-End Semantic Proto-role Labeling
AU - Spaulding, Elizabeth
AU - Kazantsev, Gary
AU - Dredze, Mark
N1 - Funding Information:
We thank Elias Stengel-Eskin, Benjamin Van Durme, Igor Malioutov, and Leslie Barrett for their helpful comments and feedback throughout conversations surrounding the project. We additionally acknowledge anonymous Bloomberg employees for assistance in reviewing the paper. Finally, we thank the ACL reviewers for their careful consideration and invaluable feedback.
Publisher Copyright:
© 2023 Association for Computational Linguistics.
PY - 2023
Y1 - 2023
N2 - Semantic proto-role labeling (SPRL) assigns properties to arguments based on a series of binary labels. While multiple studies have evaluated various approaches to SPRL, it has only been studied in-depth as a standalone task using gold predicate/argument pairs. How do SPRL systems perform as part of an information extraction pipeline? We model SPRL jointly with predicate-argument extraction using a deep transformer model. We find that proto-role labeling is surprisingly robust in this setting, with only a small decrease when using predicted arguments. We include a detailed analysis of each component of the joint system, and an error analysis to understand correlations in errors between system stages. Finally, we study the effects of annotation errors on SPRL.
AB - Semantic proto-role labeling (SPRL) assigns properties to arguments based on a series of binary labels. While multiple studies have evaluated various approaches to SPRL, it has only been studied in-depth as a standalone task using gold predicate/argument pairs. How do SPRL systems perform as part of an information extraction pipeline? We model SPRL jointly with predicate-argument extraction using a deep transformer model. We find that proto-role labeling is surprisingly robust in this setting, with only a small decrease when using predicted arguments. We include a detailed analysis of each component of the joint system, and an error analysis to understand correlations in errors between system stages. Finally, we study the effects of annotation errors on SPRL.
UR - http://www.scopus.com/inward/record.url?scp=85172263364&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85172263364&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85172263364
T3 - Proceedings of the Annual Meeting of the Association for Computational Linguistics
SP - 723
EP - 736
BT - Short Papers
PB - Association for Computational Linguistics (ACL)
T2 - 61st Annual Meeting of the Association for Computational Linguistics, ACL 2023
Y2 - 9 July 2023 through 14 July 2023
ER -