Improving Cascade Decoding with Syntax-aware Aggregator and Contrastive Learning for Event Extraction

  • Zeyu Sheng
  • , Yuanyuan Liang
  • , Yunshi Lan*
  • *Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Cascade decoding framework has shown superior performance on event extraction tasks. However, it treats a sentence as a sequence and neglects the potential benefits of the syntactic structure of sentences. In this paper, we improve cascade decoding with a novel module and a self-supervised task. Specifically, we propose a syntax-aware aggregator module to model the syntax of a sentence based on cascade decoding framework such that it captures event dependencies as well as syntactic information. Moreover, we design a type discrimination task to learn better syntactic representations of different event types, which could further boost the performance of event extraction. Experimental results on two widely used event extraction datasets demonstrate that our method could improve the original cascade decoding framework by up to 2.2% percentage points of F1 score and outperform a number of competitive baseline methods.

Original languageEnglish
Title of host publicationProceedings of the 22nd Chinese National Conference on Computational Linguistics, CCL 2023
EditorsMaosong Sun, Bing Qin, Xipeng Qiu, Jing Jiang, Xianpei Han
PublisherAssociation for Computational Linguistics (ACL)
Pages748-760
Number of pages13
ISBN (Electronic)9781713876229
StatePublished - 2023
Event22nd Chinese National Conference on Computational Linguistics, CCL 2023 - Harbin, China
Duration: 3 Aug 20235 Aug 2023

Publication series

NameProceedings of the 22nd Chinese National Conference on Computational Linguistics, CCL 2023
Volume1

Conference

Conference22nd Chinese National Conference on Computational Linguistics, CCL 2023
Country/TerritoryChina
CityHarbin
Period3/08/235/08/23

Fingerprint

Dive into the research topics of 'Improving Cascade Decoding with Syntax-aware Aggregator and Contrastive Learning for Event Extraction'. Together they form a unique fingerprint.

Cite this