Exploring Pre-trained Language Models for Event Extraction and Generation

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 5284–5294 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 5284 Exploring Pre-trained Language Models for Event Extraction and Generation Sen Yang†, Dawei Feng†, Linbo Qiao, Zhigang Kan, Dongsheng Li‡ National University of Defense Technology, Changsha, China {sen yang,linbo.qiao,kanzhigang13}@nudt.edu.cn davyfeng.c@gmail.com, lds1201@163.com Abstract Traditional approaches to the task of ACE event extraction usually depend on manually annotated data, which is often laborious to cre- ate and limited in size. Therefore, in addi- tion to the difficulty of event extraction itself, insufficient training data hinders the learning process as well. To promote event extraction, we first propose an event extraction model to overcome the roles overlap problem by sep- arating the argument prediction in terms of roles. Moreover, to address the problem of in- sufficient training data, we propose a method to automatically generate labeled data by edit- ing prototypes and screen out generated sam- ples by ranking the quality. Experiments on the ACE2005 dataset demonstrate that our ex- traction model can surpass most existing ex- traction methods. Besides, incorporating our generation method exhibits further significant improvement. It obtains new state-of-the-art results on the event extraction task, including pushing the F1 score of trigger classification to 81.1%, and the F1 score of argument classifi- cation to 58.9%. 1 Introduction Event extraction is a key and challenging task for many NLP applications. It targets to detect event trigger and arguments. Figure 1 illustrates a sen- tence containing an event of type Meet triggered by ”meeting”, with two arguments: ”President Bush” and ”several Arab leaders”, both of which play the role ”Entity”. There are two interesting issues in event ex- traction that require more efforts. On the one hand, roles in an event vary greatly in frequency (Figure 2), and they can overlap on some words, †These two authors contributed equally. ‡Corresponding Author. [Trigger] Event type: Meet Sentence : President Bush is going to be meeting with several Arab leaders [Entity] [Entity] Figure 1: An event of type Meet is highlighted in the sentence, including one trigger and two arguments. even sharing the same argument (the roles over- lap problem). For example, in sentence ”The explosion killed the bomber and three shoppers”, ”killed” triggers an Attack event, while argument ”the bomber” plays the role ”Attacker” as well as the role ”Victim” at the same time. There are about 10% events in the ACE2005 dataset (Dod- dington et al., 2004) having the roles overlap prob- lem. However, despite the evidence of the roles overlap problem, few attentions have been paid to it. On the contrary, it is often simplified in evalu- ation settings of many approaches. For example, in most previous works, if an argument plays mul- tiple roles in an event simultaneously, the model classifies correctly as long as the prediction hits any one of them, which is obviously far from ac- curate to apply to the real world. Therefore, we design an effective mechanism to solve …

Content truncated to 500 words