Global Constraints with Prompting for Zero-Shot Event Argument Classification

Findings of the Association for Computational Linguistics: EACL 2023, pages 2527–2538 May 2-6, 2023 ©2023 Association for Computational Linguistics Global Constraints with Prompting for Zero-Shot Event Argument Classification Zizheng Lin1, Hongming Zhang2 & Yangqiu Song1 1Department of Computer Science and Engineering, HKUST 2Tencent AI Lab, Seattle {zlinai, yqsong}@cse.ust.hk; hongmzhang@global.tencent.com Abstract Determining the role of event arguments is a crucial subtask of event extraction. Most pre- vious supervised models leverage costly anno- tations, which is not practical for open-domain applications. In this work, we propose to use global constraints with prompting to ef- fectively tackles event argument classification without any annotation and task-specific train- ing. Specifically, given an event and its as- sociated passage, the model first creates sev- eral new passages by prefix prompts and cloze prompts, where prefix prompts indicate event type and trigger span, and cloze prompts con- nect each candidate role with the target argu- ment span. Then, a pre-trained language model scores the new passages, making the initial pre- diction. Our novel prompt templates can easily adapt to all events and argument types with- out manual effort. Next, the model regularizes the prediction by global constraints exploiting cross-task, cross-argument, and cross-event re- lations. Extensive experiments demonstrate our model’s effectiveness: it outperforms the best zero-shot baselines by 12.5% and 10.9% F1 on ACE and ERE with given argument spans and by 4.3% and 3.3% F1, respectively, without given argument spans. We have made our code publicly available.1 1 Introduction Event Argument Classification2 (EAC), finding the roles of event arguments, is an important and chal- lenging event extraction sub-task. As shown in Figure 1, a “Transfer-Money” event whose trig- ger is “acquiring” has several argument spans (e.g., “Daily Planet”). By determining the role of these ar- guments (e.g., “Daily Planet” as “Beneficiary”), we 1https://github.com/HKUST-KnowComp/Constraints- with-Prompting-for-Zero-Shot-EAC 2We focus on event argument because existing zero-shot trigger extraction models like Zhang et al. (2021) are already strong enough, but the arguments remain a challenge. Our argument identification approach is described in Section 3.1. Figure 1: An example of EAC. The trigger is in bold face. Arguments are underlined and connected to their roles by arrows. can obtain a better understanding of the event, thus benefiting related applications like stock price pre- diction (Ding et al., 2015) and biomedical research (Zhao et al., 2021). Many previous EAC works require numerous annotations to train their models (Lin et al., 2020; Hsu et al., 2022; Liu et al., 2022), which is not only costly as the annotations are labor-intensive but also difficult to be generalized to datasets of novel domains. Accordingly, some EAC models adopt a few-shot learning paradigm (Ma et al., 2022; Hsu et al., 2022). However, they are sensitive to the few- shot example selection and they still require costly task-specific training, which hinders their real-life deployment. There have been some zero-shot EAC models based on transfer learning (Huang et al., 2018), or label semantics (Zhang et al., 2021; Wang et al., 2022), or prompt learning (Liu et al., 2020; Lyu et …

Content truncated to 500 words