Improving Event Causality Identification via Self-Supervised Representation Learning on External Causal Statement

Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pages 2162–2172 August 1–6, 2021. ©2021 Association for Computational Linguistics 2162 Improving Event Causality Identification via Self-Supervised Representation Learning on External Causal Statement Xinyu Zuo1,2, Pengfei Cao1,2, Yubo Chen1,2, Kang Liu1,2, Jun Zhao1,2, Weihua Peng3 and Yuguang Chen3 1National Laboratory of Pattern Recognition, Institute of Automation, CAS, Beijing, China 2School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China 3Beijing Baidu Netcom Science Technology Co., Ltd {xinyu.zuo,pengfei.cao,yubo.chen,kliu,jzhao}@nlpr.ia.ac.cn {pengweihua,chenyuguang}@baidu.com Abstract Current models for event causality identifica- tion (ECI) mainly adopt a supervised frame- work, which heavily rely on labeled data for training. Unfortunately, the scale of current annotated datasets is relatively limited, which cannot provide sufficient support for models to capture useful indicators from causal state- ments, especially for handing those new, un- seen cases. To alleviate this problem, we propose a novel approach, shortly named CauSeRL, which leverages external causal statements for event causality identification. First of all, we design a self-supervised frame- work to learn context-specific causal patterns from external causal statements. Then, we adopt a contrastive transfer strategy to incor- porate the learned context-specific causal pat- terns into the target ECI model. Experimen- tal results show that our method significantly outperforms previous methods on EventSto- ryLine and Causal-TimeBank (+2.0 and +3.4 points on F1 value respectively). 1 Introduction Event causality identification (ECI) aims to identify causal relations between events in texts, which can provide crucial clues for deep textual understanding (Girju, 2003; Oh et al., 2013, 2017). For example in Figure 1, an ECI system should identify two causal relations in S1 with mentioned events: noticedE1 cause −→ alertedE3 and alertedE3 cause −→ ranE2. To date, most existing methods regard this task as a classification problem and usually train ECI models on annotated data (Hashimoto et al., 2014; Riaz and Girju, 2014b; Mirza and Tonelli, 2016; Hu and Walker, 2017b; Gao et al., 2019). However, the scale of current annotated datasets are relatively limited, where the so far largest dataset EventSto- ryLine (Caselli and Vossen, 2017) only contains 258 documents, 4316 sentences, and 1770 causal The captain noticedE1 the pirates five minutes ago, he ranE2 to the deck and alertedE3 the crew of the emergency. Billy finds his childhood teddy bear. >Causes/Enables> Billy gives his childhood teddy bear to his daughter. [Entity] find/notice/feel/... [Entity] >Causes/Enables> [Entity] call/give/alert/... [Entity] Context-specific Causal Pattern noticedE1 alertedE3 Unseen Case noticedE1 alertedE3 Prediction Figure 1: S1 is a labeled data that contains unseen causal events and their statement when training; S2 is an external causal statement; The bottom illustrates the context-specific causal pattern in S2 could help identify the causality of unseen events in S1. event pairs. As a result, on the limited annotated examples, existing ECI models could not easily capture useful indicators from causal statements, especially for handing those new, unseen cases. To address this problem, Liu et al. (2020) em- ployed external event-related knowledge bases (KBs) to enhance the causality inference, where those KBs store inherent causal relations between …

Content truncated to 500 words