Salience-Aware Event Chain Modeling for Narrative Understanding

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 1418–1428 November 7–11, 2021. c⃝2021 Association for Computational Linguistics 1418 Salience-Aware Event Chain Modeling for Narrative Understanding Xiyang Zhang, Muhao Chen, and Jonathan May Information Sciences Institute University of Southern California {xiyangzh, muhao, jonmay}@isi.edu Abstract Storytelling, whether via fables, news reports, documentaries, or memoirs, can be thought of as the communication of interesting and related events that, taken together, form a concrete process. It is desirable to extract the event chains that represent such processes. However, this extraction remains a challenging problem. We posit that this is due to the nature of the texts from which chains are discovered. Natural language text interleaves a narrative of concrete, salient events with background infor- mation, contextualization, opinion, and other elements that are important for a variety of necessary discourse and pragmatics acts but are not part of the principal chain of events being communicated. We introduce methods for extracting this principal chain from natu- ral language text, by filtering away non-salient events and supportive sentences. We demon- strate the effectiveness of our methods at iso- lating critical event chains by comparing their effect on downstream tasks. We show that by pre-training large language models on our extracted chains, we obtain improvements in two tasks that benefit from a clear understand- ing of event chains: narrative prediction and event-based temporal question answering. The demonstrated improvements and ablative stud- ies confirm that our extraction method isolates critical event chains. 1 1 Introduction Human languages always communicate about evolving events. Hence, it is important for NLP systems to understand how events are procedurally described in text. In this context, identifying pat- terns of event chains is important but challenging, as it requires knowledge of inter-event relations such as temporal or causal relations. Modeling high-quality event chain patterns from text would be a first step toward the more general goal of 1Our code and data are available at https://github. com/juvezxy/Salience-Event-Chain. Figure 1: An illustration of event chains and salience- aware and discourse-aware filtering on an example text. The words in italic are the events and the arrows be- tween them show the temporal relation given by TEAR (pointing from previous to next). Salient events are shown in bold; the blue sentence is in the Main Event category whereas the red sentence is in the Evaluation category. event schema induction, which involves generating high-level representations of event relations and structures. The extracted chain patterns can in turn represent useful information for core natural lan- guage tasks such as question answering (Reddy et al., 2019), semantic role labeling (Cheng and Erk, 2018), story generation (Yao et al., 2019), and reading comprehension (Ostermann et al., 2019). Generally speaking, “events” correspond to what we perceive as happening around us. According to the theory of embodied cognition (Wilson, 2002), our understanding of the world is shaped by vari- ous aspects of our entire bodies, involving also our language comprehension. However, it is currently difficult to gather enough data …

Content truncated to 500 words