Fine-Grained Temporal Relation Extraction

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 2906–2919 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 2906 Fine-Grained Temporal Relation Extraction Siddharth Vashishtha University of Rochester Benjamin Van Durme Johns Hopkins University Aaron Steven White University of Rochester Abstract We present a novel semantic framework for modeling temporal relations and event dura- tions that maps pairs of events to real-valued scales. We use this framework to construct the largest temporal relations dataset to date, covering the entirety of the Universal Depen- dencies English Web Treebank. We use this dataset to train models for jointly predicting fine-grained temporal relations and event du- rations. We report strong results on our data and show the efficacy of a transfer-learning ap- proach for predicting categorical relations. 1 Introduction Natural languages provide a myriad of formal and lexical devices for conveying the temporal struc- ture of complex events—e.g. tense, aspect, auxil- iaries, adverbials, coordinators, subordinators, etc. Yet, these devices are generally insufficient for de- termining the fine-grained temporal structure of such events. Consider the narrative in (1). (1) At 3pm, a boy broke his neighbor’s window. He was running away, when the neighbor rushed out to confront him. His parents were called but couldn’t arrive for two hours be- cause they were still at work. Most native English speakers would have little dif- ficulty drawing a timeline for these events, likely producing something like that in Figure 1. But how do we know that the breaking, the run- ning away, the confrontation, and the calling were short, while the parents being at work was not? And why should the first four be in sequence, with the last containing the others? The answers to these questions likely involve a complex interplay between linguistic information, on the one hand, and common sense knowledge about events and their relationships, on the other run away arrive rush out confront call be at work break 5pm 3pm 4pm Figure 1: A typical timeline for the narrative in (1). (Minsky, 1975; Schank and Abelson, 1975; Lam- port, 1978; Allen and Hayes, 1985; Hobbs et al., 1987; Hwang and Schubert, 1994). But it remains a question how best to capture this interaction. A promising line of attack lies in the task of temporal relation extraction. Prior work in this domain has approached this task as a classification problem, labeling pairs of event- referring expressions—e.g. broke or be at work in (1)—and time-referring expressions—e.g. 3pm or two hours—with categorical temporal rela- tions (Pustejovsky et al., 2003; Styler IV et al., 2014; Minard et al., 2016). The downside of this approach is that time-referring expressions must be relied upon to express duration information. But as (1) highlights, nearly all temporal dura- tion information can be left implicit without hin- dering comprehension, meaning these approaches only explicitly encode duration information when that information is linguistically realized. In this paper, we develop a novel framework for temporal relation representation that puts event duration front and …

Content truncated to 500 words