GATE-Time: Extraction of Temporal Expressions and Events

GATE-Time: Extraction of Temporal Expressions and Events Leon Derczynski,♥ Jannik Str¨otgen,♠ Diana Maynard,♥ Mark A. Greenwood,♥ Manuel Jung ♣ ♥ University of Sheffield; S1 4DP, UK ♠ Max-Planck-Institut f¨ur Informatik; 66123 Saarbr¨ucken, Germany ♣ Universit¨at Heidelberg; INF 348, 69120 Heidelberg, Germany Abstract GATE is a widely used open-source solution for text processing with a large user community. It contains components for several natural language processing tasks. However, temporal information extraction functionality within GATE has been rather limited so far, despite being a prerequisite for many application scenarios in the areas of natural language processing and information retrieval. This paper presents an integrated approach to temporal information processing. We take state-of-the-art tools in temporal expression and event recognition and bring them together to form an openly-available resource within the GATE infrastructure. GATE-Time provides annotation in the form of TimeML events and temporal expressions complying with this mature ISO standard for temporal semantic annotation of documents. Major advantages of GATE-Time are (i) that it relies on HeidelTime for temporal tagging, so that temporal expressions can be extracted and normalized in multiple languages and across different domains, (ii) it includes a modern, fast event recognition and classification tool, and (iii) that it can be combined with different linguistic pre-processing annotations, and is thus not bound to license restricted preprocessing components. Keywords: Temporal tagging, Event extraction, GATE 1. Introduction Time is critical to the meaning of language. Without it, we cannot discuss past events, change, or plans. Therefore, the annotation and extraction of temporal information from language is very important. Much of the information and many of the assertions made in a text are bounded in time. For example, the sky was not always blue; George W. Bush’s presidency was confined to an eight-year interval. An understanding of time in natu- ral language text is critical to effective communication and analysis and must be accounted for in natural language pro- cessing and understanding. Over recent years, research has been published in the area of temporal annotation resulting in several publicly avail- able tools for temporal information extraction. Unfortu- nately, many of the tools rely on license-restricted linguistic preprocessing components. In this paper, we present GATE-Time, an ISO- TimeML (Pustejovsky et al., 2010) approach to temporal information extraction in GATE (Cunningham et al., 2002) which brings together a number of these existing tools, as well as an upgraded approach to event annotation, within GATE. We use HeidelTime, a state-of-the-art, actively maintained, multilingual and domain-sensitive temporal expression tagger, breaking it out of its prepackaged pipeline and making the core component available in GATE. We present a novel event annotation tool, based on the earlier Evita system, using statistical learning with uneven margins; this yields state-of-the-art results. As GATE-Time can make use of whatever linguistic pre- processing annotations are provided in GATE, it is not bound to license restricted preprocessing components. The overall approach to integration thus adds flexibility to ex- isting and novel temporal information extraction tools and makes temporal annotation accessible to the large commu- …

Content truncated to 500 words