Joint Extraction of Events and Entities within a Document Context

Proceedings of NAACL-HLT 2016, pages 289–299, San Diego, California, June 12-17, 2016. c⃝2016 Association for Computational Linguistics Joint Extraction of Events and Entities within a Document Context Bishan Yang Carnegie Mellon University 5000 Forbes Avenue Pittsburgh, PA, 15213 bishan@cs.cmu.edu Tom Mitchell Carnegie Mellon University 5000 Forbes Avenue Pittsburgh, PA, 15213 tom.mitchell@cs.cmu.edu Abstract Events and entities are closely related; entities are often actors or participants in events and events without entities are uncommon. The interpretation of events and entities is highly contextually dependent. Existing work in in- formation extraction typically models events separately from entities, and performs infer- ence at the sentence level, ignoring the rest of the document. In this paper, we propose a novel approach that models the dependen- cies among variables of events, entities, and their relations, and performs joint inference of these variables across a document. The goal is to enable access to document-level contextual information and facilitate context- aware predictions. We demonstrate that our approach substantially outperforms the state- of-the-art methods for event extraction as well as a strong baseline for entity extraction. 1 Introduction Events are things that happen or occur; they in- volve entities (people, objects, etc.) who perform or are affected by the events and spatio-temporal as- pects of the world. Understanding events and their descriptions in text is necessary for any generally- applicable machine reading systems. It is also essen- tial in facilitating practical applications such as news summarization, information retrieval, and knowl- edge base construction. The interpretation of event descriptions is highly contextually dependent. To make correct predic- tions, a model needs to account for mentions of events and entities together with the discourse con- text. Consider, for example, the following excerpt from a news report: “On Thursday, there was a massive U.S. aerial bombardment in which more than 300 Tomahawk cruise missiles rained down on Baghdad. Earlier Saturday, Baghdad was again targeted. ...” The excerpt describes two U.S. attacks on Baghdad. The two event anchors (triggers) are boldfaced and the mentions of entities and spatio-temporal infor- mation are italicized. The first event anchor “aerial bombardment” along with its surrounding entity mentions — “U.S.”, “Tomahawk cruise missiles”, and “Baghdad”, describe an attack from the U.S. on Baghdad with Tomahawk cruise missiles being the weapon. The second sentence on its own contains little event-related information, but together with the context of the previous sentence, it indicates another U.S. attack on Baghdad. State-of-the-art event extraction systems have dif- ficulties inferring such information due to two main reasons. First, they extract events and entities in separate stages: entities such as people, organiza- tion, and locations are first extracted by a named entity tagger, and then these extracted entities are used as inputs for extracting events and their argu- ments (Li et al., 2013). This often causes errors to propagate. In the above example, if the entity tagger mistakenly identifies “Baghdad” as a person, then the event extractor will fail to extract “Baghdad” as the place where the attack happened. In fact, previ- ous work …

Content truncated to 500 words