Distinguishing Past, On-going, and Future Events: The EventStatus Corpus

Abstract

Determining whether a major societal event has already happened, is still ongoing , or may occur in the future is crucial for event prediction, timeline generation, and news summarization. We introduce a new task and a new corpus, EventStatus, which has 4500 English and Spanish articles about civil unrest events labeled as PAST, ONGOING , or FUTURE. We show that the temporal status of these events is difficult to classify because local tense and aspect cues are often lacking, time expressions are insufficient, and the linguistic contexts have rich semantic compositionality. We explore two approaches for event status classification: (1) a feature-based SVM classifier augmented with a novel induced lexicon of future-oriented verbs, such as "threatened" and "planned", and (2) a convolutional neural net. Both types of classifiers improve event status recognition over a state-of-the-art TempEval model, and our analysis offers linguistic insights into the semantic compositionality challenges for this new task.

Cross-Context News Corpus for Protest Event-Related Knowledge Base Construction

“…Consequently, protest-event specific annotation schemas and data sets were proposed for creating automated or semi-automated event knowledge bases [4,11,24,25]. In their corpus, Makarov et al [26] used the ontology of CP events which is very similar to ours and identified ten event types.…”

Distinguishing Between Foreground and Background Events in News

“…Both were focused only on identifying a single central event, whereas we seek to label all events in a document as either Foreground, Background, or Other. Huang et al (2016) demonstrated an approach to placing events in news articles into three coarse temporal categories: Past events that have already occurred; On-Going events that are currently happening; and Future events that may happen. In that work, the temporal category was relative to the document creation time (DCT) and did not distinguish between foreground and background events.…”

Comprehensive Annotation and Analysis of Temporal Information of Events

“…The other approach is anchoring events to the time axis. Huang et al (2016) annotated one of five temporal status categories with events in newspaper articles on civil unrest: Past, On-going, Future Planned, Future Alert, Future Possible. Reimers et al (2016) anchored with finer granularity.…”

“…(1) 1 (Reimers et al 2016) (2) TimeBank (Asahara, Kato, Konishi, Imada, and Maekawa 2014) (BCCWJ) TimeBank Corpus (Kolomiyets, Bethard, and Moens 2012) TimeBank Dense Corpus (Cassidy et al 2014) 1 EventStatus Corpus (Huang, Cases, Jurafsky, Condoravdi, and Riloff 2016) Past, On-going, Future Planned, Future Alert, Future Possible 5 (Asakura, Hangyo, and Komachi 2016) PAST, PRESENT, FUTURE 3 (high probability, low probability, unmentioned) (Reimers et al 2016) TimeBank beginPoint=after 1984-10-01 before 1984-10-31 endPoint=after 1984 1 1 3 1 4 1 5 1 6 1 7 1 3 5 7 7 4 27 2 0 0 4 4 7 17 2 1 2 2 3 2 12 1 2 4 3 1 0 11 10 0 0 0 0 0 10 4 0 1 4 0 0 9 7 0 0 0 0 0 7 0 0 0 4 2 0 6 1 0 0 4 1 0 6 1 1 1 0 1 1 5 0 0 2 1 0 0 3 1 3 3 1 2 2 1 2 2…”