Abstract
News describe real-world events of varying granularity, and recognition of internal structure of events is important for automated reasoning over events. We propose an approach for constructing coherent event hierarchies from news by enforcing document-level coherence over pairwise decisions of spatiotemporal containment. Evaluation on a news corpus annotated with event hierarchies shows that enforcing global spatiotemporal coreference of events leads to significant improvements (7.6% F1-score) in the accuracy of pairwise decisions.
Cite
CITATION STYLE
Glavaš, G., & Šnajder, J. (2014). Constructing coherent event hierarchies from news stories. In Proceedings of TextGraphs@EMNLP 2014: The 9th Workshop on Graph-Based Methods for Natural Language Processing (pp. 34–38). The Association for Computer Linguistics. https://doi.org/10.3115/v1/w14-3705
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
