Using compositional semantics and discourse consistency to improve Chinese trigger identification

Due to the special characteristics and challenges in Chinese language, event extraction in Chinese is much more difficult than that in English. In particular, the state-of-the-art Chinese event extraction systems suffer much from the low recall in trigger identification due to the failure in identifying unknown triggers and the inconsistency in identifying trigger mentions. To resolve these two issues, this paper proposes an inference mechanism to infer unknown triggers via the compositional semantics inside Chinese words and another inference mechanism to recover trigger mentions via the discourse consistency between Chinese trigger mentions. Here, various morphological structures are explored to better represent the compositional semantics inside Chinese triggers and automatically identify the head morpheme as the governing sememe of a trigger in inferring unknown triggers. Evaluation on the ACE 2005 Chinese corpus justifies the effectiveness of our approach over a strong baseline in Chinese event extraction, in particular trigger identification.