Exploiting Rich Event Representation to Improve Event Causality Recognition

Event causality identification is an essential task for information extraction that has attracted growing attention. Early researchers were accustomed to combining the convolutional neural network or recurrent models with external causal knowledge, but these methods ignore importance of rich semantic representation event. The event more structured, so it abundant representation. We argue elements event, interaction two events, and context between events can enrich event’s help identify causality. Therefore, effective in recognition deserves further study. To verify effectiveness identification, we proposed a model exploiting improve recognition. Our based on multi-column networks, which integrate representation, including tensor context-aware designed various experimental conducted experiments Chinese emergency corpus, most comprehensive annotation elements, enabling us study from all aspects. extensive showed achieved significant performance improvement over baseline recognition, indicating plays important role