Based on the shallow tree and dependency tree, the authors introduce the structural representations, NPST (new phrase-based shallow tree) and NPDT (new phrase-based dependency tree) to Tree-LSTM to compute sentence similarity. Experimental results manifest that the proposed approach achieves a higher performance than the baseline.

Currently, most pairwise resolution models for event co-reference focused on classification or clustering approaches, which ignored the relations between events in a document. A global optimization model for event co-reference resolution was proposed to resolve the inconsistent event chains in classifier-based approaches. This model regarded co-reference resolution as a integer linear program problem and introduced various kinds of constraints, such as symmetry, transitivity, triggers, argument roles, event distances, to further improve the performance. The experimental results show that the proposed model outperforms the local classifier by 4.20% in F1-measure.

Previous Chinese argument extraction approaches mainly focus on feature engineering and trigger expansion, which cannot exploit inner relation between trigger mentions in same document. To address this issue, the authors bring forward a novel trigger inference mechanism based on Markov logic network. Head morpheme, the probabilities of a trigger mention fulfilling true and pseudo events from the training set and the relationships between trigger mentions are used to infer those trigger mentions with lack of effective context information or low confidences in testing set. Experimental results on the ACE 2005 Chinese corpus show that the proposed approach outperforms the baseline, with the F1 improvements of 3.65% and 2.51% in trigger identification and event type classification respectively.