Robust Integrated Models for Chinese Predicate-Argument Structure Analysis

Predicate-Argument (PA) structure analysis is often divided into three subtasks: predicate sense disambiguation, argument identification and argument classification. To date, they have mostly been modeled in isolation. However, this approach neglects logical constraints between them. We therefore exploite integrating predicate sense disambiguation with the latter two subtasks respectively, which verifies that the automatic predicate sense disambiguation could help the semantic role labeling task. In addition, a dual decomposition algorithm is used to alleviate the error propagation between argument identification subtask and argument classification subtask by benefitting the argument identification subtask greatly. Experiment results show that our approach leads to a better performance with PA analysis than other pipeline approaches.