A Stack LSTM Transition-Based Dependency Parser with Context Enhancement and K-best Decoding

Fuxiang Wu, Minghui Dong, Zhengchen Zhang, Fugen Zhou

Abstract: Transition-based parsing is useful for many NLP tasks. For improving the parsing accuracy, this paper proposes the following two enhancements based on a transition-based dependency parser with stack long short-term memory: using the context of a word in a sentence, and applying K-best decoding to expand the searching space. The experimental results show that the unlabeled and labeled attachment accuracies of our parser improve 0.70% and 0.87% over those of the baseline parser for English respectively, and are 0.82% and 0.86% higher than those of the baseline parser for Chinese respectively.