The integration of dependency relation classification and semantic role labeling using bilayer maximum entropy Markov models
Weiwei Sun, Hongzhan Li, Zhifang Sui · 2008
This paper describes a system to solve the joint learning of syntactic and semantic dependencies.An directed graphical model is put forward to integrate dependency relation classification and semantic role labeling.We present a bilayer directed graph to express probabilistic relationships between syntactic and semantic relations.Maximum Entropy Markov Models are implemented to estimate conditional probability distribution and to do inference.The submitted model yields 76.28% macro-average F1 performance, for the joint task, 85.75% syntactic dependencies LAS and 66.61% semantic dependencies F1.