N-ary Relation Extraction using Graph State LSTM

Linfeng Song, et.al., “N-ary Relation Extraction using Graph State LSTM”, EMNLP 2018

  • 姓名:Graph State LSTM
  • 机构:University of Rochester
  • 任务:cross-sentence N-ary Relation Extraction
  • 动机:nanyun peng seperate hte graph into 2 DAGs, in which step important informations will be lost
  • 方法:change Graph LSTM into multiple timesteps, update one node for one timestep. During updating, the in-edgesand out-edges are considered.
  • 性能:a huge margin(~5%) compared to nanyun peng’s method