提出了SDP-LSTM,利用两个entities之间的SDP,multichannel recurrent neural networks, with long short term memory (LSTM) units, pick up heterogeneous information along the
SDP。
贡献:
SDP能删除无关的word
多通道的LSTM能有效整合从不同源头收集来的dependency path (multichannel LSTM networks allow effective information integration from heterogeneous sources over the dependency paths) 。什么是不同源头?如何整合?
一句话总结:
提出了SDP-LSTM,利用两个entities之间的SDP,multichannel recurrent neural networks, with long short term memory (LSTM) units, pick up heterogeneous information along the SDP。
贡献:
资源:
论文信息:
笔记:
模型图:
结果:
接下来要看的论文: