Majumder N, Poria S, Hazarika D, et al. Dialoguernn: An attentive rnn for emotion detection in conversations[C]//Proceedings of the AAAI Conference on Artificial Intelligence. 2019, 33(01): 6818-6825.
解决的问题
论文原文: Currently systems do not treat the parties in the conversation individually by adapting to the speaker of each utterance.
在这篇论文之前很多dialogue emotion detection是不考虑对话中的角色问题的,这篇论文将对话角色信息包含进去,进行emotion detection。建模了individual parties & party states.
从3个方面建模:
- speaker
- context(preceding utterance)
- emotion(preceding utterance)
解决方法
Global-GRU获取的是一个全局的信息(speaker state + utterance)
speaker state是由Party-GRU生成的,Party-G