Neural Response Generation with Relevant Emotions for Short Text Conversation

摘要: Human conversations are often embedded with emotions. To simulate human conversations, the response generated by a chatbot not only has to be topically relevant post, but should also carry an appropriate emotion. In this paper, we conduct analysis based on social media data investigate how emotions influence conversation generation. Based observation, propose methods determine included in and generate responses The encoder-decoder architecture is extended incorporate We two implementations which train steps separately or jointly. An empirical study public dataset from STC at NTCIR-12 shows that our models outperform both retrieval-based method generation model without emotion, indicating importance of short text effectiveness approach.