Human–machine dialogue modelling with the fusion of word- and sentence-level emotions

作者：

Highlights：

• Emotional features of word- and sentence-level are fused into the dialogue model.

• A structure with a latent variable and discriminator is proposed.

• The latent variable is used to realize the global control of emotion.

• Extensive experiments demonstrate the generated responses have good results.

摘要

•Emotional features of word- and sentence-level are fused into the dialogue model.•A structure with a latent variable and discriminator is proposed.•The latent variable is used to realize the global control of emotion.•Extensive experiments demonstrate the generated responses have good results.

论文关键词：Emotional intelligence,Human–machine dialogue system,Dictionary matching,Machine learning

论文评审过程：Received 1 May 2019, Revised 14 November 2019, Accepted 29 November 2019, Available online 4 December 2019, Version of Record 24 February 2020.

论文官网地址：https://doi.org/10.1016/j.knosys.2019.105319