Mitigating Negative Style Transfer in Hybrid Dialogue System


  • Shimin Li Fudan University
  • Qinyuan Cheng Fudan University
  • Linyang Li Fudan University
  • Xipeng Qiu Fudan University



SNLP: Conversational AI/Dialogue Systems


As the functionality of dialogue systems evolves, hybrid dialogue systems that accomplish user-specific goals and participate in open-topic chitchat with users are attracting growing attention. Existing research learns both tasks concurrently utilizing a multi-task fusion technique but ignores the negative transfer phenomenon induced by the unique textual style differences. Therefore, contrastive learning based on the latent variable model is used to decouple the various textual genres in the latent space. We devise supervised and self-supervised positive and negative sample constructions for diverse datasets. In addition, to capitalize on the style information contained in the decoupled latent variables, we employ a style prefix that incorporates latent variables further to control the generation of responses with varying styles. We performed extensive experiments on three dialogue datasets, including a hybrid dialogue dataset and two task-oriented dialogue datasets. The experimental results demonstrate that our method can mitigate the negative style transfer issue and achieves state-of-the-art performance on multiple dialogue datasets.




How to Cite

Li, S., Cheng, Q., Li, L., & Qiu, X. (2023). Mitigating Negative Style Transfer in Hybrid Dialogue System. Proceedings of the AAAI Conference on Artificial Intelligence, 37(11), 13103-13111.



AAAI Technical Track on Speech & Natural Language Processing