Contrastive Learning Reduces Hallucination in Conversations


  • Weiwei Sun Shandong University
  • Zhengliang Shi Shandong University
  • Shen Gao Shandong University
  • Pengjie Ren Shandong University
  • Maarten de Rijke University of Amsterdam
  • Zhaochun Ren Shandong University



SNLP: Conversational AI/Dialogue Systems, SNLP: Applications, SNLP: Generation, SNLP: Language Models


Pre-trained language models (LMs) store knowledge in their parameters and can generate informative responses when used in conversational systems. However, LMs suffer from the problem of “hallucination:” they may generate plausible-looking statements that are irrelevant or factually incorrect. To address this problem, we propose a contrastive learning scheme, named MixCL. A novel mixed contrastive objective is proposed to explicitly optimize the implicit knowledge elicitation process of LMs, and thus reduce their hallucination in conversations. We also examine negative sampling strategies of retrieved hard negatives and model-generated negatives. We conduct experiments on Wizard-of-Wikipedia, a public, open-domain knowledge-grounded dialogue benchmark, and assess the effectiveness of MixCL. MixCL effectively reduces the hallucination of LMs in conversations and achieves the highest performance among LM-based dialogue agents in terms of relevancy and factuality. We show that MixCL achieves comparable performance to state-of-the-art KB-based approaches while enjoying notable advantages in terms of efficiency and scalability.




How to Cite

Sun, W., Shi, Z., Gao, S., Ren, P., de Rijke, M., & Ren, Z. (2023). Contrastive Learning Reduces Hallucination in Conversations. Proceedings of the AAAI Conference on Artificial Intelligence, 37(11), 13618-13626.



AAAI Technical Track on Speech & Natural Language Processing