Bi-directional Contrastive Distillation for Multi-behavior Recommendation

Abstract

Multi-behavior recommendation leverages auxiliary behaviors (e.g., view, add-to-cart) to improve the prediction for target behaviors (e.g., buy). Most existing works are built upon the assumption that all the auxiliary behaviors are positively correlated with target behaviors. However, we empirically find that such an assumption may not hold in real-world datasets. In fact, some auxiliary feedback is too noisy to be helpful, and it is necessary to restrict its influence for better performance. To this end, in this paper we propose a Bi-directional Contrastive Distillation (BCD) model for multi-behavior recommendation, aiming to distill valuable knowledge (about user preference) from the interplay of multiple user behaviors. Specifically, we design a forward distillation to distill the knowledge from auxiliary behaviors to help model target behaviors, and then a backward distillation to distill the knowledge from target behaviors to enhance the modelling of auxiliary behaviors. Through this circular learning, we can better extract the common knowledge from multiple user behaviors, where noisy auxiliary behaviors will not be involved. The experimental results on two real-world datasets show that our approach outperforms other counterparts in accuracy.