Multi-granularity Item-Based Contrastive Recommendation

Ruobing Xie, Zhijie Qiu, Bo Zhang, Leyu Lin

Abstract: Contrastive learning (CL) has shown its power in recommendation. However, most CL-based models build their CL tasks merely focusing on the user’s aspects, ignoring systematically modeling the rich and diverse information in items. In this work, we propose a novel Multi-granularity item-based contrastive learning (MicRec) for the matching stage in recommendation, which systematically introduces multi-aspect item-related correlations to representation learning via CL. Specifically, we build three item-based CL tasks as a set of plug-and-play auxiliary objectives to capture item correlations in feature, semantic and session levels. In experiments, we conduct both offline and online evaluations on real-world datasets, verifying the effectiveness and universality of three proposed CL tasks. Currently, MicRec has been deployed on a real-world recommender system of WeChat Top Stories, affecting millions of users.