HICL: Hierarchical Intent Contrastive Learning for sequential recommendation

Yan Kang, Yancong Yuan, Bin Pu, Yun Yang, Lei Zhao, Jing Guo

Abstract: Highlights•We propose a model HICL fuses sequence contrastive learning (CL) and graph CL.•The HICL can capture the dynamic and sparse intention in sequential recommendation.•Experimental results demonstrate the effectiveness of mining users’ intent.