Multi-Behavior Collaborative Filtering with Partial Order Graph Convolutional Networks
Yijie Zhang, Yuanchen Bei, Hao Chen, Qijie Shen, Zheng Yuan, Huan Gong, Senzhang Wang, Feiran Huang, Xiao Huang
TL;DR
The paper tackles the seesaw problem in multi-behavior collaborative filtering by introducing Partial Order Recommendation Graphs (POG) to merge separate behavior graphs into a unified, ranked structure. Building on this, it proposes Partial Order Graph Convolutional Networks (POGCN) and a tailored Partial Order BPR (POBPR) training strategy to learn a single user/item embedding that benefits all behaviors. The approach is formalized with a graded partial order over behavior combinations, a matrix- and message-passing GCN on the POG, and a probabilistic, data-driven sampling scheme for training across behavior combinations. Empirically, POGCN outperforms state-of-the-art baselines on three public datasets and shows significant online gains in an Alibaba deployment, including improvements in Recall/NDCG and business metrics such as CVR, GMV, and stay time, demonstrating its practical impact for billion-scale recommender systems.
Abstract
Representing information of multiple behaviors in the single graph collaborative filtering (CF) vector has been a long-standing challenge. This is because different behaviors naturally form separate behavior graphs and learn separate CF embeddings. Existing models merge the separate embeddings by appointing the CF embeddings for some behaviors as the primary embedding and utilizing other auxiliaries to enhance the primary embedding. However, this approach often results in the joint embedding performing well on the main tasks but poorly on the auxiliary ones. To address the problem arising from the separate behavior graphs, we propose the concept of Partial Order Recommendation Graphs (POG). POG defines the partial order relation of multiple behaviors and models behavior combinations as weighted edges to merge separate behavior graphs into a joint POG. Theoretical proof verifies that POG can be generalized to any given set of multiple behaviors. Based on POG, we propose the tailored Partial Order Graph Convolutional Networks (POGCN) that convolute neighbors' information while considering the behavior relations between users and items. POGCN also introduces a partial-order BPR sampling strategy for efficient and effective multiple-behavior CF training. POGCN has been successfully deployed on the homepage of Alibaba for two months, providing recommendation services for over one billion users. Extensive offline experiments conducted on three public benchmark datasets demonstrate that POGCN outperforms state-of-the-art multi-behavior baselines across all types of behaviors. Furthermore, online A/B tests confirm the superiority of POGCN in billion-scale recommender systems.