SiReN: Sign-Aware Recommendation Using Graph Neural Networks

• Published in: IEEE transaction on neural networks and learning systems 2024-04, Vol.35 (4), p.4729-4743
• Main Authors: Seo, Changwon, Jeong, Kyeong-Joong, Lim, Sungsu, Shin, Won-Yong
• Format: Article
• Language: English
• Subjects: Bayesian analysis; Bayesian personalized ranking (BPR) loss; Bipartite graph; Business process re-engineering; Embedding; graph neural network (GNN); Graph neural networks; Graph theory; Graphical representations; Graphs; Information processing; Mathematical models; Multilayer perceptrons; Negative feedback; network embedding (NE); Neural networks; Optimization; Probability theory; Ratings; recommender system; Recommender systems; signed bipartite graph; Sparse matrices
• ISSN: 2162-237X; 2162-2388
• DOI: 10.1109/TNNLS.2022.3175772

In recent years, many recommender systems using network embedding (NE) such as graph neural networks (GNNs) have been extensively studied in the sense of improving recommendation accuracy. However, such attempts have focused mostly on utilizing only the information of positive user-item interactions with high ratings. Thus, there is a challenge on how to make use of low rating scores for representing users' preferences since low ratings can be still informative in designing NE-based recommender systems. In this study, we present Si ReN, a new Si gn-aware Re commender system based on GN N models. Specifically, SiReN has three key components: 1) constructing a signed bipartite graph for more precisely representing users' preferences, which is split into two edge-disjoint graphs with positive and negative edges each; 2) generating two embeddings for the partitioned graphs with positive and negative edges via a GNN model and a multilayer perceptron (MLP), respectively, and then using an attention model to obtain the final embeddings; and 3) establishing a sign-aware Bayesian personalized ranking (BPR) loss function in the process of optimization. Through comprehensive experiments, we empirically demonstrate that SiReN consistently outperforms state-of-the-art NE-aided recommendation methods.