Efficient Bipartite Graph Embedding Induced by Clustering Constraints

TL;DR

The proposed Clustering Constraints induced BIpartite graph Embedding (CCBIE) significantly enhances user-item collaborative relation modeling by integrating adaptive clustering for relationship learning, thereby markedly improving prediction performance.

Abstract

Bipartite graph embedding (BGE) maps nodes to compressed embedding vectors that can reflect the hidden topological features of the network, and learning high-quality BGE is crucial for facilitating downstream applications such as recommender systems. However, most existing methods either struggle to efficiently learn embeddings suitable for users and items with fewer interactions, or exhibit poor scalability to handle large-scale networks. In this paper, we propose a Clustering Constraints induced BIpartite graph Embedding (CCBIE) as an integrated solution to both problems. CCBIE facilitates automatic and dynamic soft clustering of items in a top-down manner, and capturing macro-preference information of users through clusters. Specifically, by leveraging the cluster embedding matrix of items, CCBIE calculates the cluster assignment matrix for items and also captures the extent of user preferences across different clusters, thereby elucidating the similarity between users and items on a macro-scale level. CCBIE effectively preserves the global properties of bipartite graphs, maintaining the cluster structure of isomorphic nodes and accounting for long-range dependencies among heterogeneous nodes. Our approach significantly enhances user-item collaborative relation modeling by integrating adaptive clustering for relationship learning, thereby markedly improving prediction performance, particularly benefiting cold users and items. Extensive experiments indicate that CCBIE consistently and significantly improves accuracy over baseline models, particularly on sparse graphs, while also enhancing training speed and reducing memory requirements on large-scale bipartite graphs.

Figures (7)

• Figure 1: An illustrative example of user-item bipartite graph, along with its distinctive local and global properties, is provided. Locally, solid lines represent explicit inter-type connections, which are derived from historical interaction data. On a global scale, the dotted boxes denote potential interconnections among items, reflecting the community structure wherein items may exhibit shared genres. Additionally, the red dashed lines signify the long-term dependency between users and items.
• Figure 2: Overall framework of the proposed CCBIE model. The orange-shaded area enclosed by the elliptical box represents a underlying cluster structure. The thickness of the edges reflects the correlation score between nodes. A thicker edge suggests a stronger correlation and a higher likelihood of link formation between the nodes.
• Figure 3: Workflow of the proposed cluster-aware cross-category integration module. In addition to the user and item embedding matrix, we propose the incorporation of a cluster embedding matrix. This additional matrix is designed to capture and elucidate the nuanced connections between items and user preferences at the meso level. It can be updated through error backpropagation during subsequent training, thereby enhancing prediction accuracy and automating the clustering process.
• Figure 4: Running time.
• Figure 5: Training curves.