GCIB: Graph Contrastive Information Bottleneck for Multi-Behavior Recommendation
2026-05-25 • Information Retrieval
Information Retrieval
AI summaryⓘ
The authors propose GCIB, a new method to improve recommendation systems by cleaning up extra user behavior data that can be noisy or irrelevant. They use a special technique to keep useful patterns from these extra behaviors while removing unhelpful parts, making the system focus on information that matches the main behavior of interest. Additionally, they compare cleaned features from different behaviors to make user and item representations richer. Their experiments show that GCIB creates better recommendations by handling noise and capturing relevant information more effectively.
Multi-behavior learningRecommender systemsAuxiliary behaviorGraph Information BottleneckContrastive LearningUser embeddingsItem embeddingsData sparsityGraph Neural NetworksCollaborative filtering
Authors
Likang Wu, Zihao Chen, Jianxin Zhang, Sangqi Zhu, Yuanyuan Ge, Haipeng Yang, Lei Zhang
Abstract
With the rapid emergence of multi-behavior learning in recommender systems, leveraging auxiliary user behaviors has proven effective for mitigating target-behavior data sparsity. Yet auxiliary behavior graphs frequently contain noisy or irrelevant interactions that do not align with the target task, impeding the learning of accurate user and item embeddings. Moreover, the scarcity of direct supervision from the target behavior complicates the extraction of informative collaborative signals. In this paper, we introduce GCIB (Graph Contrastive Information Bottleneck), a novel framework that denoises auxiliary behavior information and enriches target behavior representations at both the structural and feature levels. At the structural level, GCIB employs a Graph Information Bottleneck (GIB) objective to maximize mutual information between the denoised auxiliary graph and the target-behavior graph while minimizing mutual information with the original auxiliary graph. This formulation preserves task-relevant structural patterns and suppresses spurious interactions. At the feature level, we propose a cross-behavior Graph Contrastive Learning (GCL) scheme in which denoised auxiliary features and target-behavior features serve as complementary views for both users and items. By contrasting these views, GCIB enriches sparse target-behavior representations with semantics distilled from auxiliary behaviors. Extensive experiments demonstrate that GCIB outperforms state-of-the-art baselines, highlighting its ability to learn noise-resilient and target-aware representations for multi-behavior recommendation.