Non-parametric Graph Convolution for Re-ranking in Recommendation Systems

2025-07-14

AI Summary

To address the low efficiency of graph knowledge utilization and high computational overhead of neighborhood retrieval in the ranking stage of recommender systems, this paper proposes a non-parametric graph convolutional re-ranking method (NGCR). NGCR dynamically constructs local graph structures during inference, requiring only real-time sparse neighborhood retrieval for target items—thereby eliminating the expensive graph convolution computations and model coupling inherent in training-phase graph neural networks. It is plug-and-play compatible with existing ranking systems. Technically, NGCR fully decouples graph structural modeling from training, integrating context-aware neighborhood aggregation with a distributed-friendly, real-time graph inference mechanism. Evaluated on four public benchmarks, NGCR achieves an average 8.1% improvement in ranking performance with only 0.5% additional latency, striking a strong balance among effectiveness, efficiency, and deployment versatility.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Graph knowledge has been proven effective in enhancing item rankings in recommender systems (RecSys), particularly during the retrieval stage. However, its application in the ranking stage, especially when richer contextual information in user-item interactions is available, remains underexplored. A major challenge lies in the substantial computational cost associated with repeatedly retrieving neighborhood information from billions of items stored in distributed systems. This resource-intensive requirement makes it difficult to scale graph-based methods in practical RecSys. To bridge this gap, we first demonstrate that incorporating graphs in the ranking stage improves ranking qualities. Notably, while the improvement is evident, we show that the substantial computational overheads entailed by graphs are prohibitively expensive for real-world recommendations. In light of this, we propose a non-parametric strategy that utilizes graph convolution for re-ranking only during test time. Our strategy circumvents the notorious computational overheads from graph convolution during training, and utilizes structural knowledge hidden in graphs on-the-fly during testing. It can be used as a plug-and-play module and easily employed to enhance the ranking ability of various ranking layers of a real-world RecSys with significantly reduced computational overhead. Through comprehensive experiments across four benchmark datasets with varying levels of sparsity, we demonstrate that our strategy yields noticeable improvements (i.e., 8.1% on average) during testing time with little to no additional computational overheads (i.e., 0.5 on average). Code: https://github.com/zyouyang/RecSys2025_NonParamGC.git

Problem

• Enhancing item rankings in recommender systems using graph knowledge.
• Reducing computational costs of graph-based methods in ranking stages.
• Proposing a non-parametric graph convolution for efficient re-ranking.

Innovation

• Non-parametric graph convolution for re-ranking
• Plug-and-play module reduces computational overhead
• On-the-fly structural knowledge utilization during testing