Ultra Fast Warm Start Solution for Graph Recommendations

2025-09-01

AI Summary

To address the challenge of real-time recommendation updates in graph neural collaborative filtering models (e.g., UltraGCN) under dynamic user preferences and frequent data arrivals, this paper proposes a warm-start method based on low-rank linearized approximation. Our approach innovatively incorporates low-rank matrix approximation into the graph convolutional propagation process, circumventing costly iterative retraining and enabling efficient incremental parameter updates. While preserving the original UltraGCN architecture, our method achieves rapid recommendation refresh via a single linear transformation—accelerating inference by up to 30×—while outperforming baselines in AUC, Recall, and other key metrics. Extensive experiments on large-scale industrial datasets demonstrate superior scalability and robustness. The proposed framework establishes a lightweight, efficient, and production-ready paradigm for real-time recommender systems operating under high-frequency update constraints.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Recommender Systems
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

In this work, we present a fast and effective Linear approach for updating recommendations in a scalable graph-based recommender system UltraGCN. Solving this task is extremely important to maintain the relevance of the recommendations under the conditions of a large amount of new data and changing user preferences. To address this issue, we adapt the simple yet effective low-rank approximation approach to the graph-based model. Our method delivers instantaneous recommendations that are up to 30 times faster than conventional methods, with gains in recommendation quality, and demonstrates high scalability even on the large catalogue datasets.

Problem

• Ultra fast warm start for graph recommendations
• Updating recommendations with new data changes
• Maintaining relevance despite shifting user preferences

Innovation

• Linear approach for updating graph recommendations
• Low-rank approximation to graph-based model
• Instantaneous recommendations 30 times faster