Task-Aware Retrieval Augmentation for Dynamic Recommendation

2025-11-16

AI Summary

In dynamic recommendation, pre-trained dynamic graph neural networks (GNNs) suffer from degraded generalization during fine-tuning due to temporal distribution shift, hindering effective modeling of evolving user preferences. To address this, we propose TarDGR, a task-aware retrieval-augmented framework. First, it introduces a task-aware evaluation mechanism to automatically retrieve semantically relevant historical subgraphs. Then, a graph Transformer jointly models subgraph structural topology and semantic relevance, integrating retrieved information to enhance current prediction. This work is the first to incorporate retrieval augmentation into generalization optimization for dynamic graph recommendation, simultaneously preserving temporal consistency and capturing structural-semantic dependencies. Extensive experiments on multiple large-scale dynamic graph datasets demonstrate that TarDGR significantly improves both recommendation accuracy and cross-temporal generalization performance, validating its effectiveness and robustness.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Planning, Routing, and Scheduling: Temporal Planning

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

Dynamic recommendation systems aim to provide personalized suggestions by modeling temporal user-item interactions across time-series behavioral data. Recent studies have leveraged pre-trained dynamic graph neural networks (GNNs) to learn user-item representations over temporal snapshot graphs. However, fine-tuning GNNs on these graphs often results in generalization issues due to temporal discrepancies between pre-training and fine-tuning stages, limiting the model's ability to capture evolving user preferences. To address this, we propose TarDGR, a task-aware retrieval-augmented framework designed to enhance generalization capability by incorporating task-aware model and retrieval-augmentation. Specifically, TarDGR introduces a Task-Aware Evaluation Mechanism to identify semantically relevant historical subgraphs, enabling the construction of task-specific datasets without manual labeling. It also presents a Graph Transformer-based Task-Aware Model that integrates semantic and structural encodings to assess subgraph relevance. During inference, TarDGR retrieves and fuses task-aware subgraphs with the query subgraph, enriching its representation and mitigating temporal generalization issues. Experiments on multiple large-scale dynamic graph datasets demonstrate that TarDGR consistently outperforms state-of-the-art methods, with extensive empirical evidence underscoring its superior accuracy and generalization capabilities.

Problem

• Addresses temporal generalization issues in dynamic recommendation systems
• Enhances GNN fine-tuning by incorporating task-aware retrieval augmentation
• Mitigates evolving user preference modeling challenges through subgraph fusion

Innovation

• Task-Aware Evaluation Mechanism identifies relevant historical subgraphs
• Graph Transformer-based Model integrates semantic and structural encodings
• Retrieval-augmented framework fuses task-aware subgraphs during inference