Hierarchical Frequency-Decomposition Graph Neural Networks for Road Network Representation Learning

2025-11-16

AI Summary

Road network representation learning faces the challenge of jointly modeling spatial structure and frequency characteristics: existing graph neural networks either emphasize local spatial smoothing—leading to distortion—or focus on global spectral analysis—neglecting local dynamics—resulting in spatial-spectral mismatch. To address this, we propose HiFiNet, the first hierarchical frequency-domain decomposition framework. It constructs a multi-scale frequency hierarchy via virtual nodes and implements a “decompose-update-reconstruct” pipeline. Furthermore, it introduces a topology-aware Graph Transformer to enable separate modeling and adaptive fusion of high- and low-frequency signals. Evaluated on four real-world traffic datasets across multiple downstream tasks, HiFiNet consistently outperforms state-of-the-art methods. It achieves fine-grained characterization of both global trends and local fluctuations, thereby enhancing representation discriminability and cross-scenario generalizability.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Machine Learning: Graph-based Machine Learning
• Computer Vision: Representation Learning for Vision

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks

Abstract

Road networks are critical infrastructures underpinning intelligent transportation systems and their related applications. Effective representation learning of road networks remains challenging due to the complex interplay between spatial structures and frequency characteristics in traffic patterns. Existing graph neural networks for modeling road networks predominantly fall into two paradigms: spatial-based methods that capture local topology but tend to over-smooth representations, and spectral-based methods that analyze global frequency components but often overlook localized variations. This spatial-spectral misalignment limits their modeling capacity for road networks exhibiting both coarse global trends and fine-grained local fluctuations. To bridge this gap, we propose HiFiNet, a novel hierarchical frequency-decomposition graph neural network that unifies spatial and spectral modeling. HiFiNet constructs a multi-level hierarchy of virtual nodes to enable localized frequency analysis, and employs a decomposition-updating-reconstruction framework with a topology-aware graph transformer to separately model and fuse low- and high-frequency signals. Theoretically justified and empirically validated on multiple real-world datasets across four downstream tasks, HiFiNet demonstrates superior performance and generalization ability in capturing effective road network representations.

Problem

• Modeling complex spatial-frequency interplay in road networks
• Overcoming over-smoothing in spatial and spectral graph methods
• Capturing both global trends and local traffic fluctuations

Innovation

• Hierarchical frequency-decomposition graph neural network
• Multi-level virtual nodes for localized frequency analysis
• Decomposition-updating-reconstruction framework with graph transformer