UniDyG: A Unified and Effective Representation Learning Approach for Large Dynamic Graphs

2025-02-23

AI Summary

This paper addresses the challenge of unified modeling for both continuous-time and discrete-time dynamic graphs (CTDGs/DTDGs). We propose the first Fourier Graph Attention Mechanism (FGAT) designed for temporal consistency, which employs complex-domain selective aggregation and energy-gated filtering to suppress high-frequency temporal noise, thereby enhancing robustness in structural evolution modeling. Furthermore, we introduce a frequency-domain enhanced linear dynamic update module to jointly capture local abrupt changes and global evolutionary patterns. Theoretically, we establish temporal consistency guarantees grounded in approximation theory; practically, our framework enables efficient large-scale graph training. Evaluated on nine benchmark datasets, our method achieves an average performance gain of 14.4% over 16 strong baselines, significantly improving dynamic representation learning under heterogeneous temporal granularities.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Planning, Routing, and Scheduling: Temporal Planning

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Web Mining and Content Analysis: Models for Web evolution

Abstract

Dynamic graphs are formulated in continuous-time or discrete-time dynamic graphs. They differ in temporal granularity: Continuous-Time Dynamic Graphs (CTDGs) exhibit rapid, localized changes, while Discrete-Time Dynamic Graphs (DTDGs) show gradual, global updates. This difference leads to isolated developments in representation learning for each type. To advance representation learning, recent research attempts to design a unified model capable of handling both CTDGs and DTDGs. However, it typically focuses on local dynamic propagation for temporal structure learning in the time domain, failing to accurately capture the structural evolution associated with each temporal granularity. In addition, existing works-whether specific or unified-often overlook the issue of temporal noise, compromising the model robustness and effectiveness. To better model both types of dynamic graphs, we propose UniDyG, a unified and effective representation learning approach, which scales to large dynamic graphs. We first propose a novel Fourier Graph Attention (FGAT) mechanism that can model local and global structural correlations based on recent neighbors and complex-number selective aggregation, while theoretically ensuring consistent representations of dynamic graphs over time. Based on approximation theory, we demonstrate that FGAT is well-suited to capture the underlying structures in CTDGs and DTDGs. We further enhance FGAT to resist temporal noise by designing an energy-gated unit, which adaptively filters out high-frequency noise according to the energy. Last, we leverage our FGAT mechanisms for temporal structure learning and employ the frequency-enhanced linear function for node-level dynamic updates, facilitating the generation of high-quality temporal embeddings. Extensive experiments show that our UniDyG achieves an average improvement of 14.4% over sixteen baselines across nine dynamic graphs.

Problem

• Unifies representation learning for dynamic graphs.
• Addresses temporal noise in dynamic graph models.
• Improves structural evolution capture across temporal granularities.

Innovation

• Unified model for dynamic graphs
• Fourier Graph Attention mechanism
• Energy-gated noise filtering unit