ScaDyG:A New Paradigm for Large-scale Dynamic Graph Learning

TL;DR

ScaDyG introduces a scalable, time-aware paradigm for dynamic graph learning that improves efficiency and performance in large-scale, evolving graph scenarios by reformulating topology, encoding temporal information, and using hypernetworks for message aggregation.

Contribution

The paper presents a novel scalable learning framework for dynamic graphs that addresses efficiency issues of traditional methods through innovative topology reformulation and temporal encoding.

Findings

Achieves comparable or better performance than SOTA methods on 12 datasets.

Uses fewer learnable parameters and offers higher efficiency.

Effective in both node and link-level tasks.

Abstract

Dynamic graphs (DGs), which capture time-evolving relationships between graph entities, have widespread real-world applications. To efficiently encode DGs for downstream tasks, most dynamic graph neural networks follow the traditional message-passing mechanism and extend it with time-based techniques. Despite their effectiveness, the growth of historical interactions introduces significant scalability issues, particularly in industry scenarios. To address this limitation, we propose ScaDyG, with the core idea of designing a time-aware scalable learning paradigm as follows: 1) Time-aware Topology Reformulation: ScaDyG first segments historical interactions into time steps (intra and inter) based on dynamic modeling, enabling weight-free and time-aware graph propagation within pre-processing. 2) Dynamic Temporal Encoding: To further achieve fine-grained graph propagation within time steps, ScaDyG integrates temporal encoding through a combination of exponential functions in a scalable manner. 3) Hypernetwork-driven Message Aggregation: After obtaining the propagated features (i.e., messages), ScaDyG utilizes hypernetwork to analyze historical dependencies, implementing node-wise representation by an adaptive temporal fusion. Extensive experiments on 12 datasets demonstrate that ScaDyG performs comparably well or even outperforms other SOTA methods in both node and link-level downstream tasks, with fewer learnable parameters and higher efficiency.