Dynamic Localisation of Spatial-Temporal Graph Neural Network

2025-01-08

AI Summary

To address the inefficiency in modeling dynamic spatial dependencies and the high communication overhead in distributed settings for spatiotemporal graph learning, this paper proposes DynAGS, a Dynamic Adaptive Graph Learning framework. DynAGS explicitly models spatial dependency evolution as a time-varying process, jointly leveraging a dynamic graph generator and cross-attention mechanisms to enable spatiotemporal-aware graph structure adaptation. It further introduces a lightweight, node-agnostic historical fusion module and personalized local subgraph partitioning to construct sparse, dynamic spatial graphs. Evaluated on nine real-world spatiotemporal datasets, DynAGS consistently outperforms state-of-the-art methods—including ASTGNN—achieving higher prediction accuracy, reducing distributed inference latency, increasing graph sparsity by 3–5×, and decreasing inter-node communication overhead by over 40%.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Agentic search

Abstract

Spatial-temporal data, fundamental to many intelligent applications, reveals dependencies indicating causal links between present measurements at specific locations and historical data at the same or other locations. Within this context, adaptive spatial-temporal graph neural networks (ASTGNNs) have emerged as valuable tools for modelling these dependencies, especially through a data-driven approach rather than pre-defined spatial graphs. While this approach offers higher accuracy, it presents increased computational demands. Addressing this challenge, this paper delves into the concept of localisation within ASTGNNs, introducing an innovative perspective that spatial dependencies should be dynamically evolving over time. We introduce extit{DynAGS}, a localised ASTGNN framework aimed at maximising efficiency and accuracy in distributed deployment. This framework integrates dynamic localisation, time-evolving spatial graphs, and personalised localisation, all orchestrated around the Dynamic Graph Generator, a light-weighted central module leveraging cross attention. The central module can integrate historical information in a node-independent manner to enhance the feature representation of nodes at the current moment. This improved feature representation is then used to generate a dynamic sparse graph without the need for costly data exchanges, and it supports personalised localisation. Performance assessments across two core ASTGNN architectures and nine real-world datasets from various applications reveal that extit{DynAGS} outshines current benchmarks, underscoring that the dynamic modelling of spatial dependencies can drastically improve model expressibility, flexibility, and system efficiency, especially in distributed settings.

Problem

• Graph Neural Networks
• Spatio-Temporal Data
• Distributed Systems

Innovation

• DynAGS
• Adaptive Spatio-Temporal Graph Neural Networks
• Cross-Attention Mechanism