What question did this study set out to answer?

To design a scalable, computationally efficient dynamic graph neural network framework that mitigates temporal embedding drift and preserves spatiotemporal topological structure.

August 16, 2026Mesopotamian Journal of Big DataOpen Access

Topology-Aware Graph Neural Network with Spatiotemporal Encoding and Contrastive Learning for Scalable Dynamic Network Representation

View Full Paper

Ask AI

Bookmark

Authors

WHWassan S. HayaleShaqra University RARasha S. AliShaqra University RHRaghda Abd Ul Rab Abd Ul HasanIraqi University

Discussion

Member takes

Implication

Computational study demonstrates improved accuracy and efficiency in dynamic graph representation, indicating enhanced scalability and stability for complex evolving networks.

Key Points

To design a scalable, computationally efficient dynamic graph neural network framework that mitigates temporal embedding drift and preserves spatiotemporal topological structure.
Integrated graph Laplacian feature transformations and position encodings to capture both global and local structural graph properties.
Combined gated recurrent units (GRUs) with a self-attention mechanism to model complex temporal relationships.
Applied neighbor sampling, hierarchical clustering, and mini-batch training to reduce computational costs on social, traffic flow, and biological network datasets.
TAGNN achieved a 92.4% node-classification accuracy, outperforming dynamic graph baselines including TGN, EvolveGCN, and DySAT.
Demonstrated a 32-fold reduction in training time, scaled to process up to 1 million edges, and improved temporal stability by approximately 9.1%.

Cite This Study

Hayale et al. (2026) studied this question.

synapsesocial.com/papers/6a817ab4f2fb91fc834aeb1e https://doi.org/10.58496/2958-6453.1084

View Full Paper

Ask AI

Bookmark

Implication

Computational study demonstrates improved accuracy and efficiency in dynamic graph representation, indicating enhanced scalability and stability for complex evolving networks.

Topology-Aware Graph Neural Network with Spatiotemporal Encoding and Contrastive Learning for Scalable Dynamic Network Representation

Discussion

Member takes

Implication

Key Points

Cite This Study

Topology-Aware Graph Neural Network with Spatiotemporal Encoding and Contrastive Learning for Scalable Dynamic Network Representation

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider