EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

An approach of dual attention-driven multi-levelled graph neural architecture for rich-structural social graph analysis and representation learning

Linh Nguyen Thi My iD, Tham Vo iD

PLOS One · 26.06.2026 · 21 (6) · e0351922

DOI: 10.1371/journal.pone.0351922 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS One, online erschienen am 26.06.2026. Verfasst von Linh Nguyen Thi My und 1 weiteren Autor:innen. Beteiligte Einrichtung: Faculty of Information Technology, Van Lang School of Technology, Van Lang University, Ho Chi Minh City, Vietnam. Hauptthemen laut MEDLINE: Graph Neural Networks, Attention. Die Arbeit nennt 27 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Van Lang University, Ho Chi Minh City, Vietnam, Nguyen Tat Thanh University, Ho Chi Minh City, Vietnam.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In recent years, graph neural networks (GNNs) have emerged as a dominant paradigm for learning from graph-structured data, achieving strong performance across a wide range of graph mining tasks, including node classification, clustering, and relational inference. Despite these advances, most existing GNN architectures primarily rely on message-passing mechanisms that aggregate smoothed high-order local neighborhood information. While effective for capturing local proximity patterns, such designs often struggle to model richer multi-level graph structures, particularly in scenarios where both fine-grained local interactions and holistic global organization are essential. From a structural perspective, graph representation learning inherently involves information distributed across multiple levels. Local proximity cues, such as node degrees and common neighbors, capture immediate relational patterns, whereas global structural characteristics, including long-range dependencies and overall graph topology, provide complementary high-order information. However, conventional GNN models typically entangle these heterogeneous signals within a unified aggregation process, leading to suboptimal representations. To address this limitation, we propose DAGRL, a dual attention-based graph representation learning framework that explicitly disentangles and integrates multi-level structural information. The proposed model employs dedicated graph neural components to learn structure-specific representations and utilizes a dual attention mechanism to adaptively fuse them into a unified embedding space. Extensive experiments demonstrate that DAGRL consistently outperforms existing methods across multiple benchmark datasets.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Graph Neural NetworksAttentionRepresentation Machine LearningAlgorithmsHumans

Autor:innen und Institutionen

  • Faculty of Information Technology, Van Lang School of Technology, Van Lang University, Ho Chi Minh City, Vietnam.
  • Faculty of Information Technology, Nguyen Tat Thanh University, Ho Chi Minh City, Vietnam.

ORCID-Profile: Linh Nguyen Thi My, Tham Vo

Förderung

  • Van Lang University, Ho Chi Minh City, Vietnam
  • Nguyen Tat Thanh University, Ho Chi Minh City, Vietnam

Zitieren

APA

My, L. N. T., & Vo, T. (2026). An approach of dual attention-driven multi-levelled graph neural architecture for rich-structural social graph analysis and representation learning. PLOS One, 21(6), e0351922. https://doi.org/10.1371/journal.pone.0351922

Vancouver

My LNT, Vo T. An approach of dual attention-driven multi-levelled graph neural architecture for rich-structural social graph analysis and representation learning. PLOS One. 2026;21(6):e0351922. doi:10.1371/journal.pone.0351922

BibTeX
@article{My2026,
  author = {My, Linh Nguyen Thi and Vo, Tham},
  title = {An approach of dual attention-driven multi-levelled graph neural architecture for rich-structural social graph analysis and representation learning},
  journal = {PLOS One},
  year = {2026},
  volume = {21},
  number = {6},
  pages = {e0351922},
  doi = {10.1371/journal.pone.0351922}
}
Kostenlos starten →