Dense multi-scale graph token transformer for local-global graph representation learning

2026-08-06

PRE

Dongdong Li *

Jinchen Han

Bohan Yang

Zhe Wang *

DOI：10.1016/j.eswa.2026.133767

Original

Original request for help

Save

Abstract

En 中文

• The framework unifies dense multi-scale graph encoding and topology-aware attention. • Scale-aware token learning adaptively fuses representations across depths. • Topology-guided attention captures long-range interactions among graph nodes. • The framework outperforms representative baselines across four benchmark datasets.

Keywords:

Graph representation learning

Dense graph convolution

Graph token learning

Topology-aware aggregation

EEG emotion recognition

Dense multi-scale graph token transformer for local-global graph representation learning

Abstract

Journal

Organization

Cited Papers

Citing Papers