GT-A <sup>2</sup> T: Graph Tensor Alliance Attention Network
Ling Wang, Kechen Liu, Ye Yuan
Abstract
Dear Editor, This letter proposes the graph tensor alliance attention network (GT-A<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>T) to represent a dynamic graph (DG) precisely. Its main idea includes 1) Establishing a unified spatio-temporal message propagation framework on a DG via the tensor product for capturing the complex cohesive spatio-temporal interdependencies precisely and 2) Acquiring the alliance attention scores by node features and favorable high-order structural correlations. Empirical studies on DG benchmark datasets indicate that the proposed GT-A<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>T consistently outperforms the state-of-the-art models in the field of missing link weight estimation and link prediction.