MMODELYST
Papers/Gaussian Relational Graph Transformer
PAP

Gaussian Relational Graph Transformer

May 15, 2026

arXiv
Abstract

Relational graph learning models relational databases as graphs and has demonstrated superior performance on a wide range of relational predictive tasks. However, existing methods struggle to capture long-range dependencies due to information decay in their message-passing mechanisms, and recent relational graph transformers remain limited in jointly modeling structural, semantic, and temporal information. In this paper, we propose GelGT, a Gaussian relational graph transformer that explicitly addresses these challenges. GelGT introduces a structure-semantic collaborative sampling strategy to preserve structural connectivity while filtering irrelevant semantic information, and incorporates a Gaussian graph attention mechanism with a learnable Gaussian bias on the sampled subgraphs to dynamically encode temporal dependencies. Extensive experiments on various real-world datasets demonstrate that GelGT achieves state-of-the-art downstream task performance, with up to a 13.8% improvement in predictive performance.

Select text to highlight · click a highlight to remove · saved in this browser only
Authors
Zezhong Ding, Jin Li, Xugang Wang, Xike Xie
Your notes (browser-local)
saved
arXiv:2605.15575