Neural networks : the official journal of the International Neural Network Society
MoGL: A mixture of heterogeneous experts for collaborative graph learning.
Gonghai Zhou, Zhiwei Xu, Kaixuan Yang, Jiatai Wang, Zhengxian Lu, Tao Li
Published: 202510.1016/j.neunet.2025.108503
Abstract
Graph Neural Networks (GNNs) have demonstrated remarkable success, yet they often exhibit limitations in capturing the complex local heterogeneity inherent in real-world graphs. While the Mixture of Experts (MoE) paradigm offers a promising direction…
Preview only. Read the full abstract at the source