“Fully Distributed Online Training of Graph Neural Networks in Networked Systems” submitted to IEEE ICMLCN 2025.

Our paper “Fully Distributed Online Training of Graph Neural Networks in Networked Systems” was submitted to IEEE ICMLCN 2025, the first author is Rostyslav Olshevskyi, whom I co-advise with Santiago Segarra. This is also the first paper submitted by Rostyslav Olshevskyi on his PhD program at Rice University.

• Preprint available https://arxiv.org/pdf/2412.06105
• Source code available soon

Abstract

Graph neural networks (GNNs) are powerful tools for developing scalable, decentralized artificial intelligence in large-scale networked systems, such as wireless networks, power grids, and transportation networks. Currently, GNNs in networked systems mostly follow a paradigm of ‘centralized training, distributed execution’, which limits their adaptability and slows down their development cycles. In this work, we fill this gap for the first time by developing a communication-efficient, fully distributed online training approach for GNNs applied to large networked systems. For a mini-batch with $B$ samples, our approach of training an $L$-layer GNN only adds $L$ rounds of message passing to the $LB$ rounds required by GNN inference, with doubled message sizes. Through numerical experiments in graph-based node regression, power allocation, and link scheduling in wireless networks, we demonstrate the effectiveness of our approach in training GNNs under supervised, unsupervised, and reinforcement learning paradigms.