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.

翻译：图神经网络（GNNs）是开发大规模网络化系统（如无线网络、电网和交通网络）中可扩展、去中心化人工智能的强大工具。目前，网络化系统中的GNNs主要遵循“集中式训练，分布式执行”的范式，这限制了其适应性并延缓了其开发周期。在本工作中，我们首次填补了这一空白，为应用于大规模网络化系统的GNNs开发了一种通信高效的、完全分布式的在线训练方法。对于一个包含$B$个样本的小批量，我们的方法训练一个$L$层GNN仅需在GNN推理所需的$LB$轮消息传递基础上增加$L$轮，且消息大小加倍。通过在基于图的节点回归、无线网络中的功率分配和链路调度等数值实验中，我们证明了我们的方法在监督学习、无监督学习和强化学习范式下训练GNNs的有效性。