arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2501.14440cs.LGcs.DMcs.NAcs.SImath.NA
收藏

线性图神经网络基于梯度的训练的收敛性

Convergence of gradient based training for linear Graph Neural Networks

RWTH Aachen University(亚琛工业大学)

Dhiraj Patel, Anton Savostianov, Michael T. Schaub

AI总结:

本文证明线性图神经网络在均方损失下梯度流训练以指数速率收敛至全局最小值,速率依赖初始权重和图移位算子,并验证于合成与真实数据集,同时探讨了梯度下降的收敛性。

摘要:

图神经网络(GNNs)是解决图结构学习问题的强大工具,在分子生物学和社交网络等领域有广泛应用。然而,其经验性能背后的理论基础尚未被充分理解。本文考察了线性GNN训练中梯度动力学的收敛性。具体而言,我们证明了在均方损失下,线性GNN的梯度流训练以指数速率收敛到全局最小值。收敛速率明确依赖于初始权重和图移位算子,我们在来自著名图模型的合成数据集和真实世界数据集上验证了这一点。此外,我们讨论了在全局最小值处最小化总权重的梯度流。除梯度流外,我们还研究了梯度下降训练下线性GNN的收敛性,这是一种被视为梯度流离散化的迭代方案。

英文摘要:

Graph Neural Networks (GNNs) are powerful tools for addressing learning problems on graph structures, with a wide range of applications in molecular biology and social networks. However, the theoretical foundations underlying their empirical performance are not well understood. In this article, we examine the convergence of gradient dynamics in the training of linear GNNs. Specifically, we prove that the gradient flow training of a linear GNN with mean squared loss converges to the global minimum at an exponential rate. The convergence rate depends explicitly on the initial weights and the graph shift operator, which we validate on synthetic datasets from well-known graph models and real-world datasets. Furthermore, we discuss the gradient flow that minimizes the total weights at the global minimum. In addition to the gradient flow, we study the convergence of linear GNNs under gradient descent training, an iterative scheme viewed as a discretization of gradient flow.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑