Genetic-GNN: Evolutionary architecture search for Graph Neural Networks
Genetic-GNN: Evolutionary architecture search for Graph Neural Networks
复制标题
DOI：
10.1016/j.knosys.2022.108752
复制
发表时间：
2022-07-08
影响因子：
8.8
通讯作者：
Liu, Jianxun
中科院分区：
文献类型：
--
作者：
Shi, Min;Tang, Yufei;Liu, Jianxun
神经体系结构搜索(NAS)在整个计算智能研究界引起了极大的关注，并推动了许多神经模型的发展，以处理文本和图像等网格状数据。然而，致力于非结构化网络数据的图形神经网络(GNN)模型所做的工作很少。鉴于诸如聚合器和激活函数等组件的大量选择和组合，确定适用于特定问题的GNN模型通常需要大量的专业知识和艰苦的试验。此外，学习速率和辍学率等超参数的适度变化将极大地影响GNN模型的学习能力。在本文中，我们通过在大型GNN架构搜索空间中个体模型的进化，提出了一种新的框架。不是简单地优化模型结构，而是在GNN模型结构和超参数之间进行交替进化过程，以动态逼近彼此的最优拟合度。实验和验证表明，进化NAS能够与现有最先进的强化学习方法相匹配，用于引导和归纳图的表示学习和节点分类。(C)2022爱思唯尔B.V.保留所有权利。
Neural architecture search (NAS) has seen significant attention throughout the computational intelligence research community and has pushed forward the state-of-the-art of many neural models to address grid-like data such as texts and images. However, little work has been done on Graph Neural Network (GNN) models dedicated to unstructured network data. Given the huge number of choices and combinations of components such as aggregators and activation functions, determining the suitable GNN model for a specific problem normally necessitates tremendous expert knowledge and laborious trials. In addition, the moderate change of hyperparameters such as the learning rate and dropout rate would dramatically impact the learning capacity of a GNN model. In this paper, we propose a novel framework through the evolution of individual models in a large GNN architecture searching space. Instead of simply optimizing the model structures, an alternating evolution process is performed between GNN model structures and hyperparameters to dynamically approach the optimal fit of each other. Experiments and validations demonstrate that evolutionary NAS is capable of matching existing state-of-the-art reinforcement learning methods for both transductive and inductive graph representation learning and node classification. (c) 2022 Elsevier B.V. All rights reserved.