[论文解读] Graph Neural Networks: Methods, Applications, and Opportunities

Lilapati Waikhom, Ripon Patgiri|arXiv (Cornell University)|Aug 24, 2021

Advanced Graph Neural Networks参考文献 192被引用 31

一句话总结

对图神经网络（GNN）在各种学习设置的全面综述，提供分类、架构、数据集、应用、挑战与未来方向。

ABSTRACT

In the last decade or so, we have witnessed deep learning reinvigorating the machine learning field. It has solved many problems in the domains of computer vision, speech recognition, natural language processing, and various other tasks with state-of-the-art performance. The data is generally represented in the Euclidean space in these domains. Various other domains conform to non-Euclidean space, for which graph is an ideal representation. Graphs are suitable for representing the dependencies and interrelationships between various entities. Traditionally, handcrafted features for graphs are incapable of providing the necessary inference for various tasks from this complex data representation. Recently, there is an emergence of employing various advances in deep learning to graph data-based tasks. This article provides a comprehensive survey of graph neural networks (GNNs) in each learning setting: supervised, unsupervised, semi-supervised, and self-supervised learning. Taxonomy of each graph based learning setting is provided with logical divisions of methods falling in the given learning setting. The approaches for each learning task are analyzed from both theoretical as well as empirical standpoints. Further, we provide general architecture guidelines for building GNNs. Various applications and benchmark datasets are also provided, along with open challenges still plaguing the general applicability of GNNs.

研究动机与目标

• 定义图及基于图的任务的基本术语与变体。
• 提出适用于每种学习设置（监督、无监督、半监督、自监督）的 GNN 变体分类。
• 从理论与经验角度分析跨学习设置的 GNN 方法。
• 提供设计 GNN 架构的一般指南。
• 调研数据集、应用与开放挑战，为后续研究提供指引。

提出的方法

• 描述 GNNs 中使用的图表示及信息传递框架。
• 为跨学习设置（监督、无监督、半监督、自监督）开发 GNN 变体分类。
• 将节点、边、图及子图层面的任务进行分类并讨论相应方法。
• 总结 GNN 研究中的数据集、应用与基准实践。
• 概述可扩展与鲁棒的 GNN 架构的设计指南与考量。
• 深入讨论开放挑战与未来研究方向。

实验结果

研究问题

• RQ1 在不同学习设置（监督、无监督、半监督、自监督）中，主流的 GNN 方法有哪些？
• RQ2 图变体和任务层级如何影响跨领域的 GNN 设计与性能？
• RQ3 哪些理论与经验见解能解释 GNN 的成功与局限，未来研究方向是什么？
• RQ4 哪些数据集与基准最好地反映 GNN 在各任务中的能力与挑战？
• RQ5 可以为构建有效且可扩展的 GNN 架构制定哪些指南？

主要发现

• 为每种学习设置提供了新的 GNN 分类体系。
• 提供跨任务与领域的 GNN 方法的理论与经验分析。
• 提供构建 GNN 的一般架构指南。
• 记录了在基于图的学习中广泛使用的数据集与应用。
• 深入识别开放挑战并提出未来研究方向（深度、可扩展性、高阶结构、鲁棒性）。