【论文阅读】Model Stealing Attacks Against Inductive Graph Neural Networks（2021）

摘要

Many real-world data（真实世界的数据） come in the form of graphs（以图片的形式）. Graph neural networks (GNNs 图神经网络), a new family of machine learning (ML) models, have been proposed to fully leverage graph data（充分利用图数据） to build powerful applications（构建强大的应用程序）. In particular（特别是）, the inductive GNNs（归纳 GNNs）, which can generalize to unseen data（泛化到不可见的数据）, become mainstream in this direction（成为各个方向的主流）. Machine learning models have shown great potential（巨大的潜力） in various tasks（各种任务上） and have been deployed（部署） in many real-world scenarios（在许多现实场景）. To train a good model（为了训练一个好的模型）, a large amount of data（大量的数据） as well as computational resources（计算资源） are needed, leading to（产生） valuable intellectual property（宝贵的知识产权）. Previous researc