arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2103.13944cs.LG
收藏

维护、提升还是攻击?GNN的拓扑扰动解释

Preserve, Promote, or Attack? GNN Explanation via Topology Perturbation

MIT(麻省理工学院);IBM Research(IBM研究院);Michigan State University(密歇根州立大学)

Yi Sun, Abel Valente, Sijia Liu, Dakuo Wang

摘要:

关于图神经网络(GNN)的形式化解释的先前工作集中于单个用例-通过识别重要的边和节点来保留预测结果。在本文中,我们开发了一个多用途的解释框架,通过获取一个掩码,表示输入图的拓扑扰动。我们将该框架打包成一个交互式可视化系统(GNNViz),该系统可以实现多种目的:保存,提升或攻击GNN的预测。我们通过三个案例来说明我们的方法的新颖性和有效性:第一,GNNViz可以帮助非专家用户轻松地探索图拓扑和GNN决策之间的关系(保存),或操纵预测第二,在Pokec社交网络数据集上,我们的框架可以发现不公平和人口统计偏见;最后,它与合成数据集上最先进的GNN解释器基线进行了比较。

英文摘要:

Prior works on formalizing explanations of a graph neural network (GNN) focus on a single use case - to preserve the prediction results through identifying important edges and nodes. In this paper, we develop a multi-purpose interpretation framework by acquiring a mask that indicates topology perturbations of the input graphs. We pack the framework into an interactive visualization system (GNNViz) which can fulfill multiple purposes: Preserve,Promote, or Attack GNN's predictions. We illustrate our approach's novelty and effectiveness with three case studies: First, GNNViz can assist non expert users to easily explore the relationship between graph topology and GNN's decision (Preserve), or to manipulate the prediction (Promote or Attack) for an image classification task on MS-COCO; Second, on the Pokec social network dataset, our framework can uncover unfairness and demographic biases; Lastly, it compares with state-of-the-art GNN explainer baseline on a synthetic dataset.

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。

↑