论文链接: Robustness of Graph Neural Networks at Scale (neurips.cc)作者: Abstract 鉴于图神经网络 (GNN) 的受欢迎程度和应用的多样性,图神经网络 (GNN) 变得越来越重要。然而,现有对其对抗性攻击脆弱性的…
文章Characterizing Malicious Edges targeting on Graph Neural Networks提出基于图生成模型的不同方法,并将链接预测作为预处理来检测潜在的恶意边;文章Comparing and detection adversarial attacks for graph deep learning关注受Nettack计算出的拓扑扰动的节点的检测问题,观察由Nettack产生的 v_i 和v_i 邻居一阶距离信息...
“measure by which a geometrical object deviates from being flat”17. Although there are multiple notions of graph curvature18, this work focuses on the Ricci curvature as formulated by Ollivier19, because of its positive correlation with the robustness of a network and because of its natural ...
Figure 2 shows the change of the robustness λ and the small-world effect E versus iterations t in scale-free networks and ER random graphs, respectively. We find that optimizing λ will reduce E obviously while increasing E also leads to a great loss of robustness λ. It means that ...
Deep neural networks (DNNs) have achieved comparable performance to humans on many tasks. However, there are two problems in deploying a DDN to terminal devices of a real scene. First, DNNs consume huge computing resources and memory; therefore, it is difficult to directly apply DNNs to resourc...
2020), which design a convolutional neural network for image watermarking. However, compared with 2D data, the convolution on 3D mesh has more difficulties because of the irregularity and complexity. And 3D mesh watermarking suffers from more threats with the increased dimensional space. As graph ...
of acto-myosin within specific subcellular domains drives specific changes in cell shape. These forces, generated at a single-cell level can propagate from one cell to another throughadherens junctionor diffusible biomechanical signals (Lecuit and Yap, 2015), ultimately leading to large-scale changes...
we improve upon the speed of convergence by a factor of n, the number of players. Both the scope of settings and the class of algorithms for which our analysis provides fast convergence are considerably broader than in previous work. Our framework applies to dynamic population games via a low...
Genotype networks are sets of genotypes connected by small mutational changes that share the same phenotype. They facilitate evolutionary innovation by enabling the exploration of different neighborhoods in genotype space. Genotype networks, first sugges
two states: normal (N) and abnormal(A). The normal state denotes that the chip trade between countries (or regions) is completed at the rated scale. The abnormal state indicates that a country (or region) being undersupplied with exports owing to the weakened utility of the import channel....