Oversquashing可以理解为,当前节点i没有受到与节点i距离为r的节点s的某些输入特征的影响。理论上,作者从雅可比 \partial h_{i}^{(r)} / \partial x_{s} \\ 来度量过度挤压效应。 「Lemma 1. Let i, s \in V with s \in S_{r+1}(i). If \left|\nabla \phi_{\ell}\right| \leq \alpha...
在图神经网络(GNN)中,"over-smoothing" 和 "over-squashing" 是两种不同的问题,它们影响网络的性能和学习能力。 1. Over-Smoothing: 定义:Over-smoothing 是指随着图神经网络层数的增加,节点特征变得越来越相似,最终在高层次上收敛到一个相似或相同的状态。这导致不同节点之间的特征区分度降低,使得GNN难以捕捉到...
Graph Neural Networks (GNNs) have succeeded in various computer science applications, yet deep GNNs underperform their shallow counterparts despite deep learning's success in other domains. Over-smoothing and over-squashing are key challenges when stacking graph convolutional layers, hindering deep represe...
Repository of the paper "On the Trade-off between Over-smoothing and Over-squashing in Deep Graph Neural Networks" published in ACM CIKM 2023 - jhonygiraldo/SJLR
More focus on bug squashing over feature adding 09-30-2016 04:24 PM Another month, another monthly update... While it is nice to see some features being added here and there (even though I am puzzled on the criteria of which features take priority -- as it certainly isn'...