Description
GCNII is an extension of a Graph Convolution Networks with two new techniques, initial residual and identify mapping, to tackle the problem of oversmoothing -- where stacking more layers and adding non-linearity tends to degrade performance. At each layer, initial residual constructs a skip connection from the input layer, while identity mapping adds an identity matrix to the weight matrix.
Papers Using This Method
Understanding and Improving Deep Graph Neural Networks: A Probabilistic Graphical Model Perspective2023-01-25Multi-duplicated Characterization of Graph Structures using Information Gain Ratio for Graph Neural Networks2022-12-24Every Node Counts: Improving the Training of Graph Neural Networks on Node Classification2022-11-29wsGAT: Weighted and Signed Graph Attention Networks for Link Prediction2021-09-21Layer-wise Adaptive Graph Convolution Networks Using Generalized Pagerank2021-08-24Extract the Knowledge of Graph Neural Networks and Go Beyond it: An Effective Knowledge Distillation Framework2021-03-04Simple and Deep Graph Convolutional Networks2020-07-04