1 code implementation • 5 Dec 2023 • Jiahang Li, Yakun Song, Xiang Song, David Paul Wipf
In this paper, we analyze the variance of forward and backward propagation across GNN layers and show that the variance instability of GNN initializations comes from the combined effect of the activation function, hidden dimension, graph structure and message passing.
no code implementations • 11 Apr 2022 • Aditya Chattopadhyay, Xi Zhang, David Paul Wipf, Himanshu Arora, Rene Vidal
The architecture consists of a graph encoder that maps the input graph to a structured latent space, and a graph decoder that generates a furniture graph, given a latent code and the room graph.
no code implementations • 23 Nov 2021 • Xiang Song, Runjie Ma, Jiahang Li, Muhan Zhang, David Paul Wipf
However, wider hidden layers can easily lead to overfitting, and incrementally adding more GNN layers can potentially result in over-smoothing. In this paper, we present a model-agnostic methodology, namely Network In Graph Neural Network (NGNN ), that allows arbitrary GNN models to increase their model capacity by making the model deeper.
Ranked #1 on Link Property Prediction on ogbl-citation2