GraphMix: Improved Training of GNNs for Semi-Supervised Learning
This work addresses the problem of improving training efficiency and generalization for GNNs in semi-supervised learning, which is incremental as it builds on existing GNN architectures with a new regularization technique.
The paper tackles semi-supervised object classification with Graph Neural Networks (GNNs) by introducing GraphMix, a regularization method that jointly trains a fully-connected network with the GNN via parameter sharing and interpolation-based regularization, resulting in consistent improvements or close matches to state-of-the-art performance across multiple datasets, including Cora, Citeseer, and Pubmed.
We present GraphMix, a regularization method for Graph Neural Network based semi-supervised object classification, whereby we propose to train a fully-connected network jointly with the graph neural network via parameter sharing and interpolation-based regularization. Further, we provide a theoretical analysis of how GraphMix improves the generalization bounds of the underlying graph neural network, without making any assumptions about the "aggregation" layer or the depth of the graph neural networks. We experimentally validate this analysis by applying GraphMix to various architectures such as Graph Convolutional Networks, Graph Attention Networks and Graph-U-Net. Despite its simplicity, we demonstrate that GraphMix can consistently improve or closely match state-of-the-art performance using even simpler architectures such as Graph Convolutional Networks, across three established graph benchmarks: Cora, Citeseer and Pubmed citation network datasets, as well as three newly proposed datasets: Cora-Full, Co-author-CS and Co-author-Physics.