MM-GCN: Multi-Mutual Learning Networks of Graph Convolution for Node Classification

Graph Convolutional Network (GCN) has been proved to be an effective method to process graph-structured data. Concurrently, deep mutual learning has shown significant improvements in network performance. In this paper, we propose a model: Multi-Mutual learning Networks of GCN (MM-GCN), which combines these two lines of work. At its core, MM-GCN trains multiple GCNs, which may be the same or different, and the final loss function is jointly determined by these networks, which can be used for backpropagation to train the network. Our experiments show that the effect of MM-GCN proposed by us improves state-of-the-art baselines on node classification tasks.