Label-GCN: An Effective Method for Adding Label Propagation to Graph Convolutional Networks
This incremental improvement addresses classification accuracy in graph-based learning for both transductive and inductive settings.
The paper tackles the problem of improving Graph Convolutional Networks (GCNs) for classification by modifying the first layer to propagate label information across neighbor nodes, resulting in substantial performance gains compared to standard GCNs, especially with varying label availability and imbalanced datasets.
We show that a modification of the first layer of a Graph Convolutional Network (GCN) can be used to effectively propagate label information across neighbor nodes, for binary and multi-class classification problems. This is done by selectively eliminating self-loops for the label features during the training phase of a GCN. The GCN architecture is otherwise unchanged, without any extra hyper-parameters, and can be used in both a transductive and inductive setting. We show through several experiments that, depending on how many labels are available during the inference phase, this strategy can lead to a substantial improvement in the model performance compared to a standard GCN approach, including with imbalanced datasets.