Graph Convolutional Networks for Classification with a Structured Label Space
This work addresses classification tasks where labels have inherent relationships, offering a method to leverage such structures for better consistency, though it appears incremental in its approach.
The paper tackles the problem of multi-class classification by exploiting known graph structures among labels using a graph convolutional network (GCN) augmented neural network, resulting in improved graph-theoretic metrics compared to a baseline that ignores label structures.
It is a usual practice to ignore any structural information underlying classes in multi-class classification. In this paper, we propose a graph convolutional network (GCN) augmented neural network classifier to exploit a known, underlying graph structure of labels. The proposed approach resembles an (approximate) inference procedure in, for instance, a conditional random field (CRF). We evaluate the proposed approach on document classification and object recognition and report both accuracies and graph-theoretic metrics that correspond to the consistency of the model's prediction. The experiment results reveal that the proposed model outperforms a baseline method which ignores the graph structures of a label space in terms of graph-theoretic metrics.