Directed Graph Auto-Encoders
This work addresses the challenge of modeling directed graphs for researchers in graph representation learning, though it appears incremental as it extends existing auto-encoder and GCN methods to directed cases.
The authors tackled the problem of learning latent representations for directed graphs by introducing a new class of auto-encoders based on an extension of the Weisfeiler-Leman algorithm, achieving superior performance on directed link prediction tasks across several network datasets.
We introduce a new class of auto-encoders for directed graphs, motivated by a direct extension of the Weisfeiler-Leman algorithm to pairs of node labels. The proposed model learns pairs of interpretable latent representations for the nodes of directed graphs, and uses parameterized graph convolutional network (GCN) layers for its encoder and an asymmetric inner product decoder. Parameters in the encoder control the weighting of representations exchanged between neighboring nodes. We demonstrate the ability of the proposed model to learn meaningful latent embeddings and achieve superior performance on the directed link prediction task on several popular network datasets.