Exploring Adaptive Structure Learning for Heterophilic Graphs
This work addresses a domain-specific problem for improving graph representation learning on heterophilic graphs, but it is incremental as it builds on existing GCN methods with limited success.
The paper tackled the problem of capturing long-range dependencies in heterophilic graphs using Graph Convolutional Networks (GCNs) by proposing structure learning to rewire edges in shallow GCNs, but it performed inconsistently on node classification tasks and lacked generalizability across different graph structures.
Graph Convolutional Networks (GCNs) gained traction for graph representation learning, with recent attention on improving performance on heterophilic graphs for various real-world applications. The localized feature aggregation in a typical message-passing paradigm hinders the capturing of long-range dependencies between non-local nodes of the same class. The inherent connectivity structure in heterophilic graphs often conflicts with information sharing between distant nodes of same class. We propose structure learning to rewire edges in shallow GCNs itself to avoid performance degradation in downstream discriminative tasks due to oversmoothing. Parameterizing the adjacency matrix to learn connections between non-local nodes and extend the hop span of shallow GCNs facilitates the capturing of long-range dependencies. However, our method is not generalizable across heterophilic graphs and performs inconsistently on node classification task contingent to the graph structure.