Addressing Over-Smoothing in Graph Neural Networks via Deep Supervision
This addresses a key bottleneck for researchers and practitioners using deep GNNs in domains like social networks or bioinformatics, though it is an incremental improvement over existing methods.
The paper tackles the problem of over-smoothing in deep graph neural networks, where performance degrades with more layers, by proposing deeply-supervised GNNs that use representations from all layers for training, resulting in improved resilience and outperforming benchmarks on node and graph prediction tasks.
Learning useful node and graph representations with graph neural networks (GNNs) is a challenging task. It is known that deep GNNs suffer from over-smoothing where, as the number of layers increases, node representations become nearly indistinguishable and model performance on the downstream task degrades significantly. To address this problem, we propose deeply-supervised GNNs (DSGNNs), i.e., GNNs enhanced with deep supervision where representations learned at all layers are used for training. We show empirically that DSGNNs are resilient to over-smoothing and can outperform competitive benchmarks on node and graph property prediction problems.