Simple yet Effective Gradient-Free Graph Convolutional Networks
This work addresses over-smoothing issues in graph representation learning for researchers and practitioners, offering an incremental improvement over existing linearized GNN methods.
The paper tackles over-smoothing in linearized Graph Neural Networks by linking it to vanishing gradients and proposes a gradient-free training framework, resulting in better and more stable node classification performance with significantly reduced training time.
Linearized Graph Neural Networks (GNNs) have attracted great attention in recent years for graph representation learning. Compared with nonlinear Graph Neural Network (GNN) models, linearized GNNs are much more time-efficient and can achieve comparable performances on typical downstream tasks such as node classification. Although some linearized GNN variants are purposely crafted to mitigate ``over-smoothing", empirical studies demonstrate that they still somehow suffer from this issue. In this paper, we instead relate over-smoothing with the vanishing gradient phenomenon and craft a gradient-free training framework to achieve more efficient and effective linearized GNNs which can significantly overcome over-smoothing and enhance the generalization of the model. The experimental results demonstrate that our methods achieve better and more stable performances on node classification tasks with varying depths and cost much less training time.