Over-smoothing is the core bottleneck of deep neural network construction. In order to alleviate the over-smoothing phenomenon in the process of graph learning, we propose a graph convolution network ...