We present an attention-based spatial graph convolution (AGC) for graph neural networks (GNNs). Existing AGCs focus on only using node-wise features and utilizing one type of attention function when calculating attention weights. Instead, we propose two methods to improve the representational power of AGCs by utilizing 1) structural information in a high-dimensional space and 2) multiple attention functions when calculating their weights. The first method computes a local structure representation of a graph in a high-dimensional space. The second method utilizes multiple attention functions simultaneously in one AGC. Both approaches can be combined. We also propose a GNN for the classification of point clouds and that for the prediction of point labels in a point cloud based on the proposed AGC. According to experiments, the proposed GNNs perform better than existing methods. Our codes open at https://github.com/liyang-tuat/SFAGC.

翻译：我们提出了一种基于注意力的空间图卷积（AGC）用于图神经网络（GNN）。现有的AGC在计算注意力权重时仅关注节点特征并使用单一类型的注意力函数。相反，我们提出了两种方法来提升AGC的表征能力：1）利用高维空间中的结构信息；2）在计算权重时使用多种注意力函数。第一种方法计算图在高维空间中的局部结构表示。第二种方法在一个AGC中同时利用多种注意力函数。两种方法可以结合使用。我们还基于所提出的AGC，分别提出了用于点云分类和点云中标签预测的GNN。实验表明，所提出的GNN性能优于现有方法。我们的代码开源在https://github.com/liyang-tuat/SFAGC。