Graph Node-Feature Convolution for Representation Learning
This work addresses a specific bottleneck in graph representation learning for researchers and practitioners, offering an incremental improvement over existing GCN methods.
The paper tackles the problem of graph convolutional networks (GCNs) treating all features equally during neighbor aggregation, which can lead to suboptimal node representations. It introduces a new convolution operation on feature maps from sampled nodes, followed by a standard GCN, and reports improved performance in semi-supervised node classification tasks.
Graph convolutional network (GCN) is an emerging neural network approach. It learns new representation of a node by aggregating feature vectors of all neighbors in the aggregation process without considering whether the neighbors or features are useful or not. Recent methods have improved solutions by sampling a fixed size set of neighbors, or assigning different weights to different neighbors in the aggregation process, but features within a feature vector are still treated equally in the aggregation process. In this paper, we introduce a new convolution operation on regular size feature maps constructed from features of a fixed node bandwidth via sampling to get the first-level node representation, which is then passed to a standard GCN to learn the second-level node representation. Experiments show that our method outperforms competing methods in semi-supervised node classification tasks. Furthermore, our method opens new doors for exploring new GCN architectures, particularly deeper GCN models.