Violin: Virtual Overbridge Linking for Enhancing Semi-supervised Learning on Graphs with Limited Labels

Siyue Xie, Da Sun Handason Tam, Wing Cheong Lau

Abstract: Graph Neural Networks (GNNs) is a family of promising tools for graph semi-supervised learn- ing. However, in training, most existing GNNs rely heavily on a large amount of labeled data, which is rare in real-world scenarios. Unlabeled data with useful information are usually under- exploited, which limits the representation power of GNNs. To handle these problems, we propose Virtual Overbridge Linking (Violin), a generic framework to enhance the learning capacity of common GNNs. By learning to add virtual over- bridges between two nodes that are estimated to be semantic-consistent, labeled and unlabeled data can be correlated. Supervised information can be well utilized in training while simultaneously inducing the model to learn from unlabeled data. Discrim- inative relation patterns extracted from unlabeled nodes can also be shared with other nodes even if they are remote from each other. Motivated by re- cent advances in data augmentations, we addition- ally integrate Violin with the consistency regular- ized training. Such a scheme yields node represen- tations with better robustness, which significantly enhances a GNN. Violin can be readily extended to a wide range of GNNs without introducing addi- tional learnable parameters. Extensive experiments on six datasets demonstrate that our method is ef- fective and robust under low-label rate scenarios, where Violin can boost some GNNs’ performance by over 10% on node classifications.