ALG: Fast and Accurate Active Learning Framework for Graph Convolutional Networks

Graph Convolutional Networks (GCNs) and their variants, which have become the de facto methods in many supervised and semi-supervised graph representation learning scenarios suchas node classification and link prediction.In order to achieve satisfactory performance, GCNs require a sufficient amount oflabeled data. However, in the real world scenarios, labeled data is often expensive to obtain.Therefore, we propose ALG, a novel \textbf{A}ctive \textbf{L}earning framework for \textbf{G}CNs, which employs domain-specific intelligence to achieve much higher performance and efficiency compared to the generic AL frameworks.First, by decoupling GCN models, ALG serves as an effective and efficient AL framework for measuring and combining node representativeness and informativeness.Second, by exploiting the characteristic of receptive field in GCNs, ALG considers both the importance and correlation of nodes through proposing a new node selection metric that maximizes the effective receptive field (ERF). We prove that this ERF maximization problem is an NP-hard and provide an efficient algorithm accompanied with provable approximationguarantee.The empirical studies on four public datasets demonstrate that ERF can significantly improve both the performance and efficiency of active learning for GCNs.Especially on Reddit dataset, the proposed ALG can outperform the competitive baseline - AGE by at least two orders of magnitude in our evaluation.