Mitigating Semantic Confusion from Hostile Neighborhood for Graph Active Learning
This addresses a specific bottleneck in graph active learning for researchers/practitioners using GNNs, representing a novel method rather than incremental improvement.
The paper tackles the problem of semantic confusion in Graph Active Learning (GAL) caused by noisy inter-class edges during message-passing, which reduces node classification performance. The proposed Semantic-aware Active learning framework for Graphs (SAG) significantly improves performance on benchmark and real-world financial datasets, consistently outperforming previous methods.
Graph Active Learning (GAL), which aims to find the most informative nodes in graphs for annotation to maximize the Graph Neural Networks (GNNs) performance, has attracted many research efforts but remains non-trivial challenges. One major challenge is that existing GAL strategies may introduce semantic confusion to the selected training set, particularly when graphs are noisy. Specifically, most existing methods assume all aggregating features to be helpful, ignoring the semantically negative effect between inter-class edges under the message-passing mechanism. In this work, we present Semantic-aware Active learning framework for Graphs (SAG) to mitigate the semantic confusion problem. Pairwise similarities and dissimilarities of nodes with semantic features are introduced to jointly evaluate the node influence. A new prototype-based criterion and query policy are also designed to maintain diversity and class balance of the selected nodes, respectively. Extensive experiments on the public benchmark graphs and a real-world financial dataset demonstrate that SAG significantly improves node classification performances and consistently outperforms previous methods. Moreover, comprehensive analysis and ablation study also verify the effectiveness of the proposed framework.