Metapath aggregation graph neural network for fusion coding and counterattack

The heterogeneous information network (HIN) has broad application prospects in practical problems due to its inclusion of different types of nodes and edges. The objective of representation learning models for HIN is to find an effective modeling method that represents the nodes in the heterogeneous information network as low-dimensional vectors while preserving the heterogeneous information in the network as much as possible. Existing representation learning models still have limitations in insufficient utilization of heterogeneous information. We propose a fusion encoding and adversarial attack meta-path aggregation graph neural network (FAMAGNN). The model consists of three module components, namely, node content transformation, intra-meta-path aggregation, and inter-meta-path aggregation, which aim to solve the problem of insufficient feature extraction in existing heterogeneous information network representation learning methods. At the same time, the model introduces a fused meta-path instance encoder to extract rich structural and semantic information in the heterogeneous information network. In addition, we introduce FGM adversarial training to perform adversarial attacks during model training to improve the robustness of the model. The outstanding performance in downstream tasks such as node classification and node clustering proves the effectiveness of this method. © 2024 Systems Engineering Society of China. All rights reserved.