Geometric-Contextual Mutual Infomax Path Aggregation for Relation Reasoning on Knowledge Graph
Geometric-Contextual Mutual Infomax Path Aggregation for Relation Reasoning on Knowledge Graph is a scholarly work, published in 2024 in ''IEEE Transactions on Knowledge and Data Engineering''. The main subjects of the publication include path, graph, artificial intelligence, graph theory, relation, Infomax, theoretical computer science, graph neural network, cognitive science, knowledge graph, mutual information, Semantic Web, and computer science. Extensive experiments on 32 real-world relation reasoning tasks demonstrate that the method significantly outperforms 8 state-of-the-art baselines in terms of AP and AUC.