knowledge-graph-embedding

This paper proposes PogRE, a method to address pattern over-generalization in knowledge graph embedding by using dense linear transformations, improving link prediction performance on standard benchmarks.

FlowNeg is a GFlowNet-based method for diverse hard negative sampling in knowledge graph embedding, improving performance by generating context-conditioned negatives that balance hardness and diversity without treating structural similarity as absolute truth.

TeRoR introduces a novel temporal knowledge graph embedding method that decouples entity temporal evolution with independent rotation transformations and uses relational circular regions to model diverse relation mapping properties, achieving competitive performance on four datasets.