Temporal Knowledge Graph Forecasting with Neural ODE
This addresses the challenge of modeling continuous temporal changes in knowledge graphs for improved link prediction, representing an incremental advance over discrete methods.
The paper tackled the problem of forecasting future links on temporal knowledge graphs by proposing a continuum model using neural ODEs to handle continuous-time dynamics, achieving superior performance on five benchmark datasets.
There has been an increasing interest in inferring future links on temporal knowledge graphs (KG). While links on temporal KGs vary continuously over time, the existing approaches model the temporal KGs in discrete state spaces. To this end, we propose a novel continuum model by extending the idea of neural ordinary differential equations (ODEs) to multi-relational graph convolutional networks. The proposed model preserves the continuous nature of dynamic multi-relational graph data and encodes both temporal and structural information into continuous-time dynamic embeddings. In addition, a novel graph transition layer is applied to capture the transitions on the dynamic graph, i.e., edge formation and dissolution. We perform extensive experiments on five benchmark datasets for temporal KG reasoning, showing our model's superior performance on the future link forecasting task.