TEILP: Time Prediction over Knowledge Graphs via Logical Reasoning

• DOI: 10.48550/arxiv.2312.15816
• Publication Date: 2023-01-01
• Authors: Siheng Xiong; Yuan Yang; Ali Payani; James C Kerce; Faramarz Fekri
• Keywords: Robustness; Knowledge graph

Conventional embedding-based models approach event time prediction in temporal knowledge graphs (TKGs) as a ranking problem. However, they often fall short capturing essential relationships such order and distance. In this paper, we propose TEILP, logical reasoning framework that naturally integrates elements into graph predictions. We first convert TKGs (TEKG) which has more explicit representation of term nodes the graph. The TEKG equips us to develop differentiable random walk prediction. Finally, introduce conditional probability density functions, associated with rules involving query interval, using arrive at compare TEILP state-of-the-art methods on five benchmark datasets. show our model achieves significant improvement over baselines while providing interpretable explanations. particular, consider several scenarios where training samples are limited, types imbalanced, forecasting future events based only past is desired. all these cases, outperforms terms robustness.