Temporal Knowledge Graphs for Multi-Hop Pattern Recognition

TLogic is a symbolic AI framework for forecasting future events in temporal knowledge graphs (tKGs). It learns human-readable temporal logical rules through time-aware random walks, enabling interpretable and time-consistent predictions. While TLogic achieves state-of-the-art performance compared to existing models, our current research investigates whether Graph Neural Network (GNN)-based models can outperform large language models (LLMs) and symbolic approaches like TLogic on this task, particularly in terms of scalability, generalization, and explainability.