Learning on Temporal Knowledge Graphs

Abstract

Temporal Knowledge Graphs (TKG) extend Knowledge Graphs by incorporating temporal information, enabling the representation of facts that occur, recur, or evolve over time. This talk provides an overview of TKG, with a particular focus on TKG forecasting—the task of predicting future links based on historical graph snapshots. After reviewing existing forecasting methods, I will discuss key challenges in their evaluation, including inconsistent protocols and the use of limited, small-scale datasets. Finally, I will present steps toward more rigorous evaluation: a joint evaluation protocol, a standardized framework, large-scale benchmark datasets, and a simple baseline for comparison.

Bio

Julia Gastinger (she/her) is a Ph.D. student at Mannheim University, supervised by Professor Heiner Stuckenschmidt. Previously, she was a Research Scientist in the AI Innovations group at NEC Laboratories Europe. Her research primarily focuses on graph-based Machine Learning – she is interested in how to incorporate the time aspect in knowledge graph representations. She served as a Reviewing Chair and Co-Organizer in Temporal Graph Learning Workshop @ NeurIPS 2023 and Temporal Graph Learning Workshop @ KDD 2025, and co-organizes the weekly online Temporal Graph Learning Reading Group.
Google Scholar
LinkedIn
Personal website