aiqbench analysis
CountTRuCoLa introduces a symbolic rule-based framework for temporal knowledge graph forecasting, offering interpretable predictions traceable to specific logical rules rather than opaque neural weights. The method competes with state-of-the-art black-box models while maintaining stability on large-scale datasets where other approaches suffer from memory or runtime failures. This advancement highlights the potential of combining formal logic with machine learning to solve complex prediction tasks in dynamic knowledge graphs.
CountTRuCoLa: Rule Learning for Interpretable Temporal Knowledge Graph Forecasting
arXiv:2509.09474v2 Announce Type: replace Abstract: We address the task of temporal knowledge graph forecasting with an inherently interpretable method based on symbolic rules. Motivated by recent work proposing a strong baseline based on recurrent facts, our approach learns four simple rule types, including temporal rules with confidence functions that combine both recency and frequency. Evaluated on nine datasets, our method achieves performance that is competitive with state-of-the-art models and outperforms the majority of them, while each prediction remains directly traceable to the rules and observations that produced it. Moreover, our approach remains functional on very large datasets, where other methods encounter runtime or memory failures.
Abstract:We address the task of temporal knowledge graph forecasting with an inherently interpretable method based on symbolic rules. Motivated by recent work proposing a strong baseline based on recurrent facts, our approach learns four simple rule types, including temporal rules with confidence functions that combine both recency and frequency. Evaluated on nine datasets, our method achieves performance that is competitive with state-of-the-art models and outperforms the majority of them, while each prediction remains directly traceable to the rules and observations that produced it. Moreover, our approach remains functional on very large datasets, where other methods encounter runtime or memory failures.
Machine Learning (cs.LG)
arXiv:2509.09474 [cs.LG]
(or arXiv:2509.09474v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2509.09474 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 192 words
Related coverage
Links #5: 2026/07
LessWrong · Jul 31, 2026
General Semantic Knowledge Infusion for Spatio-Temporal Traffic Forecasting
arXiv · Aug 20, 2026
Sera: Semantic Representation Aggregation for Reliable and Interpretable Battery Health Forecasting
arXiv · Oct 10, 2026
Scene-Conditioned Relation Routing for urban cellular activity forecasting
arXiv · Sep 18, 2026