Goal-Driven Reasoning in DatalogMTL with Magic Sets

• DOI: 10.1609/aaai.v39i14.33668
• Publication Date: 2025-04-11T12:28:15Z
• Authors: Shaoyu Wang; Kaiyue Zhao; Dongliang Wei; Przemysław Andrze...; Dingmin Wang; Hongming Cai; Pan Hu

DatalogMTL is a powerful rule-based language for temporal reasoning. Due to its high expressive power and flexible modeling capabilities, it suitable wide range of applications, including tasks from industrial financial sectors. However, due computational complexity, practical reasoning in highly challenging. To address this difficulty, we introduce new method which exploits the magic sets technique—a rewriting approach developed (non-temporal) Datalog simulate top-down evaluation with bottom-up We have implemented evaluated on publicly available benchmarks, showing that proposed significantly consistently outperformed state-of-the-art techniques.