0581803 - ÚI 2025 RIV US eng C - Konferenční příspěvek (zahraniční konf.)
Cerna, David M. - Cropper, A.
Generalisation through Negation and Predicate Invention.
Proceedings of the 38th AAAI Conference on Artificial Intelligence. Washington, DC: AAAI Press, 2024, s. 10467-10475. ISBN 978-1-57735-887-9. ISSN 2159-5399.
[AAAI 2024: The Annual Conference on Artificial Intelligence /38./. Vancouver (CA), 20.02.2024-27.02.2024]
Grant CEP: GA ČR(CZ) GF22-06414L
Institucionální podpora: RVO:67985807
Klíčová slova: KRR * Logic Programming * ML: Statistical Relational/Logic Learning
Obor OECD: Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)
Web výsledku:
https://doi.org/10.1609/aaai.v38i9.28915DOI: https://doi.org/10.1609/aaai.v38i9.28915

The ability to generalise from a small number of examples is a fundamental challenge in machine learning. To tackle this challenge, we introduce an inductive logic programming (ILP) approach that combines negation and predicate invention. Combining these two features allows an ILP system to generalise better by learning rules with universally quantified body-only variables. We implement our idea in NOPI, which can learn normal logic programs with predicate invention, including Datalog programs with stratified negation. Our experimental results on multiple domains show that our approach can improve predictive accuracies and learning times.
Trvalý link: https://hdl.handle.net/11104/0349954
Vědecká data: Preprint - ArXiv.org