City Research Online

Efficient Predicate Invention using Shared NeMuS

Mota, E., Howe, J. M. ORCID: 0000-0001-8013-6941, Schramm, A. and d'Avila Garcez, A. S. (2019). Efficient Predicate Invention using Shared NeMuS. Paper presented at the 14th International Workshop on Neural-Symbolic Learning and Reasoning, 10 - 16 August 2019, Macau, China.

Abstract

Amao is a cognitive agent framework that tacklesthe invention of predicates with a different strat-egy as compared to recent advances in InductiveLogic Programming (ILP) approaches like Meta-Intepretive Learning (MIL) technique. It uses aNeural Multi-Space (NeMuS) graph structure toanti-unify atoms from the Herbrand base, whichpasses in the inductive momentum check. Induc-tive Clause Learning (ICL), as it is called, is ex-tended here by using the weights of logical compo-nents, already present in NeMuS, to support induc-tive learning by expanding clause candidates withanti-unified atoms. An efficient invention mecha-nism is achieved, including the learning of recur-sive hypotheses, while restricting the shape of thehypothesis by adding bias definitions or idiosyn-crasies of the language.

Publication Type: Conference or Workshop Item (Paper)
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Q Science > QA Mathematics > QA76 Computer software
Departments: School of Mathematics, Computer Science & Engineering > Computer Science
URI: http://openaccess.city.ac.uk/id/eprint/22287
[img]
Preview
Text - Accepted Version
Download (798kB) | Preview

Export

Downloads

Downloads per month over past year

View more statistics

Actions (login required)

Admin Login Admin Login