Generalization in order-sorted theories with any combination of associativity (A), commutativity (C), and unity (U) algebraic axioms is finitary. However, existing tools for computing generalizers (also called "anti-unifiers") of two typed structures in such theories do not currently scale to real size problems. This paper describes the ACUOS2 system that achieves high performance when computing a complete and minimal set of least general generalizations in these theories. We discuss how it can be used to address artificial intelligence (AI) problems that are representable as order-sorted ACU generalization, e.g., generalization in lists, trees, (multi-)sets, and typical hierarchical/structural relations. Experimental results are also given to demonstrate that ACUOS2 greatly outperforms the predecessor tool ACUOS by running up to five orders of magnitude faster.

$$ extsf ACUOS^mathbf 2$$ : A High-Performance System for Modular ACU Generalization with Subtyping and Inheritance

Ballis D.;
2019

Abstract

Generalization in order-sorted theories with any combination of associativity (A), commutativity (C), and unity (U) algebraic axioms is finitary. However, existing tools for computing generalizers (also called "anti-unifiers") of two typed structures in such theories do not currently scale to real size problems. This paper describes the ACUOS2 system that achieves high performance when computing a complete and minimal set of least general generalizations in these theories. We discuss how it can be used to address artificial intelligence (AI) problems that are representable as order-sorted ACU generalization, e.g., generalization in lists, trees, (multi-)sets, and typical hierarchical/structural relations. Experimental results are also given to demonstrate that ACUOS2 greatly outperforms the predecessor tool ACUOS by running up to five orders of magnitude faster.
978-3-030-19569-4
978-3-030-19570-0
File in questo prodotto:
File Dimensione Formato  
JELIA!9-Editorial.pdf

non disponibili

Tipologia: Versione Editoriale (PDF)
Licenza: Non pubblico
Dimensione 249.38 kB
Formato Adobe PDF
249.38 kB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11390/1150835
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 7
  • ???jsp.display-item.citation.isi??? 6
social impact