Meta-interpretive learning of higher-order dyadic datalog: predicate invention revisited

Verfasser / Beitragende:
[Stephen Muggleton, Dianhuan Lin, Alireza Tamaddoni-Nezhad]
Ort, Verlag, Jahr:
2015
Enthalten in:
Machine Learning, 100/1(2015-07-01), 49-73
Format:
Artikel (online)
ID: 605478201
LEADER caa a22 4500
001 605478201
003 CHVBK
005 20210128100404.0
007 cr unu---uuuuu
008 210128e20150701xx s 000 0 eng
024 7 0 |a 10.1007/s10994-014-5471-y  |2 doi 
035 |a (NATIONALLICENCE)springer-10.1007/s10994-014-5471-y 
245 0 0 |a Meta-interpretive learning of higher-order dyadic datalog: predicate invention revisited  |h [Elektronische Daten]  |c [Stephen Muggleton, Dianhuan Lin, Alireza Tamaddoni-Nezhad] 
520 3 |a Since the late 1990s predicate invention has been under-explored within inductive logic programming due to difficulties in formulating efficient search mechanisms. However, a recent paper demonstrated that both predicate invention and the learning of recursion can be efficiently implemented for regular and context-free grammars, by way of metalogical substitutions with respect to a modified Prolog meta-interpreter which acts as the learning engine. New predicate symbols are introduced as constants representing existentially quantified higher-order variables. The approach demonstrates that predicate invention can be treated as a form of higher-order logical reasoning. In this paper we generalise the approach of meta-interpretive learning (MIL) to that of learning higher-order dyadic datalog programs. We show that with an infinite signature the higher-order dyadic datalog class $$H^2_2$$ H 2 2 has universal Turing expressivity though $$H^2_2$$ H 2 2 is decidable given a finite signature. Additionally we show that Knuth-Bendix ordering of the hypothesis space together with logarithmic clause bounding allows our MIL implementation Metagol $$_{D}$$ D to PAC-learn minimal cardinality $$H^2_2$$ H 2 2 definitions. This result is consistent with our experiments which indicate that Metagol $$_{D}$$ D efficiently learns compact $$H^2_2$$ H 2 2 definitions involving predicate invention for learning robotic strategies, the East-West train challenge and NELL. Additionally higher-order concepts were learned in the NELL language learning domain. The Metagol code and datasets described in this paper have been made publicly available on a website to allow reproduction of results in this paper. 
540 |a The Author(s), 2015 
690 7 |a Induction  |2 nationallicence 
690 7 |a Abduction  |2 nationallicence 
690 7 |a Meta-interpretation  |2 nationallicence 
690 7 |a Predicate invention  |2 nationallicence 
690 7 |a Learning recursion  |2 nationallicence 
700 1 |a Muggleton  |D Stephen  |u Department of Computing, Imperial College London, London, UK  |4 aut 
700 1 |a Lin  |D Dianhuan  |u Department of Computing, Imperial College London, London, UK  |4 aut 
700 1 |a Tamaddoni-Nezhad  |D Alireza  |u Department of Computing, Imperial College London, London, UK  |4 aut 
773 0 |t Machine Learning  |d Springer US; http://www.springer-ny.com  |g 100/1(2015-07-01), 49-73  |x 0885-6125  |q 100:1<49  |1 2015  |2 100  |o 10994 
856 4 0 |u https://doi.org/10.1007/s10994-014-5471-y  |q text/html  |z Onlinezugriff via DOI 
898 |a BK010053  |b XK010053  |c XK010000 
900 7 |a Metadata rights reserved  |b Springer special CC-BY-NC licence  |2 nationallicence 
908 |D 1  |a research-article  |2 jats 
949 |B NATIONALLICENCE  |F NATIONALLICENCE  |b NL-springer 
950 |B NATIONALLICENCE  |P 856  |E 40  |u https://doi.org/10.1007/s10994-014-5471-y  |q text/html  |z Onlinezugriff via DOI 
950 |B NATIONALLICENCE  |P 700  |E 1-  |a Muggleton  |D Stephen  |u Department of Computing, Imperial College London, London, UK  |4 aut 
950 |B NATIONALLICENCE  |P 700  |E 1-  |a Lin  |D Dianhuan  |u Department of Computing, Imperial College London, London, UK  |4 aut 
950 |B NATIONALLICENCE  |P 700  |E 1-  |a Tamaddoni-Nezhad  |D Alireza  |u Department of Computing, Imperial College London, London, UK  |4 aut 
950 |B NATIONALLICENCE  |P 773  |E 0-  |t Machine Learning  |d Springer US; http://www.springer-ny.com  |g 100/1(2015-07-01), 49-73  |x 0885-6125  |q 100:1<49  |1 2015  |2 100  |o 10994