A Survey of Reinforcement Learning in Relational Domains

M. van Otterlo

Research output: Book/ReportReportProfessional

68 Downloads (Pure)

Abstract

Reinforcement learning has developed into a primary approach for learning control strategies for autonomous agents. However, most of the work has focused on the algorithmic aspect, i.e. various ways of computing value functions and policies. Usually the representational aspects were limited to the use of attribute-value or propositional languages to describe states, actions etc. A recent direction - under the general name of relational reinforcement learning - is concerned with upgrading the representation of reinforcement learning methods to the first-order case, being able to speak, reason and learn about objects and relations between objects. This survey aims at presenting an introduction to this new field, starting from the classical reinforcement learning framework. We will describe the main motivations and challenges, and give a comprehensive survey of methods that have been proposed in the literature. The aim is to give a complete survey of the available literature, of the underlying motivations and of the implications if the new methods for learning in large, relational and probabilistic environments.
Original languageUndefined
Place of PublicationEnschede
PublisherCentre for Telematics and Information Technology (CTIT)
Number of pages70
Publication statusPublished - 2005

Publication series

NameCTIT Technical Report Series
No.05-31
ISSN (Print)1381-3625

Keywords

  • METIS-227385
  • IR-53976
  • EWI-1879

Cite this

van Otterlo, M. (2005). A Survey of Reinforcement Learning in Relational Domains. (CTIT Technical Report Series; No. 05-31). Enschede: Centre for Telematics and Information Technology (CTIT).
van Otterlo, M. / A Survey of Reinforcement Learning in Relational Domains. Enschede : Centre for Telematics and Information Technology (CTIT), 2005. 70 p. (CTIT Technical Report Series; 05-31).
@book{3a17c6ca673d4607a5341674d9360221,
title = "A Survey of Reinforcement Learning in Relational Domains",
abstract = "Reinforcement learning has developed into a primary approach for learning control strategies for autonomous agents. However, most of the work has focused on the algorithmic aspect, i.e. various ways of computing value functions and policies. Usually the representational aspects were limited to the use of attribute-value or propositional languages to describe states, actions etc. A recent direction - under the general name of relational reinforcement learning - is concerned with upgrading the representation of reinforcement learning methods to the first-order case, being able to speak, reason and learn about objects and relations between objects. This survey aims at presenting an introduction to this new field, starting from the classical reinforcement learning framework. We will describe the main motivations and challenges, and give a comprehensive survey of methods that have been proposed in the literature. The aim is to give a complete survey of the available literature, of the underlying motivations and of the implications if the new methods for learning in large, relational and probabilistic environments.",
keywords = "METIS-227385, IR-53976, EWI-1879",
author = "{van Otterlo}, M.",
note = "Imported from HMI",
year = "2005",
language = "Undefined",
series = "CTIT Technical Report Series",
publisher = "Centre for Telematics and Information Technology (CTIT)",
number = "05-31",
address = "Netherlands",

}

van Otterlo, M 2005, A Survey of Reinforcement Learning in Relational Domains. CTIT Technical Report Series, no. 05-31, Centre for Telematics and Information Technology (CTIT), Enschede.

A Survey of Reinforcement Learning in Relational Domains. / van Otterlo, M.

Enschede : Centre for Telematics and Information Technology (CTIT), 2005. 70 p. (CTIT Technical Report Series; No. 05-31).

Research output: Book/ReportReportProfessional

TY - BOOK

T1 - A Survey of Reinforcement Learning in Relational Domains

AU - van Otterlo, M.

N1 - Imported from HMI

PY - 2005

Y1 - 2005

N2 - Reinforcement learning has developed into a primary approach for learning control strategies for autonomous agents. However, most of the work has focused on the algorithmic aspect, i.e. various ways of computing value functions and policies. Usually the representational aspects were limited to the use of attribute-value or propositional languages to describe states, actions etc. A recent direction - under the general name of relational reinforcement learning - is concerned with upgrading the representation of reinforcement learning methods to the first-order case, being able to speak, reason and learn about objects and relations between objects. This survey aims at presenting an introduction to this new field, starting from the classical reinforcement learning framework. We will describe the main motivations and challenges, and give a comprehensive survey of methods that have been proposed in the literature. The aim is to give a complete survey of the available literature, of the underlying motivations and of the implications if the new methods for learning in large, relational and probabilistic environments.

AB - Reinforcement learning has developed into a primary approach for learning control strategies for autonomous agents. However, most of the work has focused on the algorithmic aspect, i.e. various ways of computing value functions and policies. Usually the representational aspects were limited to the use of attribute-value or propositional languages to describe states, actions etc. A recent direction - under the general name of relational reinforcement learning - is concerned with upgrading the representation of reinforcement learning methods to the first-order case, being able to speak, reason and learn about objects and relations between objects. This survey aims at presenting an introduction to this new field, starting from the classical reinforcement learning framework. We will describe the main motivations and challenges, and give a comprehensive survey of methods that have been proposed in the literature. The aim is to give a complete survey of the available literature, of the underlying motivations and of the implications if the new methods for learning in large, relational and probabilistic environments.

KW - METIS-227385

KW - IR-53976

KW - EWI-1879

M3 - Report

T3 - CTIT Technical Report Series

BT - A Survey of Reinforcement Learning in Relational Domains

PB - Centre for Telematics and Information Technology (CTIT)

CY - Enschede

ER -

van Otterlo M. A Survey of Reinforcement Learning in Relational Domains. Enschede: Centre for Telematics and Information Technology (CTIT), 2005. 70 p. (CTIT Technical Report Series; 05-31).