Bellman goes Relational

Kristian Kersting; M. van Otterlo; Luc De Raedt

doi:10.1145/1015330.1015401

Bellman goes Relational

Kristian Kersting
, M. van Otterlo
, Luc De Raedt

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

81 Citations (Scopus)

209 Downloads (Pure)

Abstract

Motivated by the interest in relational reinforcement learning, we introduce a novel relational Bellman update operator called ReBel. It employs a constraint logic programming language to compactly represent Markov decision processes over relational domains. Using ReBel, a novel value iteration algorithm is developed in which abstraction (over states and actions) plays a major role. This framework provides new insights into relational reinforcement learning. Convergence results as well as experiments are presented.

Original language	English
Title of host publication	Proceedings of the International Conference on Machine Learning (ICML'04)
Editors	R. Greiner, D. Schuurmans
Place of Publication	New York
Publisher	University of Alberta
Pages	465-472
Number of pages	8
ISBN (Print)	1-58113-8385
DOIs	https://doi.org/10.1145/1015330.1015401
Publication status	Published - 7 Dec 2004
Event	21st International Conference on Machine Learning, ICML 2004 - Banff, Canada Duration: 4 Jul 2004 → 8 Jul 2004 Conference number: 21

Conference

Conference	21st International Conference on Machine Learning, ICML 2004
Abbreviated title	ICML
Country/Territory	Canada
City	Banff
Period	4/07/04 → 8/07/04

Keywords

EC Grant Agreement nr.: FP6/508861
HMI-IA: Intelligent Agents

Access to Document

10.1145/1015330.1015401

Kersting2004bellmanAccepted author version, 221 KB

Cite this

@inproceedings{df1b9a4afbde4c408d44a67501d3b6a1,

title = "Bellman goes Relational",

abstract = "Motivated by the interest in relational reinforcement learning, we introduce a novel relational Bellman update operator called ReBel. It employs a constraint logic programming language to compactly represent Markov decision processes over relational domains. Using ReBel, a novel value iteration algorithm is developed in which abstraction (over states and actions) plays a major role. This framework provides new insights into relational reinforcement learning. Convergence results as well as experiments are presented.",

keywords = "EC Grant Agreement nr.: FP6/508861, HMI-IA: Intelligent Agents",

author = "Kristian Kersting and \{van Otterlo\}, M. and \{De Raedt\}, Luc",

year = "2004",

month = dec,

day = "7",

doi = "10.1145/1015330.1015401",

language = "English",

isbn = "1-58113-8385",

pages = "465--472",

editor = "R. Greiner and D. Schuurmans",

booktitle = "Proceedings of the International Conference on Machine Learning (ICML'04)",

publisher = "University of Alberta",

address = "Canada",

note = "21st International Conference on Machine Learning, ICML 2004, ICML ; Conference date: 04-07-2004 Through 08-07-2004",

}

TY - GEN

T1 - Bellman goes Relational

AU - Kersting, Kristian

AU - van Otterlo, M.

AU - De Raedt, Luc

N1 - Conference code: 21

PY - 2004/12/7

Y1 - 2004/12/7

N2 - Motivated by the interest in relational reinforcement learning, we introduce a novel relational Bellman update operator called ReBel. It employs a constraint logic programming language to compactly represent Markov decision processes over relational domains. Using ReBel, a novel value iteration algorithm is developed in which abstraction (over states and actions) plays a major role. This framework provides new insights into relational reinforcement learning. Convergence results as well as experiments are presented.

AB - Motivated by the interest in relational reinforcement learning, we introduce a novel relational Bellman update operator called ReBel. It employs a constraint logic programming language to compactly represent Markov decision processes over relational domains. Using ReBel, a novel value iteration algorithm is developed in which abstraction (over states and actions) plays a major role. This framework provides new insights into relational reinforcement learning. Convergence results as well as experiments are presented.

KW - EC Grant Agreement nr.: FP6/508861

KW - HMI-IA: Intelligent Agents

U2 - 10.1145/1015330.1015401

DO - 10.1145/1015330.1015401

M3 - Conference contribution

SN - 1-58113-8385

SP - 465

EP - 472

BT - Proceedings of the International Conference on Machine Learning (ICML'04)

A2 - Greiner, R.

A2 - Schuurmans, D.

PB - University of Alberta

CY - New York

T2 - 21st International Conference on Machine Learning, ICML 2004

Y2 - 4 July 2004 through 8 July 2004

ER -

Bellman goes Relational

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this