Learning local forward models on unforgiving games

Publikation: Beitrag in Buch/Bericht/Sammelwerk/KonferenzbandAufsatz in KonferenzbandForschungPeer-Review

Autoren

Externe Organisationen

  • Otto-von-Guericke-Universität Magdeburg
  • Queen Mary University of London
Forschungs-netzwerk anzeigen

Details

OriginalspracheEnglisch
Titel des SammelwerksIEEE Conference on Games 2019, CoG 2019
Herausgeber (Verlag)IEEE Computer Society
ISBN (elektronisch)9781728118840
PublikationsstatusVeröffentlicht - Aug. 2019
Extern publiziertJa
Veranstaltung2019 IEEE Conference on Games, CoG 2019 - London, Großbritannien / Vereinigtes Königreich
Dauer: 20 Aug. 201923 Aug. 2019

Publikationsreihe

NameIEEE Conference on Computatonal Intelligence and Games, CIG
Band2019-August
ISSN (Print)2325-4270
ISSN (elektronisch)2325-4289

Abstract

This paper examines learning approaches for forward models based on local cell transition functions. We provide a formal definition of local forward models for which we propose two basic learning approaches. Our analysis is based on the game Sokoban, where a wrong action can lead to an unsolvable game state. Therefore, an accurate prediction of an action's resulting state is necessary to avoid this scenario.In contrast to learning the complete state transition function, local forward models allow extracting multiple training examples from a single state transition. In this way, the Hash Set model, as well as the Decision Tree model, quickly learn to predict upcoming state transitions of both the training and the test set. Applying the model using a statistical forward planner showed that the best models can be used to satisfying degree even in cases in which the test levels have not yet been seen.Our evaluation includes an analysis of various local neighbourhood patterns and sizes to test the learners' capabilities in case too few or too many attributes are extracted, of which the latter has shown do degrade the performance of the model learner.

ASJC Scopus Sachgebiete

Zitieren

Learning local forward models on unforgiving games. / Dockhorn, Alexander; Lucas, Simon M.; Volz, Vanessa et al.
IEEE Conference on Games 2019, CoG 2019. IEEE Computer Society, 2019. 8848044 (IEEE Conference on Computatonal Intelligence and Games, CIG; Band 2019-August).

Publikation: Beitrag in Buch/Bericht/Sammelwerk/KonferenzbandAufsatz in KonferenzbandForschungPeer-Review

Dockhorn, A, Lucas, SM, Volz, V, Bravi, I, Gaina, RD & Perez-Liebana, D 2019, Learning local forward models on unforgiving games. in IEEE Conference on Games 2019, CoG 2019., 8848044, IEEE Conference on Computatonal Intelligence and Games, CIG, Bd. 2019-August, IEEE Computer Society, 2019 IEEE Conference on Games, CoG 2019, London, Großbritannien / Vereinigtes Königreich, 20 Aug. 2019. https://doi.org/10.1109/CIG.2019.8848044
Dockhorn, A., Lucas, S. M., Volz, V., Bravi, I., Gaina, R. D., & Perez-Liebana, D. (2019). Learning local forward models on unforgiving games. In IEEE Conference on Games 2019, CoG 2019 Artikel 8848044 (IEEE Conference on Computatonal Intelligence and Games, CIG; Band 2019-August). IEEE Computer Society. https://doi.org/10.1109/CIG.2019.8848044
Dockhorn A, Lucas SM, Volz V, Bravi I, Gaina RD, Perez-Liebana D. Learning local forward models on unforgiving games. in IEEE Conference on Games 2019, CoG 2019. IEEE Computer Society. 2019. 8848044. (IEEE Conference on Computatonal Intelligence and Games, CIG). doi: 10.1109/CIG.2019.8848044
Dockhorn, Alexander ; Lucas, Simon M. ; Volz, Vanessa et al. / Learning local forward models on unforgiving games. IEEE Conference on Games 2019, CoG 2019. IEEE Computer Society, 2019. (IEEE Conference on Computatonal Intelligence and Games, CIG).
Download
@inproceedings{57fc8c60826747f8829b9df85bbf7da0,
title = "Learning local forward models on unforgiving games",
abstract = "This paper examines learning approaches for forward models based on local cell transition functions. We provide a formal definition of local forward models for which we propose two basic learning approaches. Our analysis is based on the game Sokoban, where a wrong action can lead to an unsolvable game state. Therefore, an accurate prediction of an action's resulting state is necessary to avoid this scenario.In contrast to learning the complete state transition function, local forward models allow extracting multiple training examples from a single state transition. In this way, the Hash Set model, as well as the Decision Tree model, quickly learn to predict upcoming state transitions of both the training and the test set. Applying the model using a statistical forward planner showed that the best models can be used to satisfying degree even in cases in which the test levels have not yet been seen.Our evaluation includes an analysis of various local neighbourhood patterns and sizes to test the learners' capabilities in case too few or too many attributes are extracted, of which the latter has shown do degrade the performance of the model learner.",
keywords = "Decision Tree, Forward Model Learning, Local Forward Model, Rolling Horizon Evolutionary Algorithm",
author = "Alexander Dockhorn and Lucas, {Simon M.} and Vanessa Volz and Ivan Bravi and Gaina, {Raluca D.} and Diego Perez-Liebana",
year = "2019",
month = aug,
doi = "10.1109/CIG.2019.8848044",
language = "English",
series = "IEEE Conference on Computatonal Intelligence and Games, CIG",
publisher = "IEEE Computer Society",
booktitle = "IEEE Conference on Games 2019, CoG 2019",
address = "United States",
note = "2019 IEEE Conference on Games, CoG 2019 ; Conference date: 20-08-2019 Through 23-08-2019",

}

Download

TY - GEN

T1 - Learning local forward models on unforgiving games

AU - Dockhorn, Alexander

AU - Lucas, Simon M.

AU - Volz, Vanessa

AU - Bravi, Ivan

AU - Gaina, Raluca D.

AU - Perez-Liebana, Diego

PY - 2019/8

Y1 - 2019/8

N2 - This paper examines learning approaches for forward models based on local cell transition functions. We provide a formal definition of local forward models for which we propose two basic learning approaches. Our analysis is based on the game Sokoban, where a wrong action can lead to an unsolvable game state. Therefore, an accurate prediction of an action's resulting state is necessary to avoid this scenario.In contrast to learning the complete state transition function, local forward models allow extracting multiple training examples from a single state transition. In this way, the Hash Set model, as well as the Decision Tree model, quickly learn to predict upcoming state transitions of both the training and the test set. Applying the model using a statistical forward planner showed that the best models can be used to satisfying degree even in cases in which the test levels have not yet been seen.Our evaluation includes an analysis of various local neighbourhood patterns and sizes to test the learners' capabilities in case too few or too many attributes are extracted, of which the latter has shown do degrade the performance of the model learner.

AB - This paper examines learning approaches for forward models based on local cell transition functions. We provide a formal definition of local forward models for which we propose two basic learning approaches. Our analysis is based on the game Sokoban, where a wrong action can lead to an unsolvable game state. Therefore, an accurate prediction of an action's resulting state is necessary to avoid this scenario.In contrast to learning the complete state transition function, local forward models allow extracting multiple training examples from a single state transition. In this way, the Hash Set model, as well as the Decision Tree model, quickly learn to predict upcoming state transitions of both the training and the test set. Applying the model using a statistical forward planner showed that the best models can be used to satisfying degree even in cases in which the test levels have not yet been seen.Our evaluation includes an analysis of various local neighbourhood patterns and sizes to test the learners' capabilities in case too few or too many attributes are extracted, of which the latter has shown do degrade the performance of the model learner.

KW - Decision Tree

KW - Forward Model Learning

KW - Local Forward Model

KW - Rolling Horizon Evolutionary Algorithm

UR - http://www.scopus.com/inward/record.url?scp=85073106140&partnerID=8YFLogxK

U2 - 10.1109/CIG.2019.8848044

DO - 10.1109/CIG.2019.8848044

M3 - Conference contribution

AN - SCOPUS:85073106140

T3 - IEEE Conference on Computatonal Intelligence and Games, CIG

BT - IEEE Conference on Games 2019, CoG 2019

PB - IEEE Computer Society

T2 - 2019 IEEE Conference on Games, CoG 2019

Y2 - 20 August 2019 through 23 August 2019

ER -

Von denselben Autoren