Using Kernel Perceptrons to Learn Action Effects for Planning

Kira Mourao; Ron Petrick; Mark Steedman

Using Kernel Perceptrons to Learn Action Effects for Planning

Kira Mourao, Ron Petrick, Mark Steedman

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

We investigate the problem of learning action effects in STRIPS and ADL planning domains. Our approach is based on a kernel perceptron learning model, where action and state information is encoded in a compact vector representation as input to the learning mechanism, and resulting state changes are produced as output. Empirical results of our approach indicate efficient training and prediction times, with low average error rates (< 3%) when tested on STRIPS and ADL versions of an object manipulation scenario. This work is part of a project to integrate machine learning techniques with a planning system, as part of a larger cognitive architecture linking a high-level
reasoning component with a low-level robot/vision system.

Original language	English
Title of host publication	Proceedings of the International Conference on Cognitive Systems (CogSys 2008)
Pages	45-50
Number of pages	6
Publication status	Published - Apr 2008

Access to Document

Mourao-2008-Publication-CogsysAccepted author manuscript, 197 KB

PACO-PLUS: Perception, Action & Cognition through learning of object-action complexes
Steedman, M. (Principal Investigator)
EU government bodies
1/02/06 → 31/01/10
Project: Research

Cite this

@inproceedings{a4c452ba4d7a4f58865cafbfe5005a68,

title = "Using Kernel Perceptrons to Learn Action Effects for Planning",

abstract = "We investigate the problem of learning action effects in STRIPS and ADL planning domains. Our approach is based on a kernel perceptron learning model, where action and state information is encoded in a compact vector representation as input to the learning mechanism, and resulting state changes are produced as output. Empirical results of our approach indicate efficient training and prediction times, with low average error rates (< 3\%) when tested on STRIPS and ADL versions of an object manipulation scenario. This work is part of a project to integrate machine learning techniques with a planning system, as part of a larger cognitive architecture linking a high-level reasoning component with a low-level robot/vision system.",

author = "Kira Mourao and Ron Petrick and Mark Steedman",

year = "2008",

month = apr,

language = "English",

pages = "45--50",

booktitle = "Proceedings of the International Conference on Cognitive Systems (CogSys 2008)",

}

TY - GEN

T1 - Using Kernel Perceptrons to Learn Action Effects for Planning

AU - Mourao, Kira

AU - Petrick, Ron

AU - Steedman, Mark

PY - 2008/4

Y1 - 2008/4

N2 - We investigate the problem of learning action effects in STRIPS and ADL planning domains. Our approach is based on a kernel perceptron learning model, where action and state information is encoded in a compact vector representation as input to the learning mechanism, and resulting state changes are produced as output. Empirical results of our approach indicate efficient training and prediction times, with low average error rates (< 3%) when tested on STRIPS and ADL versions of an object manipulation scenario. This work is part of a project to integrate machine learning techniques with a planning system, as part of a larger cognitive architecture linking a high-level reasoning component with a low-level robot/vision system.

AB - We investigate the problem of learning action effects in STRIPS and ADL planning domains. Our approach is based on a kernel perceptron learning model, where action and state information is encoded in a compact vector representation as input to the learning mechanism, and resulting state changes are produced as output. Empirical results of our approach indicate efficient training and prediction times, with low average error rates (< 3%) when tested on STRIPS and ADL versions of an object manipulation scenario. This work is part of a project to integrate machine learning techniques with a planning system, as part of a larger cognitive architecture linking a high-level reasoning component with a low-level robot/vision system.

M3 - Conference contribution

SP - 45

EP - 50

BT - Proceedings of the International Conference on Cognitive Systems (CogSys 2008)

ER -

Using Kernel Perceptrons to Learn Action Effects for Planning

Abstract

Access to Document

Fingerprint

Projects

PACO-PLUS: Perception, Action & Cognition through learning of object-action complexes

Cite this