Refining incomplete planning domain models through plan traces

Hankz Hankui Zhuo, Tuan Nguyen, Subbarao Kambhampati

Research output: Chapter in Book/Report/Conference proceedingConference contribution

18 Citations (Scopus)

Abstract

Most existing work on learning planning models assumes that the entire model needs to be learned from scratch. A more realistic situation is that the planning agent has an incomplete model which it needs to refine through learning. In this paper we propose and evaluate a method for doing this. Our method takes as input an incomplete model (with missing preconditions and effects in the actions), as well as a set of plan traces that are known to be correct. It outputs a "refined" model that not only captures additional precondition/effect knowledge about the given actions, but also "macro actions". We use a MAX-SAT framework for learning, where the constraints are derived from the executability of the given plan traces, as well as the preconditions/ effects of the given incomplete model. Unlike traditional macro-action learners which use macros to increase the efficiency of planning (in the context of a complete model), our motivation for learning macros is to increase the accuracy (robustness) of the plans generated with the refined model. We demonstrate the effectiveness of our approach through a systematic empirical evaluation.

Original languageEnglish (US)
Title of host publicationIJCAI International Joint Conference on Artificial Intelligence
Pages2451-2457
Number of pages7
StatePublished - 2013
Event23rd International Joint Conference on Artificial Intelligence, IJCAI 2013 - Beijing, China
Duration: Aug 3 2013Aug 9 2013

Other

Other23rd International Joint Conference on Artificial Intelligence, IJCAI 2013
CountryChina
CityBeijing
Period8/3/138/9/13

Fingerprint

Refining
Planning
Macros

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this

Zhuo, H. H., Nguyen, T., & Kambhampati, S. (2013). Refining incomplete planning domain models through plan traces. In IJCAI International Joint Conference on Artificial Intelligence (pp. 2451-2457)

Refining incomplete planning domain models through plan traces. / Zhuo, Hankz Hankui; Nguyen, Tuan; Kambhampati, Subbarao.

IJCAI International Joint Conference on Artificial Intelligence. 2013. p. 2451-2457.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Zhuo, HH, Nguyen, T & Kambhampati, S 2013, Refining incomplete planning domain models through plan traces. in IJCAI International Joint Conference on Artificial Intelligence. pp. 2451-2457, 23rd International Joint Conference on Artificial Intelligence, IJCAI 2013, Beijing, China, 8/3/13.
Zhuo HH, Nguyen T, Kambhampati S. Refining incomplete planning domain models through plan traces. In IJCAI International Joint Conference on Artificial Intelligence. 2013. p. 2451-2457
Zhuo, Hankz Hankui ; Nguyen, Tuan ; Kambhampati, Subbarao. / Refining incomplete planning domain models through plan traces. IJCAI International Joint Conference on Artificial Intelligence. 2013. pp. 2451-2457
@inproceedings{39e61ebd77404b38acf4d2912ae44f84,
title = "Refining incomplete planning domain models through plan traces",
abstract = "Most existing work on learning planning models assumes that the entire model needs to be learned from scratch. A more realistic situation is that the planning agent has an incomplete model which it needs to refine through learning. In this paper we propose and evaluate a method for doing this. Our method takes as input an incomplete model (with missing preconditions and effects in the actions), as well as a set of plan traces that are known to be correct. It outputs a {"}refined{"} model that not only captures additional precondition/effect knowledge about the given actions, but also {"}macro actions{"}. We use a MAX-SAT framework for learning, where the constraints are derived from the executability of the given plan traces, as well as the preconditions/ effects of the given incomplete model. Unlike traditional macro-action learners which use macros to increase the efficiency of planning (in the context of a complete model), our motivation for learning macros is to increase the accuracy (robustness) of the plans generated with the refined model. We demonstrate the effectiveness of our approach through a systematic empirical evaluation.",
author = "Zhuo, {Hankz Hankui} and Tuan Nguyen and Subbarao Kambhampati",
year = "2013",
language = "English (US)",
isbn = "9781577356332",
pages = "2451--2457",
booktitle = "IJCAI International Joint Conference on Artificial Intelligence",

}

TY - GEN

T1 - Refining incomplete planning domain models through plan traces

AU - Zhuo, Hankz Hankui

AU - Nguyen, Tuan

AU - Kambhampati, Subbarao

PY - 2013

Y1 - 2013

N2 - Most existing work on learning planning models assumes that the entire model needs to be learned from scratch. A more realistic situation is that the planning agent has an incomplete model which it needs to refine through learning. In this paper we propose and evaluate a method for doing this. Our method takes as input an incomplete model (with missing preconditions and effects in the actions), as well as a set of plan traces that are known to be correct. It outputs a "refined" model that not only captures additional precondition/effect knowledge about the given actions, but also "macro actions". We use a MAX-SAT framework for learning, where the constraints are derived from the executability of the given plan traces, as well as the preconditions/ effects of the given incomplete model. Unlike traditional macro-action learners which use macros to increase the efficiency of planning (in the context of a complete model), our motivation for learning macros is to increase the accuracy (robustness) of the plans generated with the refined model. We demonstrate the effectiveness of our approach through a systematic empirical evaluation.

AB - Most existing work on learning planning models assumes that the entire model needs to be learned from scratch. A more realistic situation is that the planning agent has an incomplete model which it needs to refine through learning. In this paper we propose and evaluate a method for doing this. Our method takes as input an incomplete model (with missing preconditions and effects in the actions), as well as a set of plan traces that are known to be correct. It outputs a "refined" model that not only captures additional precondition/effect knowledge about the given actions, but also "macro actions". We use a MAX-SAT framework for learning, where the constraints are derived from the executability of the given plan traces, as well as the preconditions/ effects of the given incomplete model. Unlike traditional macro-action learners which use macros to increase the efficiency of planning (in the context of a complete model), our motivation for learning macros is to increase the accuracy (robustness) of the plans generated with the refined model. We demonstrate the effectiveness of our approach through a systematic empirical evaluation.

UR - http://www.scopus.com/inward/record.url?scp=84896062946&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84896062946&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9781577356332

SP - 2451

EP - 2457

BT - IJCAI International Joint Conference on Artificial Intelligence

ER -