Compressing pattern databases with learning

Mehdi Samadi; Maryam Siabani; Ariel Felner; Robert Holte

doi:10.3233/978-1-58603-891-5-495

Compressing pattern databases with learning

Mehdi Samadi, Maryam Siabani, Ariel Felner, Robert Holte

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

11 Scopus citations

Abstract

A pattern database (PDB) is a heuristic function implemented as a lookup table. It stores the lengths of optimal solutions for instances of subproblems. Most previous PDBs had a distinct entry in the table for each subproblem instance. In this paper we apply learning techniques to compress PDBs by using neural networks and decision trees thereby reducing the amount of memory needed. Experiments on the sliding tile puzzles and the TopSpin puzzle show that our compressed PDBs significantly outperforms both uncompressed PDBs as well as previous compressing methods. Our full compressing system reduced the size of memory needed by a factor of up to 63 at a cost of no more than a factor of 2 in the search effort.

Original language	English
Title of host publication	Frontiers in Artificial Intelligence and Applications
Publisher	IOS Press BV
Pages	495-499
Number of pages	5
ISBN (Print)	978158603891
DOIs	https://doi.org/10.3233/978-1-58603-891-5-495
State	Published - Jun 2008
Externally published	Yes
Event	18th European Conference on Artificial Intelligence, ECAI 2008 - Patras, Greece Duration: 21 Jul 2008 → 25 Jul 2008

Publication series

Name	Frontiers in Artificial Intelligence and Applications
Volume	178
ISSN (Print)	0922-6389
ISSN (Electronic)	1879-8314

Conference

Conference	18th European Conference on Artificial Intelligence, ECAI 2008
Country/Territory	Greece
City	Patras
Period	21/07/08 → 25/07/08

Bibliographical note

Publisher Copyright:
© 2008 The authors and IOS Press. All rights reserved.

Access to Document

10.3233/978-1-58603-891-5-495

Cite this

@inproceedings{f63219fe152249b687a4caec820f0d7f,

title = "Compressing pattern databases with learning",

abstract = "A pattern database (PDB) is a heuristic function implemented as a lookup table. It stores the lengths of optimal solutions for instances of subproblems. Most previous PDBs had a distinct entry in the table for each subproblem instance. In this paper we apply learning techniques to compress PDBs by using neural networks and decision trees thereby reducing the amount of memory needed. Experiments on the sliding tile puzzles and the TopSpin puzzle show that our compressed PDBs significantly outperforms both uncompressed PDBs as well as previous compressing methods. Our full compressing system reduced the size of memory needed by a factor of up to 63 at a cost of no more than a factor of 2 in the search effort.",

author = "Mehdi Samadi and Maryam Siabani and Ariel Felner and Robert Holte",

note = "Publisher Copyright: {\textcopyright} 2008 The authors and IOS Press. All rights reserved.; 18th European Conference on Artificial Intelligence, ECAI 2008 ; Conference date: 21-07-2008 Through 25-07-2008",

year = "2008",

month = jun,

doi = "10.3233/978-1-58603-891-5-495",

language = "אנגלית",

isbn = "978158603891",

series = "Frontiers in Artificial Intelligence and Applications",

publisher = "IOS Press BV",

pages = "495--499",

booktitle = "Frontiers in Artificial Intelligence and Applications",

address = "הולנד",

}

Samadi, M, Siabani, M, Felner, A & Holte, R 2008, Compressing pattern databases with learning. in Frontiers in Artificial Intelligence and Applications. Frontiers in Artificial Intelligence and Applications, vol. 178, IOS Press BV, pp. 495-499, 18th European Conference on Artificial Intelligence, ECAI 2008, Patras, Greece, 21/07/08. https://doi.org/10.3233/978-1-58603-891-5-495

TY - GEN

T1 - Compressing pattern databases with learning

AU - Samadi, Mehdi

AU - Siabani, Maryam

AU - Felner, Ariel

AU - Holte, Robert

PY - 2008/6

Y1 - 2008/6

N2 - A pattern database (PDB) is a heuristic function implemented as a lookup table. It stores the lengths of optimal solutions for instances of subproblems. Most previous PDBs had a distinct entry in the table for each subproblem instance. In this paper we apply learning techniques to compress PDBs by using neural networks and decision trees thereby reducing the amount of memory needed. Experiments on the sliding tile puzzles and the TopSpin puzzle show that our compressed PDBs significantly outperforms both uncompressed PDBs as well as previous compressing methods. Our full compressing system reduced the size of memory needed by a factor of up to 63 at a cost of no more than a factor of 2 in the search effort.

AB - A pattern database (PDB) is a heuristic function implemented as a lookup table. It stores the lengths of optimal solutions for instances of subproblems. Most previous PDBs had a distinct entry in the table for each subproblem instance. In this paper we apply learning techniques to compress PDBs by using neural networks and decision trees thereby reducing the amount of memory needed. Experiments on the sliding tile puzzles and the TopSpin puzzle show that our compressed PDBs significantly outperforms both uncompressed PDBs as well as previous compressing methods. Our full compressing system reduced the size of memory needed by a factor of up to 63 at a cost of no more than a factor of 2 in the search effort.

UR - http://www.scopus.com/inward/record.url?scp=79957870304&partnerID=8YFLogxK

U2 - 10.3233/978-1-58603-891-5-495

DO - 10.3233/978-1-58603-891-5-495

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:79957870304

SN - 978158603891

T3 - Frontiers in Artificial Intelligence and Applications

SP - 495

EP - 499

BT - Frontiers in Artificial Intelligence and Applications

PB - IOS Press BV

T2 - 18th European Conference on Artificial Intelligence, ECAI 2008

Y2 - 21 July 2008 through 25 July 2008

ER -

Compressing pattern databases with learning

Abstract

Publication series

Conference

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this