Combinatorial online prediction via metarounding

Takahiro Fujita, Kohei Hatano, Eiji Takimoto

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

6 Citations (Scopus)

Abstract

We consider online prediction problems of combinatorial concepts. Examples of such concepts include s-t paths, permutations, truth assignments, set covers, and so on. The goal of the online prediction algorithm is to compete with the best fixed combinatorial concept in hindsight. A generic approach to this problem is to design an online prediction algorithm using the corresponding offline (approximation) algorithm as an oracle. The current state-of-the art method, however, is not efficient enough. In this paper we propose a more efficient online prediction algorithm when the offline approximation algorithm has a guarantee of the integrality gap.

Original languageEnglish
Title of host publicationAlgorithmic Learning Theory - 24th International Conference, ALT 2013, Proceedings
Pages68-82
Number of pages15
DOIs
Publication statusPublished - Nov 18 2013
Event24th International Conference on Algorithmic Learning Theory, ALT 2013 - Singapore, Singapore
Duration: Oct 6 2013Oct 9 2013

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8139 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other24th International Conference on Algorithmic Learning Theory, ALT 2013
CountrySingapore
CitySingapore
Period10/6/1310/9/13

Fingerprint

Prediction
Approximation algorithms
Approximation Algorithms
Set Cover
Integrality
Permutation
Assignment
Path
Concepts
Design
Truth

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Fujita, T., Hatano, K., & Takimoto, E. (2013). Combinatorial online prediction via metarounding. In Algorithmic Learning Theory - 24th International Conference, ALT 2013, Proceedings (pp. 68-82). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8139 LNAI). https://doi.org/10.1007/978-3-642-40935-6_6

Combinatorial online prediction via metarounding. / Fujita, Takahiro; Hatano, Kohei; Takimoto, Eiji.

Algorithmic Learning Theory - 24th International Conference, ALT 2013, Proceedings. 2013. p. 68-82 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8139 LNAI).

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

Fujita, T, Hatano, K & Takimoto, E 2013, Combinatorial online prediction via metarounding. in Algorithmic Learning Theory - 24th International Conference, ALT 2013, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8139 LNAI, pp. 68-82, 24th International Conference on Algorithmic Learning Theory, ALT 2013, Singapore, Singapore, 10/6/13. https://doi.org/10.1007/978-3-642-40935-6_6
Fujita T, Hatano K, Takimoto E. Combinatorial online prediction via metarounding. In Algorithmic Learning Theory - 24th International Conference, ALT 2013, Proceedings. 2013. p. 68-82. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-40935-6_6
Fujita, Takahiro ; Hatano, Kohei ; Takimoto, Eiji. / Combinatorial online prediction via metarounding. Algorithmic Learning Theory - 24th International Conference, ALT 2013, Proceedings. 2013. pp. 68-82 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{c98bd46a2ac64a1faf7778a485fdd3b0,
title = "Combinatorial online prediction via metarounding",
abstract = "We consider online prediction problems of combinatorial concepts. Examples of such concepts include s-t paths, permutations, truth assignments, set covers, and so on. The goal of the online prediction algorithm is to compete with the best fixed combinatorial concept in hindsight. A generic approach to this problem is to design an online prediction algorithm using the corresponding offline (approximation) algorithm as an oracle. The current state-of-the art method, however, is not efficient enough. In this paper we propose a more efficient online prediction algorithm when the offline approximation algorithm has a guarantee of the integrality gap.",
author = "Takahiro Fujita and Kohei Hatano and Eiji Takimoto",
year = "2013",
month = "11",
day = "18",
doi = "10.1007/978-3-642-40935-6_6",
language = "English",
isbn = "9783642409349",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "68--82",
booktitle = "Algorithmic Learning Theory - 24th International Conference, ALT 2013, Proceedings",

}

TY - GEN

T1 - Combinatorial online prediction via metarounding

AU - Fujita, Takahiro

AU - Hatano, Kohei

AU - Takimoto, Eiji

PY - 2013/11/18

Y1 - 2013/11/18

N2 - We consider online prediction problems of combinatorial concepts. Examples of such concepts include s-t paths, permutations, truth assignments, set covers, and so on. The goal of the online prediction algorithm is to compete with the best fixed combinatorial concept in hindsight. A generic approach to this problem is to design an online prediction algorithm using the corresponding offline (approximation) algorithm as an oracle. The current state-of-the art method, however, is not efficient enough. In this paper we propose a more efficient online prediction algorithm when the offline approximation algorithm has a guarantee of the integrality gap.

AB - We consider online prediction problems of combinatorial concepts. Examples of such concepts include s-t paths, permutations, truth assignments, set covers, and so on. The goal of the online prediction algorithm is to compete with the best fixed combinatorial concept in hindsight. A generic approach to this problem is to design an online prediction algorithm using the corresponding offline (approximation) algorithm as an oracle. The current state-of-the art method, however, is not efficient enough. In this paper we propose a more efficient online prediction algorithm when the offline approximation algorithm has a guarantee of the integrality gap.

UR - http://www.scopus.com/inward/record.url?scp=84887443709&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84887443709&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-40935-6_6

DO - 10.1007/978-3-642-40935-6_6

M3 - Conference contribution

AN - SCOPUS:84887443709

SN - 9783642409349

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 68

EP - 82

BT - Algorithmic Learning Theory - 24th International Conference, ALT 2013, Proceedings

ER -