ShareBoost

Boosting for multi-view learning with performance guarantees

Jing Peng, Costin Barbu, Guna Seetharaman, Wei Fan, Xian Wu, Kannappan Palaniappan

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

18 Citations (Scopus)

Abstract

Algorithms combining multi-view information are known to exponentially quicken classification, and have been applied to many fields. However, they lack the ability to mine most discriminant information sources (or data types) for making predictions. In this paper, we propose an algorithm based on boosting to address these problems. The proposed algorithm builds base classifiers independently from each data type (view) that provides a partial view about an object of interest. Different from AdaBoost, where each view has its own re-sampling weight, our algorithm uses a single re-sampling distribution for all views at each boosting round. This distribution is determined by the view whose training error is minimal. This shared sampling mechanism restricts noise to individual views, thereby reducing sensitivity to noise. Furthermore, in order to establish performance guarantees, we introduce a randomized version of the algorithm, where a winning view is chosen probabilistically. As a result, it can be cast within a multi-armed bandit framework, which allows us to show that with high probability the algorithm seeks out most discriminant views of data for making predictions. We provide experimental results that show its performance against noise and competing techniques.

Original languageEnglish
Title of host publicationMachine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2011, Proceedings
Pages597-612
Number of pages16
EditionPART 2
DOIs
StatePublished - 9 Sep 2011
EventEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2011 - Athens, Greece
Duration: 5 Sep 20119 Sep 2011

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume6912 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

OtherEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2011
CountryGreece
CityAthens
Period5/09/119/09/11

Fingerprint

Performance Guarantee
Boosting
Resampling
Sampling
Discriminant
Multi-armed Bandit
Adaptive boosting
Sampling Distribution
AdaBoost
Prediction
Learning
Classifiers
Classifier
Partial
Experimental Results

Keywords

  • boosting
  • convergence
  • Data fusion
  • multi-view learning

Cite this

Peng, J., Barbu, C., Seetharaman, G., Fan, W., Wu, X., & Palaniappan, K. (2011). ShareBoost: Boosting for multi-view learning with performance guarantees. In Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2011, Proceedings (PART 2 ed., pp. 597-612). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6912 LNAI, No. PART 2). https://doi.org/10.1007/978-3-642-23783-6_38
Peng, Jing ; Barbu, Costin ; Seetharaman, Guna ; Fan, Wei ; Wu, Xian ; Palaniappan, Kannappan. / ShareBoost : Boosting for multi-view learning with performance guarantees. Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2011, Proceedings. PART 2. ed. 2011. pp. 597-612 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).
@inproceedings{aaa9cb4dfc884a7eb3a57fea50e1c727,
title = "ShareBoost: Boosting for multi-view learning with performance guarantees",
abstract = "Algorithms combining multi-view information are known to exponentially quicken classification, and have been applied to many fields. However, they lack the ability to mine most discriminant information sources (or data types) for making predictions. In this paper, we propose an algorithm based on boosting to address these problems. The proposed algorithm builds base classifiers independently from each data type (view) that provides a partial view about an object of interest. Different from AdaBoost, where each view has its own re-sampling weight, our algorithm uses a single re-sampling distribution for all views at each boosting round. This distribution is determined by the view whose training error is minimal. This shared sampling mechanism restricts noise to individual views, thereby reducing sensitivity to noise. Furthermore, in order to establish performance guarantees, we introduce a randomized version of the algorithm, where a winning view is chosen probabilistically. As a result, it can be cast within a multi-armed bandit framework, which allows us to show that with high probability the algorithm seeks out most discriminant views of data for making predictions. We provide experimental results that show its performance against noise and competing techniques.",
keywords = "boosting, convergence, Data fusion, multi-view learning",
author = "Jing Peng and Costin Barbu and Guna Seetharaman and Wei Fan and Xian Wu and Kannappan Palaniappan",
year = "2011",
month = "9",
day = "9",
doi = "10.1007/978-3-642-23783-6_38",
language = "English",
isbn = "9783642237829",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
number = "PART 2",
pages = "597--612",
booktitle = "Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2011, Proceedings",
edition = "PART 2",

}

Peng, J, Barbu, C, Seetharaman, G, Fan, W, Wu, X & Palaniappan, K 2011, ShareBoost: Boosting for multi-view learning with performance guarantees. in Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2011, Proceedings. PART 2 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 6912 LNAI, pp. 597-612, European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2011, Athens, Greece, 5/09/11. https://doi.org/10.1007/978-3-642-23783-6_38

ShareBoost : Boosting for multi-view learning with performance guarantees. / Peng, Jing; Barbu, Costin; Seetharaman, Guna; Fan, Wei; Wu, Xian; Palaniappan, Kannappan.

Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2011, Proceedings. PART 2. ed. 2011. p. 597-612 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6912 LNAI, No. PART 2).

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

TY - GEN

T1 - ShareBoost

T2 - Boosting for multi-view learning with performance guarantees

AU - Peng, Jing

AU - Barbu, Costin

AU - Seetharaman, Guna

AU - Fan, Wei

AU - Wu, Xian

AU - Palaniappan, Kannappan

PY - 2011/9/9

Y1 - 2011/9/9

N2 - Algorithms combining multi-view information are known to exponentially quicken classification, and have been applied to many fields. However, they lack the ability to mine most discriminant information sources (or data types) for making predictions. In this paper, we propose an algorithm based on boosting to address these problems. The proposed algorithm builds base classifiers independently from each data type (view) that provides a partial view about an object of interest. Different from AdaBoost, where each view has its own re-sampling weight, our algorithm uses a single re-sampling distribution for all views at each boosting round. This distribution is determined by the view whose training error is minimal. This shared sampling mechanism restricts noise to individual views, thereby reducing sensitivity to noise. Furthermore, in order to establish performance guarantees, we introduce a randomized version of the algorithm, where a winning view is chosen probabilistically. As a result, it can be cast within a multi-armed bandit framework, which allows us to show that with high probability the algorithm seeks out most discriminant views of data for making predictions. We provide experimental results that show its performance against noise and competing techniques.

AB - Algorithms combining multi-view information are known to exponentially quicken classification, and have been applied to many fields. However, they lack the ability to mine most discriminant information sources (or data types) for making predictions. In this paper, we propose an algorithm based on boosting to address these problems. The proposed algorithm builds base classifiers independently from each data type (view) that provides a partial view about an object of interest. Different from AdaBoost, where each view has its own re-sampling weight, our algorithm uses a single re-sampling distribution for all views at each boosting round. This distribution is determined by the view whose training error is minimal. This shared sampling mechanism restricts noise to individual views, thereby reducing sensitivity to noise. Furthermore, in order to establish performance guarantees, we introduce a randomized version of the algorithm, where a winning view is chosen probabilistically. As a result, it can be cast within a multi-armed bandit framework, which allows us to show that with high probability the algorithm seeks out most discriminant views of data for making predictions. We provide experimental results that show its performance against noise and competing techniques.

KW - boosting

KW - convergence

KW - Data fusion

KW - multi-view learning

UR - http://www.scopus.com/inward/record.url?scp=80052412839&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-23783-6_38

DO - 10.1007/978-3-642-23783-6_38

M3 - Conference contribution

SN - 9783642237829

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

SP - 597

EP - 612

BT - Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2011, Proceedings

ER -

Peng J, Barbu C, Seetharaman G, Fan W, Wu X, Palaniappan K. ShareBoost: Boosting for multi-view learning with performance guarantees. In Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2011, Proceedings. PART 2 ed. 2011. p. 597-612. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). https://doi.org/10.1007/978-3-642-23783-6_38