Information-theoretic multi-view domain adaptation: A theoretical and empirical study

Pei Yang, Wei Gao

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

Multi-view learning aims to improve classification performance by leveraging the consistency among different views of data. The incorporation of multiple views was paid little attention in the studies of domain adaptation, where the view consistency based on source data is largely violated in the target domain due to the distribution gap between different domain data. In this paper, we leverage multiple views for cross-domain document classification. The central idea is to strengthen the views' consistency on target data by identifying the associations of domain-specific features from different domains. We present an Information-theoretic Multi-view Adaptation Model (IMAM) using a multi-way clustering scheme, where word and link clusters can draw together seemingly unrelated features across domains, which boosts the consistency between document clusterings that are based on the respective word and link views. Moreover, we demonstrate that IMAM can always find the document clustering with the minimal disagreement rate to the overlap of viewbased clusterings. We provide both theoretical and empirical justifications of the proposed method. Our experiments show that IMAM significantly outperforms traditional multi-view algorithm cotraining, the co-training-based adaptation algorithm CODA, the single-view transfer model CoCC and the large-margin-based multi-view transfer model MVTL-LM.

Original languageEnglish
Pages (from-to)501-525
Number of pages25
JournalJournal of Artificial Intelligence Research
Volume49
Publication statusPublished - 1 Jan 2014

Fingerprint

Experiments

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this

Information-theoretic multi-view domain adaptation : A theoretical and empirical study. / Yang, Pei; Gao, Wei.

In: Journal of Artificial Intelligence Research, Vol. 49, 01.01.2014, p. 501-525.

Research output: Contribution to journalArticle

@article{62b5748dc02c420a87e0f786ceee5d00,
title = "Information-theoretic multi-view domain adaptation: A theoretical and empirical study",
abstract = "Multi-view learning aims to improve classification performance by leveraging the consistency among different views of data. The incorporation of multiple views was paid little attention in the studies of domain adaptation, where the view consistency based on source data is largely violated in the target domain due to the distribution gap between different domain data. In this paper, we leverage multiple views for cross-domain document classification. The central idea is to strengthen the views' consistency on target data by identifying the associations of domain-specific features from different domains. We present an Information-theoretic Multi-view Adaptation Model (IMAM) using a multi-way clustering scheme, where word and link clusters can draw together seemingly unrelated features across domains, which boosts the consistency between document clusterings that are based on the respective word and link views. Moreover, we demonstrate that IMAM can always find the document clustering with the minimal disagreement rate to the overlap of viewbased clusterings. We provide both theoretical and empirical justifications of the proposed method. Our experiments show that IMAM significantly outperforms traditional multi-view algorithm cotraining, the co-training-based adaptation algorithm CODA, the single-view transfer model CoCC and the large-margin-based multi-view transfer model MVTL-LM.",
author = "Pei Yang and Wei Gao",
year = "2014",
month = "1",
day = "1",
language = "English",
volume = "49",
pages = "501--525",
journal = "Journal of Artificial Intelligence Research",
issn = "1076-9757",
publisher = "Morgan Kaufmann Publishers, Inc.",

}

TY - JOUR

T1 - Information-theoretic multi-view domain adaptation

T2 - A theoretical and empirical study

AU - Yang, Pei

AU - Gao, Wei

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Multi-view learning aims to improve classification performance by leveraging the consistency among different views of data. The incorporation of multiple views was paid little attention in the studies of domain adaptation, where the view consistency based on source data is largely violated in the target domain due to the distribution gap between different domain data. In this paper, we leverage multiple views for cross-domain document classification. The central idea is to strengthen the views' consistency on target data by identifying the associations of domain-specific features from different domains. We present an Information-theoretic Multi-view Adaptation Model (IMAM) using a multi-way clustering scheme, where word and link clusters can draw together seemingly unrelated features across domains, which boosts the consistency between document clusterings that are based on the respective word and link views. Moreover, we demonstrate that IMAM can always find the document clustering with the minimal disagreement rate to the overlap of viewbased clusterings. We provide both theoretical and empirical justifications of the proposed method. Our experiments show that IMAM significantly outperforms traditional multi-view algorithm cotraining, the co-training-based adaptation algorithm CODA, the single-view transfer model CoCC and the large-margin-based multi-view transfer model MVTL-LM.

AB - Multi-view learning aims to improve classification performance by leveraging the consistency among different views of data. The incorporation of multiple views was paid little attention in the studies of domain adaptation, where the view consistency based on source data is largely violated in the target domain due to the distribution gap between different domain data. In this paper, we leverage multiple views for cross-domain document classification. The central idea is to strengthen the views' consistency on target data by identifying the associations of domain-specific features from different domains. We present an Information-theoretic Multi-view Adaptation Model (IMAM) using a multi-way clustering scheme, where word and link clusters can draw together seemingly unrelated features across domains, which boosts the consistency between document clusterings that are based on the respective word and link views. Moreover, we demonstrate that IMAM can always find the document clustering with the minimal disagreement rate to the overlap of viewbased clusterings. We provide both theoretical and empirical justifications of the proposed method. Our experiments show that IMAM significantly outperforms traditional multi-view algorithm cotraining, the co-training-based adaptation algorithm CODA, the single-view transfer model CoCC and the large-margin-based multi-view transfer model MVTL-LM.

UR - http://www.scopus.com/inward/record.url?scp=84898611705&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84898611705&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:84898611705

VL - 49

SP - 501

EP - 525

JO - Journal of Artificial Intelligence Research

JF - Journal of Artificial Intelligence Research

SN - 1076-9757

ER -