Online methods for multi-domain learning and adaptation

Mark Dredze; Koby Crammer

doi:10.3115/1613715.1613801

Online methods for multi-domain learning and adaptation

Research output: Contribution to conference › Paper › peer-review

Abstract

NLP tasks are often domain specific, yet systems can learn behaviors across multiple domains. We develop a new multi-domain online learning framework based on parameter combination from multiple classifiers. Our algorithms draw from multi-task learning and domain adaptation to adapt multiple source domain classifiers to a new target domain, learn across multiple similar domains, and learn across a large number of disparate domains. We evaluate our algorithms on two popular NLP domain adaptation tasks: sentiment classification and spam filtering.

Original language	English (US)
Pages	689-697
Number of pages	9
DOIs	https://doi.org/10.3115/1613715.1613801
State	Published - 2008
Externally published	Yes
Event	2008 Conference on Empirical Methods in Natural Language Processing, EMNLP 2008, Co-located with AMTA 2008 and the International Workshop on Spoken Language Translation - Honolulu, HI, United States Duration: Oct 25 2008 → Oct 27 2008

Conference

Conference	2008 Conference on Empirical Methods in Natural Language Processing, EMNLP 2008, Co-located with AMTA 2008 and the International Workshop on Spoken Language Translation
Country/Territory	United States
City	Honolulu, HI
Period	10/25/08 → 10/27/08

ASJC Scopus subject areas

Computational Theory and Mathematics
Computer Science Applications
Information Systems

Access to Document

10.3115/1613715.1613801

Cite this

Online methods for multi-domain learning and adaptation. / Dredze, Mark; Crammer, Koby.
2008. 689-697 Paper presented at 2008 Conference on Empirical Methods in Natural Language Processing, EMNLP 2008, Co-located with AMTA 2008 and the International Workshop on Spoken Language Translation, Honolulu, HI, United States.

Research output: Contribution to conference › Paper › peer-review

Dredze, M & Crammer, K 2008, 'Online methods for multi-domain learning and adaptation', Paper presented at 2008 Conference on Empirical Methods in Natural Language Processing, EMNLP 2008, Co-located with AMTA 2008 and the International Workshop on Spoken Language Translation, Honolulu, HI, United States, 10/25/08 - 10/27/08 pp. 689-697. https://doi.org/10.3115/1613715.1613801

@conference{b2be90c4ce374a37ab239a7037c3d133,

title = "Online methods for multi-domain learning and adaptation",

abstract = "NLP tasks are often domain specific, yet systems can learn behaviors across multiple domains. We develop a new multi-domain online learning framework based on parameter combination from multiple classifiers. Our algorithms draw from multi-task learning and domain adaptation to adapt multiple source domain classifiers to a new target domain, learn across multiple similar domains, and learn across a large number of disparate domains. We evaluate our algorithms on two popular NLP domain adaptation tasks: sentiment classification and spam filtering.",

author = "Mark Dredze and Koby Crammer",

year = "2008",

doi = "10.3115/1613715.1613801",

language = "English (US)",

pages = "689--697",

note = "2008 Conference on Empirical Methods in Natural Language Processing, EMNLP 2008, Co-located with AMTA 2008 and the International Workshop on Spoken Language Translation ; Conference date: 25-10-2008 Through 27-10-2008",

}

TY - CONF

T1 - Online methods for multi-domain learning and adaptation

AU - Dredze, Mark

AU - Crammer, Koby

PY - 2008

Y1 - 2008

N2 - NLP tasks are often domain specific, yet systems can learn behaviors across multiple domains. We develop a new multi-domain online learning framework based on parameter combination from multiple classifiers. Our algorithms draw from multi-task learning and domain adaptation to adapt multiple source domain classifiers to a new target domain, learn across multiple similar domains, and learn across a large number of disparate domains. We evaluate our algorithms on two popular NLP domain adaptation tasks: sentiment classification and spam filtering.

AB - NLP tasks are often domain specific, yet systems can learn behaviors across multiple domains. We develop a new multi-domain online learning framework based on parameter combination from multiple classifiers. Our algorithms draw from multi-task learning and domain adaptation to adapt multiple source domain classifiers to a new target domain, learn across multiple similar domains, and learn across a large number of disparate domains. We evaluate our algorithms on two popular NLP domain adaptation tasks: sentiment classification and spam filtering.

UR - http://www.scopus.com/inward/record.url?scp=79951490162&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79951490162&partnerID=8YFLogxK

U2 - 10.3115/1613715.1613801

DO - 10.3115/1613715.1613801

M3 - Paper

AN - SCOPUS:79951490162

SP - 689

EP - 697

T2 - 2008 Conference on Empirical Methods in Natural Language Processing, EMNLP 2008, Co-located with AMTA 2008 and the International Workshop on Spoken Language Translation

Y2 - 25 October 2008 through 27 October 2008

ER -

Online methods for multi-domain learning and adaptation

Abstract

Conference

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this