Confidence-weighted linear classification

Mark Dredze; Koby Crammer; Fernando Pereira

doi:10.1145/1390156.1390190

Confidence-weighted linear classification

Mark Dredze, Koby Crammer, Fernando Pereira

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

Abstract

We introduce confidence-weighted linear classifiers, which add parameter confidence information to linear classifiers. Online learners in this setting update both classifier parameters and the estimate of their confidence. The particular online algorithms we study here maintain a Gaussian distribution over parameter vectors and update the mean and eovarianee of the distribution with each instance. Empirical evaluation on a range of NLP tasks show that our algorithm improves over other state of the art online and batch methods, learns faster in the online setting, and lends itself to better classifier combination after parallel training.

Original language	English (US)
Title of host publication	Proceedings of the 25th International Conference on Machine Learning
Publisher	Association for Computing Machinery (ACM)
Pages	264-271
Number of pages	8
ISBN (Print)	9781605582054
DOIs	https://doi.org/10.1145/1390156.1390190
State	Published - 2008
Externally published	Yes
Event	25th International Conference on Machine Learning - Helsinki, Finland Duration: Jul 5 2008 → Jul 9 2008

Publication series

Name	Proceedings of the 25th International Conference on Machine Learning

Conference

Conference	25th International Conference on Machine Learning
Country/Territory	Finland
City	Helsinki
Period	7/5/08 → 7/9/08

ASJC Scopus subject areas

Artificial Intelligence
Human-Computer Interaction
Software

Access to Document

10.1145/1390156.1390190

Cite this

Confidence-weighted linear classification. / Dredze, Mark; Crammer, Koby; Pereira, Fernando.
Proceedings of the 25th International Conference on Machine Learning. Association for Computing Machinery (ACM), 2008. p. 264-271 (Proceedings of the 25th International Conference on Machine Learning).

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

Dredze, M, Crammer, K & Pereira, F 2008, Confidence-weighted linear classification. in Proceedings of the 25th International Conference on Machine Learning. Proceedings of the 25th International Conference on Machine Learning, Association for Computing Machinery (ACM), pp. 264-271, 25th International Conference on Machine Learning, Helsinki, Finland, 7/5/08. https://doi.org/10.1145/1390156.1390190

@inproceedings{c2474db38a6d4e6095d67561a0f74aa6,

title = "Confidence-weighted linear classification",

abstract = "We introduce confidence-weighted linear classifiers, which add parameter confidence information to linear classifiers. Online learners in this setting update both classifier parameters and the estimate of their confidence. The particular online algorithms we study here maintain a Gaussian distribution over parameter vectors and update the mean and eovarianee of the distribution with each instance. Empirical evaluation on a range of NLP tasks show that our algorithm improves over other state of the art online and batch methods, learns faster in the online setting, and lends itself to better classifier combination after parallel training.",

author = "Mark Dredze and Koby Crammer and Fernando Pereira",

year = "2008",

doi = "10.1145/1390156.1390190",

language = "English (US)",

isbn = "9781605582054",

series = "Proceedings of the 25th International Conference on Machine Learning",

publisher = "Association for Computing Machinery (ACM)",

pages = "264--271",

booktitle = "Proceedings of the 25th International Conference on Machine Learning",

address = "United States",

note = "25th International Conference on Machine Learning ; Conference date: 05-07-2008 Through 09-07-2008",

}

TY - GEN

T1 - Confidence-weighted linear classification

AU - Dredze, Mark

AU - Crammer, Koby

AU - Pereira, Fernando

PY - 2008

Y1 - 2008

N2 - We introduce confidence-weighted linear classifiers, which add parameter confidence information to linear classifiers. Online learners in this setting update both classifier parameters and the estimate of their confidence. The particular online algorithms we study here maintain a Gaussian distribution over parameter vectors and update the mean and eovarianee of the distribution with each instance. Empirical evaluation on a range of NLP tasks show that our algorithm improves over other state of the art online and batch methods, learns faster in the online setting, and lends itself to better classifier combination after parallel training.

AB - We introduce confidence-weighted linear classifiers, which add parameter confidence information to linear classifiers. Online learners in this setting update both classifier parameters and the estimate of their confidence. The particular online algorithms we study here maintain a Gaussian distribution over parameter vectors and update the mean and eovarianee of the distribution with each instance. Empirical evaluation on a range of NLP tasks show that our algorithm improves over other state of the art online and batch methods, learns faster in the online setting, and lends itself to better classifier combination after parallel training.

UR - http://www.scopus.com/inward/record.url?scp=56449101965&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=56449101965&partnerID=8YFLogxK

U2 - 10.1145/1390156.1390190

DO - 10.1145/1390156.1390190

M3 - Conference contribution

AN - SCOPUS:56449101965

SN - 9781605582054

T3 - Proceedings of the 25th International Conference on Machine Learning

SP - 264

EP - 271

BT - Proceedings of the 25th International Conference on Machine Learning

PB - Association for Computing Machinery (ACM)

T2 - 25th International Conference on Machine Learning

Y2 - 5 July 2008 through 9 July 2008

ER -

Confidence-weighted linear classification

Abstract

Publication series

Conference

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this