Article

Agnostic active learning without constraints

Authors:

Alina Beygelzimer,

John Langford, and

Tong ZhangAuthors Info & Claims

NIPS'10: Proceedings of the 23rd International Conference on Neural Information Processing Systems - Volume 1

December 2010

Pages 199 - 207

Published: 06 December 2010 Publication History

Abstract

We present and analyze an agnostic active learning algorithm that works without keeping a version space. This is unlike all previous approaches where a restricted set of candidate hypotheses is maintained throughout learning, and only hypotheses from this set are ever returned. By avoiding this version space approach, our algorithm sheds the computational burden and brittleness associated with maintaining version spaces, yet still allows for substantial improvements over supervised learning for classification.

References

[1]

D. Cohn, L. Atlas, and R. Ladner. Improving generalization with active learning. Machine Learning, 15(2):201-221, 1994.

Digital Library

[2]

S. Dasgupta. Coarse sample complexity bounds for active learning. In Advances in Neural Information Processing Systems 18, 2005.

Digital Library

[3]

M.-F. Balcan, A. Beygelzimer, and J. Langford. Agnostic active learning. In Twenty-Third International Conference on Machine Learning, 2006.

Digital Library

[4]

S. Dasgupta, D. Hsu, and C. Monteleoni. A general agnostic active learning algorithm. In Advances in Neural Information Processing Systems 20, 2007.

Digital Library

[5]

A. Beygelzimer, S. Dasgupta, and J. Langford. Importance weighted active learning. In Twenty-Sixth International Conference on Machine Learning, 2009.

Digital Library

[6]

S. Hanneke. Adaptive rates of convergence in active learning. In Twenty-Second Annual Conference on Learning Theory, 2009.

[7]

V. Koltchinskii. Rademacher complexities and bounding the excess risk in active learning. Manuscript, 2009.

[8]

M.-F. Balcan, A. Broder, and T. Zhang. Margin based active learning. In Twentieth Annual Conference on Learning Theory, 2007.

Digital Library

[9]

R. S. Sutton and A. G. Barto. Reinforcement Learning: An Introduction. MIT Press, 1998.

Digital Library

[10]

P. Auer, N. Cesa-Bianchi, Y. Freund, and R. E. Schapire. The nonstochastic multiarmed bandit problem. SIAM Journal of Computing, 32:48-77, 2002.

Digital Library

[11]

M. Sugiyama, M. Krauledat, and K.-R. Müller. Covariate shift adaptation by importance weighted cross validation. Journal of Machine Learning Research, 8:985-1005, 2007.

Digital Library

[12]

M. Sugiyama. Active learning for misspecified models. In Advances in Neural Information Processing Systems 18, 2005.

Digital Library

[13]

F. Bach. Active learning for misspecified generalized linear models. In Advances in Neural Information Processing Systems 19, 2006.

Digital Library

[14]

S. Hanneke. A bound on the label complexity of agnostic active learning. In Twenty-Fourth International Conference on Machine Learning, 2007.

Digital Library

[15]

M. Kä äriäinen. Active learning in the non-realizable case. In Seventeenth International Conference on Algorithmic Learning Theory, 2006.

Digital Library

[16]

A. B. Tsybakov. Optimal aggregation of classifiers in statistical learning. Annals of Statistics, 32(1):135-166, 2004.

[17]

R. Castro and R. Nowak. Upper and lower bounds for active learning. In Allerton Conference on Communication, Control and Computing, 2006.

[18]

R. Castro and R. Nowak. Minimax bounds for active learning. In Twentieth Annual Conference on Learning Theory, 2007.

Digital Library

[19]

T. Zhang. Data dependent concentration bounds for sequential prediction algorithms. In Eighteenth Annual Conference on Learning Theory, 2005.

Digital Library

[20]

E. Friedman. Active learning for smooth problems. In Twenty-Second Annual Conference on Learning Theory, 2009.

[21]

L. Wang. Sufficient conditions for agnostic active learnable. In Advances in Neural Information Processing Systems 22, 2009.

Digital Library

[22]

S. Dasgupta and D. Hsu. Hierarchical sampling for active learning. In Twenty-Fifth International Conference on Machine Learning, 2008.

Digital Library

Cited By

Tifrea AClarysse JYang FKrause ABrunskill ECho KEngelhardt BSabato SScarlett J(2023)Margin-based sampling in high dimensionsProceedings of the 40th International Conference on Machine Learning10.5555/3618408.3619832(34222-34262)Online publication date: 23-Jul-2023
https://dl.acm.org/doi/10.5555/3618408.3619832
Tang YHuang SKoyejo SMohamed SAgarwal ABelgrave DCho KOh A(2022)Active learning for multiple target modelsProceedings of the 36th International Conference on Neural Information Processing Systems10.5555/3600270.3603054(38424-38435)Online publication date: 28-Nov-2022
https://dl.acm.org/doi/10.5555/3600270.3603054
Krishnamurthy AAgarwal AHuang TDaumé HLangford J(2021)Active learning for cost-sensitive classificationThe Journal of Machine Learning Research10.5555/3322706.336200620:1(2334-2383)Online publication date: 9-Mar-2021
https://dl.acm.org/doi/10.5555/3322706.3362006
Show More Cited By

Agnostic active learning without constraints
1. Computing methodologies

Recommendations

Toward Efficient Agnostic Learning
Special issue on computational learning theory, COLT'92

In this paper we initiate an investigation of generalizations of the Probably Approximately Correct (PAC) learning model that attempt to significantly weaken the target function assumptions. The ultimate goal in this direction is informally termed ...
Read More
Agnostic active learning

We state and analyze the first active learning algorithm that finds an @e-optimal hypothesis in any hypothesis class, when the underlying distribution has arbitrary forms of noise. The algorithm, A^2 (for Agnostic Active), relies only upon the ...
Read More
Smoothness, Disagreement Coefficient, and the Label Complexity of Agnostic Active Learning

We study pool-based active learning in the presence of noise, that is, the agnostic setting. It is known that the effectiveness of agnostic active learning depends on the learning problem and the hypothesis space. Although there are many cases on which ...
Read More

Comments

Information & Contributors

Information

Published In

cover image Guide Proceedings

NIPS'10: Proceedings of the 23rd International Conference on Neural Information Processing Systems - Volume 1

December 2010

2630 pages

Publisher

Curran Associates Inc.

Red Hook, NY, United States

Publication History

Published: 06 December 2010

Qualifiers

Article

Contributors

Other Metrics

View Article Metrics

Bibliometrics & Citations

Bibliometrics

Article Metrics

22
Total Citations
View Citations
0
Total Downloads

Downloads (Last 12 months)0
Downloads (Last 6 weeks)0

Other Metrics

View Author Metrics

Citations

Cited By

Tifrea AClarysse JYang FKrause ABrunskill ECho KEngelhardt BSabato SScarlett J(2023)Margin-based sampling in high dimensionsProceedings of the 40th International Conference on Machine Learning10.5555/3618408.3619832(34222-34262)Online publication date: 23-Jul-2023
https://dl.acm.org/doi/10.5555/3618408.3619832
Tang YHuang SKoyejo SMohamed SAgarwal ABelgrave DCho KOh A(2022)Active learning for multiple target modelsProceedings of the 36th International Conference on Neural Information Processing Systems10.5555/3600270.3603054(38424-38435)Online publication date: 28-Nov-2022
https://dl.acm.org/doi/10.5555/3600270.3603054
Krishnamurthy AAgarwal AHuang TDaumé HLangford J(2021)Active learning for cost-sensitive classificationThe Journal of Machine Learning Research10.5555/3322706.336200620:1(2334-2383)Online publication date: 9-Mar-2021
https://dl.acm.org/doi/10.5555/3322706.3362006
Krishnamurthy AAgarwal AHuang TDaumé HLangford J(2017)Active learning for cost-sensitive classificationProceedings of the 34th International Conference on Machine Learning - Volume 7010.5555/3305381.3305579(1915-1924)Online publication date: 6-Aug-2017
https://dl.acm.org/doi/10.5555/3305381.3305579
Yan SZhang C(2017)Revisiting perceptronProceedings of the 31st International Conference on Neural Information Processing Systems10.5555/3294771.3294872(1056-1066)Online publication date: 4-Dec-2017
https://dl.acm.org/doi/10.5555/3294771.3294872
Sourati JAkcakaya MLeen TErdogmus DDy J(2017)Asymptotic analysis of objectives based on fisher information in active learningThe Journal of Machine Learning Research10.5555/3122009.312204318:1(1123-1163)Online publication date: 1-Jan-2017
https://dl.acm.org/doi/10.5555/3122009.3122043
Qian KPopa LSen PLim EWinslett MSanderson MFu ASun JCulpepper SLo EHo JDonato DAgrawal RZheng YCastillo CSun ATseng VLi C(2017)Active Learning for Large-Scale Entity ResolutionProceedings of the 2017 ACM on Conference on Information and Knowledge Management10.1145/3132847.3132949(1379-1388)Online publication date: 6-Nov-2017
https://dl.acm.org/doi/10.1145/3132847.3132949
Awasthi PBalcan MLong P(2017)The Power of Localization for Efficiently Learning Linear Separators with NoiseJournal of the ACM10.1145/300638463:6(1-27)Online publication date: 9-Jan-2017
https://dl.acm.org/doi/10.1145/3006384
Beygelzimer AHsu DLangford JZhang C(2016)Search improves label for active learningProceedings of the 30th International Conference on Neural Information Processing Systems10.5555/3157382.3157472(3350-3358)Online publication date: 5-Dec-2016
https://dl.acm.org/doi/10.5555/3157382.3157472
Huang TLi LVartanian AAmershi SZhu X(2016)Active learning with oracle epiphanyProceedings of the 30th International Conference on Neural Information Processing Systems10.5555/3157382.3157414(2828-2836)Online publication date: 5-Dec-2016
https://dl.acm.org/doi/10.5555/3157382.3157414
Show More Cited By

View Options

View options

Media

Figures

Other

Tables

View Table of Contents