Agnostic active learning

Agnostic active learning (2006)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Maria-florina Balcan , Alina Beygelzimer , John Langford

Venue:	In ICML
Citations:	80 - 10 self

BibTeX

@INPROCEEDINGS{Balcan06agnosticactive,
    author = {Maria-florina Balcan and Alina Beygelzimer and John Langford},
    title = {Agnostic active learning},
    booktitle = {In ICML},
    year = {2006},
    pages = {65--72},
    publisher = {ICML}
}

Years of Citing Articles

Bookmark

OpenURL

Abstract

We state and analyze the first active learning algorithm which works in the presence of arbitrary forms of noise. The algorithm, A2 (for Agnostic Active), relies only upon the assumption that the samples are drawn i.i.d. from a fixed distribution. We show that A2 achieves an exponential improvement (i.e., requires only O � ln 1 ɛ samples to find an ɛ-optimal classifier) over the usual sample complexity of supervised learning, for several settings considered before in the realizable case. These include learning threshold classifiers and learning homogeneous linear separators with respect to an input distribution which is uniform over the unit sphere. 1.

Citations

946	Combining labeled and unlabeled data with co-training - Blum, Mitchell - 1998
816	On the uniform convergence of relative frequencies of events to their probabilities. Theory of Prob. and its - Vapnik, Chervonenkis - 1971
632	Text classification from labeled and unlabeled documents using - Nigram, McCallum, et al.
582	Queries and concept learning - Angluin - 1988
509	Transductive inference for text classification using support vector machines - Joachims - 1999
358	TM: Machine Learning - Mitchell - 1997
334	Improving generalization with active learning - Cohn, Atlas, et al. - 1994
256	Selective sampling using the query by committee algorithm, Machine Learning 28 - Freund, Seung, et al. - 1997
255	PL: Neural Network Learning: Theoretical Foundations - Anthony, Bartlett - 1999
202	Learning from Labeled and Unlabeled Data Using Graph Mincuts - Blum, Chawla
171	Semi-Supervised Learning - Chapelle, Schlkopf, et al. - 2006
150	An efficient membership-query algorithm for learning dnf with respect to the uniform distribution - Jackson - 1994
67	Coarse sample complexity bounds for active learning - Dasgupta - 2005
63	Unsupervised improvement of visual detectors using co-training - Levin, Viola, et al. - 2003
51	Analysis of perceptron-based active learning - Dasgupta, Kalai, et al. - 2009
50	Queries revisited - Angluin - 2001
50	Analysis of a greedy active learning strategy - Dasgupta - 2004
46	A general agnostic active learning algorithm - Dasgupta, Hsu, et al.
44	A PAC-style model for learning from labeled and unlabeled data - Balcan, Blum - 2005
40	A bound on the label complexity of agnostic active learning - Hanneke - 2007
32	Theory of Classification: a Survey of Recent Advances. ESAIM: Probability and Statistics, 9:323-375. MR2182250 Wegkamp/Lasso type classifiers with a reject option 168 - Boucheron, Bousquet, et al. - 2005
31	Minimax bounds for active learning - Castro, Nowak - 2007
30	Active learning in the non-realizable case - Kääriäinen - 2006
29	Query learning can work poorly when a human oracle is used - Baum, Lang - 1993
28	Margin based active learning - Balcan, Broder, et al. - 2007
25	Faster rates in regression via active learning - Castro, Willett, et al. - 2005
24	The true sample complexity of active learning - Balcan, Hanneke, et al.
20	Corrected co-training for statistical parsers - Hwa, Osborne, et al. - 2003
17	An interval estimation problem for controlled observations - Burnashev, Zigangirov - 1974
17	On the sample complexity of PAC learning halfspaces against the uniform distribution - Long - 1995
8	Ulam’s searching game with lies - Czyzowicz, Mundici, et al. - 1989
6	Co-trained support vector machines for large scale unstructured document classification using unlabeled data and syntactic information - Park, Zhang - 2004
6	Improving generalzation with active learning - Cohn, Atlas, et al. - 1994
4	Learning monotone dnf from a teacher that almost does not answer membership queries - Bshouty, Eiron
3	Asymptotic active learning - Balcan, Even-Dar, et al. - 2007
3	Machine learning theory - Blum - 2007
1	Agnostic Active Learning - Cohen, Atlas, et al. - 1994
1	Information, prediction, and query by comittee. Neural Information Processing Systems - Freund, Seung, et al. - 1993
1	Appendix We use the following standard Sample Complexity bound from Anthony and Bartlett - unknown authors - 1999
1	A Standard Results The following standard Sample Complexity bound is in [3]. Theorem A.1 Suppose that H is a set of functions from X to {−1, 1} with finite VC-dimension VH ≥ 1. Let D be an arbitrary, but fixed probability distribution over X × {−1, 1}. Fo - Wiley, Inc - 1998