Active learning for regression based on query by committee

Authors:
Robert Burbidge;Jem J. Rowland;Ross D. King
Affiliations:
Department of Computer Science, University of Wales, Aberystwyth, Aberystwyth, UK;Department of Computer Science, University of Wales, Aberystwyth, Aberystwyth, UK;Department of Computer Science, University of Wales, Aberystwyth, Aberystwyth, UK
Venue:
IDEAL'07 Proceedings of the 8th international conference on Intelligent data engineering and automated learning
Year:
2007

Citing 9
Cited 2

Query by committee

COLT '92 Proceedings of the fifth annual workshop on Computational learning theory
Neural networks and the bias/variance dilemma

Neural Computation
Selective Sampling Using the Query by Committee Algorithm

Machine Learning
Neural Networks for Pattern Recognition

Neural Networks for Pattern Recognition
Query Learning Strategies Using Boosting and Bagging

ICML '98 Proceedings of the Fifteenth International Conference on Machine Learning
Selective Sampling for Nearest Neighbor Classifiers

Machine Learning
Active Learning in Approximately Linear Regression Based on Conditional Expectation of Generalization Error

The Journal of Machine Learning Research
Active learning with statistical models

Journal of Artificial Intelligence Research
An active learning framework for content-based information retrieval

IEEE Transactions on Multimedia

Characterising enzymes for information processing: towards an artificial experimenter

UC'10 Proceedings of the 9th international conference on Unconventional computation
An artificial experimenter for enzymatic response characterisation

DS'10 Proceedings of the 13th international conference on Discovery science

Quantified Score

Hi-index	0.00

Visualization

Abstract

We investigate a committee-based approach for active learning of real-valued functions. This is a variance-only strategy for selection of informative training data. As such it is shown to suffer when the model class is misspecified since the learner's bias is high. Conversely, the strategy outperforms passive selection when the model class is very expressive since active minimization of the variance avoids overfitting.