Sciweavers

Share
NIPS
1994

Active Learning with Statistical Models

8 years 9 months ago
Active Learning with Statistical Models
For many types of machine learning algorithms, one can compute the statistically optimal" way to select training data. In this paper, we review how optimal data selection techniques have been used with feedforward neural networks. We then show how the same principles may be used to select data for two alternative, statistically-based learning architectures: mixtures of Gaussians and locally weighted regression. While the techniques for neural networks are computationally expensive and approximate, the techniques for mixtures of Gaussians and locally weighted regression are both e cient and accurate. Empirically, we observe that the optimality criterion sharply decreases the number of training examples the learner needs in order to achieve good performance.
David A. Cohn, Zoubin Ghahramani, Michael I. Jorda
Added 02 Nov 2010
Updated 02 Nov 2010
Type Conference
Year 1994
Where NIPS
Authors David A. Cohn, Zoubin Ghahramani, Michael I. Jordan
Comments (0)
books