Active Learning with Statistical Models
David A. Cohn, Zoubin Ghahramani, Michael I. Jordan · Journal of Artificial Intelligence Research · 1996
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 efficient 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.
Cited by 1,317 works. doi:10.1613/jair.295 Open-access PDF
Authors
References
- Neural Networks and the Bias/Variance Dilemma
- Statistical Analysis of Finite Mixture Distributions.
- A general regression neural network
- Queries and Concept Learning
- Statistical Analysis of Finite Mixture Distributions.
- Theory of Optimal Experiments.
- Learning Bayesian networks: The combination of knowledge and statistical data
- Applied Linear Regression.
- Improving Generalization with Active Learning
- Empirical Model-Building and Response Surfaces.