Adaptive Rates of Convergence in Active Learning
Abstract
We study the rates of convergence in classification error achievable by active learning in the presence of label noise. Additionally, we study the more general problem of active learning with a nested hierarchy of hypothesis classes, and propose an algorithm whose error rate provably converges to the best achievable error among classifiers in the hierarchy at a rate adaptive to both the complexity of the optimal classifier and the noise conditions. In particular, we state sufficient conditions for these rates to be dramatically faster than those achievable by passive learning.
Cite
Text
Hanneke. "Adaptive Rates of Convergence in Active Learning." Annual Conference on Computational Learning Theory, 2009.Markdown
[Hanneke. "Adaptive Rates of Convergence in Active Learning." Annual Conference on Computational Learning Theory, 2009.](https://mlanthology.org/colt/2009/hanneke2009colt-adaptive/)BibTeX
@inproceedings{hanneke2009colt-adaptive,
title = {{Adaptive Rates of Convergence in Active Learning}},
author = {Hanneke, Steve},
booktitle = {Annual Conference on Computational Learning Theory},
year = {2009},
url = {https://mlanthology.org/colt/2009/hanneke2009colt-adaptive/}
}