Diameter-Based Active Learning

Christopher Tosh; Sanjoy Dasgupta

ICML 2017poster35 citations

Diameter-Based Active Learning

Abstract

To date, the tightest upper and lower-bounds for the active learning of general concept classes have been in terms of a parameter of the learning problem called the splitting index. We provide, for the first time, an efficient algorithm that is able to realize this upper bound, and we empirically demonstrate its good performance.

BibTeX

@InProceedings{pmlr-v70-tosh17a,
  title = 	 {Diameter-Based Active Learning},
  author =       {Christopher Tosh and Sanjoy Dasgupta},
  booktitle = 	 {Proceedings of the 34th International Conference on Machine Learning},
  pages = 	 {3444--3452},
  year = 	 {2017},
  editor = 	 {Precup, Doina and Teh, Yee Whye},
  volume = 	 {70},
  series = 	 {Proceedings of Machine Learning Research},
  month = 	 {06--11 Aug},
  publisher =    {PMLR},
  pdf = 	 {http://proceedings.mlr.press/v70/tosh17a/tosh17a.pdf},
  url = 	 {https://proceedings.mlr.press/v70/tosh17a.html},
  abstract = 	 {To date, the tightest upper and lower-bounds for the active learning of general concept classes have been in terms of a parameter of the learning problem called the splitting index. We provide, for the first time, an efficient algorithm that is able to realize this upper bound, and we empirically demonstrate its good performance.}
}