Optimal and Adaptive Algorithms for Online Boosting

Abstract

We study online boosting, the task of converting any weak online learner into a strong online learner. Based on a novel and natural definition of weak online learnability, we develop two online boosting algorithms. The first algorithm is an online version of boost-by-majority. By proving a matching lower bound, we show that this algorithm is essentially optimal in terms of the number of weak learners and the sample complexity needed to achieve a specified accuracy. The second algorithm is adaptive and parameter-free, albeit not optimal.

Cite

Text

Beygelzimer et al. "Optimal and Adaptive Algorithms for Online Boosting." International Conference on Machine Learning, 2015.

Markdown

[Beygelzimer et al. "Optimal and Adaptive Algorithms for Online Boosting." International Conference on Machine Learning, 2015.](https://mlanthology.org/icml/2015/beygelzimer2015icml-optimal/)

BibTeX

@inproceedings{beygelzimer2015icml-optimal,
  title     = {{Optimal and Adaptive Algorithms for Online Boosting}},
  author    = {Beygelzimer, Alina and Kale, Satyen and Luo, Haipeng},
  booktitle = {International Conference on Machine Learning},
  year      = {2015},
  pages     = {2323-2331},
  volume    = {37},
  url       = {https://mlanthology.org/icml/2015/beygelzimer2015icml-optimal/}
}