Scale-Free Algorithms for Online Linear Optimization
Abstract
We design algorithms for online linear optimization that have optimal regret and at the same time do not need to know any upper or lower bounds on the norm of the loss vectors. We achieve adaptiveness to norms of loss vectors by scale invariance, i.e., our algorithms make exactly the same decisions if the sequence of loss vectors is multiplied by any positive constant. Our algorithms work for any decision set, bounded or unbounded. For unbounded decisions sets, these are the first truly adaptive algorithms for online linear optimization.
Cite
Text
Orabona and Pál. "Scale-Free Algorithms for Online Linear Optimization." International Conference on Algorithmic Learning Theory, 2015. doi:10.1007/978-3-319-24486-0_19Markdown
[Orabona and Pál. "Scale-Free Algorithms for Online Linear Optimization." International Conference on Algorithmic Learning Theory, 2015.](https://mlanthology.org/alt/2015/orabona2015alt-scalefree/) doi:10.1007/978-3-319-24486-0_19BibTeX
@inproceedings{orabona2015alt-scalefree,
title = {{Scale-Free Algorithms for Online Linear Optimization}},
author = {Orabona, Francesco and Pál, Dávid},
booktitle = {International Conference on Algorithmic Learning Theory},
year = {2015},
pages = {287-301},
doi = {10.1007/978-3-319-24486-0_19},
url = {https://mlanthology.org/alt/2015/orabona2015alt-scalefree/}
}