Normalized Online Learning
Machine Learning
2013-05-30 v1 Machine Learning
Abstract
We introduce online learning algorithms which are independent of feature scales, proving regret bounds dependent on the ratio of scales existent in the data rather than the absolute scale. This has several useful effects: there is no need to pre-normalize data, the test-time and test-space complexity are reduced, and the algorithms are more robust.
Cite
@article{arxiv.1305.6646,
title = {Normalized Online Learning},
author = {Stephane Ross and Paul Mineiro and John Langford},
journal= {arXiv preprint arXiv:1305.6646},
year = {2013}
}
Comments
Appears in Proceedings of the Twenty-Ninth Conference on Uncertainty in Artificial Intelligence (UAI2013)