Online Learning: Random Averages, Combinatorial Parameters, and Learnability
Abstract
We develop a theory of online learning by defining several complexity measures. Among them are analogues of Rademacher complexity, covering numbers and fat-shattering dimension from statistical learning theory. Relationship among these complexity measures, their connection to online learning, and tools for bounding them are provided. We apply these results to various learning problems. We provide a complete characterization of online learnability in the supervised setting.
Cite
Text
Rakhlin et al. "Online Learning: Random Averages, Combinatorial Parameters, and Learnability." Neural Information Processing Systems, 2010.Markdown
[Rakhlin et al. "Online Learning: Random Averages, Combinatorial Parameters, and Learnability." Neural Information Processing Systems, 2010.](https://mlanthology.org/neurips/2010/rakhlin2010neurips-online/)BibTeX
@inproceedings{rakhlin2010neurips-online,
title = {{Online Learning: Random Averages, Combinatorial Parameters, and Learnability}},
author = {Rakhlin, Alexander and Sridharan, Karthik and Tewari, Ambuj},
booktitle = {Neural Information Processing Systems},
year = {2010},
pages = {1984-1992},
url = {https://mlanthology.org/neurips/2010/rakhlin2010neurips-online/}
}