Online Learning of Non-Stationary Sequences

Abstract

We consider an online learning scenario in which the learner can make predictions on the basis of a fixed set of experts. We derive upper and lower relative loss bounds for a class of universal learning algorithms in- volving a switching dynamics over the choice of the experts. On the basis of the performance bounds we provide the optimal a priori discretiza- tion for learning the parameter that governs the switching dynamics. We demonstrate the new algorithm in the context of wireless networks.

Cite

Text

Monteleoni and Jaakkola. "Online Learning of Non-Stationary Sequences." Neural Information Processing Systems, 2003.

Markdown

[Monteleoni and Jaakkola. "Online Learning of Non-Stationary Sequences." Neural Information Processing Systems, 2003.](https://mlanthology.org/neurips/2003/monteleoni2003neurips-online/)

BibTeX

@inproceedings{monteleoni2003neurips-online,
  title     = {{Online Learning of Non-Stationary Sequences}},
  author    = {Monteleoni, Claire and Jaakkola, Tommi S.},
  booktitle = {Neural Information Processing Systems},
  year      = {2003},
  pages     = {1093-1100},
  url       = {https://mlanthology.org/neurips/2003/monteleoni2003neurips-online/}
}