Abstract
We study online linear optimization problems over concept classes which are defined in some combinatorial ways. Typically, those concept classes contain finite but exponentially many concepts and hence the complexity issue arises. In this paper, we survey some recent results on universal and efficient implementations of low-regret algorithmic frameworks such as Follow the Regularized Leader (FTRL) and Follow the Perturbed Leader (FPL). © 2013 Springer-Verlag.
Cite
CITATION STYLE
Takimoto, E., & Hatano, K. (2013). Efficient algorithms for combinatorial online prediction. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8139 LNAI, pp. 22–32). https://doi.org/10.1007/978-3-642-40935-6_3
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
