Open Problem: Adversarial Multiarmed Bandits with Limited Advice

Abstract

Adversarial multiarmed bandits with expert advice is one of the fundamental problems in studying the exploration-exploitation trade-off. It is known that if we observe the advice of all experts on every round we can achieve O\left(\sqrt{KT} \ln N\right) regret, where K is the number of arms, T is the number of game rounds, and N is the number of experts. It is also known that if we observe the advice of just one expert on every round, we can achieve regret of order O\left(\sqrt{NT}\right). Our open problem is what can be achieved by asking M experts on every round, where 1<M<N.

Cite

Text

Seldin et al. "Open Problem: Adversarial Multiarmed Bandits with Limited Advice." Annual Conference on Computational Learning Theory, 2013.

Markdown

[Seldin et al. "Open Problem: Adversarial Multiarmed Bandits with Limited Advice." Annual Conference on Computational Learning Theory, 2013.](https://mlanthology.org/colt/2013/seldin2013colt-open/)

BibTeX

@inproceedings{seldin2013colt-open,
  title     = {{Open Problem: Adversarial Multiarmed Bandits with Limited Advice}},
  author    = {Seldin, Yevgeny and Crammer, Koby and Bartlett, Peter L.},
  booktitle = {Annual Conference on Computational Learning Theory},
  year      = {2013},
  pages     = {1067-1072},
  url       = {https://mlanthology.org/colt/2013/seldin2013colt-open/}
}