Bounded Regret for Finite-Armed Structured Bandits

Abstract

We study a new type of K-armed bandit problem where the expected return of one arm may depend on the returns of other arms. We present a new algorithm for this general class of problems and show that under certain circumstances it is possible to achieve finite expected cumulative regret. We also give problem-dependent lower bounds on the cumulative regret showing that at least in special cases the new algorithm is nearly optimal.

Cite

Text

Lattimore and Munos. "Bounded Regret for Finite-Armed Structured Bandits." Neural Information Processing Systems, 2014.

Markdown

[Lattimore and Munos. "Bounded Regret for Finite-Armed Structured Bandits." Neural Information Processing Systems, 2014.](https://mlanthology.org/neurips/2014/lattimore2014neurips-bounded/)

BibTeX

@inproceedings{lattimore2014neurips-bounded,
  title     = {{Bounded Regret for Finite-Armed Structured Bandits}},
  author    = {Lattimore, Tor and Munos, Remi},
  booktitle = {Neural Information Processing Systems},
  year      = {2014},
  pages     = {550-558},
  url       = {https://mlanthology.org/neurips/2014/lattimore2014neurips-bounded/}
}