No-Regret Algorithms for Heavy-Tailed Linear Bandits

@inproceedings{Medina2016NoRegretAF,
  title={No-Regret Algorithms for Heavy-Tailed Linear Bandits},
  author={Andres Mu{\~n}oz Medina and Scott Yang},
  booktitle={ICML},
  year={2016}
}
We analyze the problem of linear bandits under heavy tailed noise. Most of the work on linear bandits has been based on the assumption of bounded or sub-Gaussian noise. This assumption however is often violated in common scenarios such as financial markets. We present two algorithms to tackle this problem: one based on dynamic truncation and one based on a median of means estimator. We show that, when the noise admits only a 1 + ✏ moment, these algorithms are still able to achieve regret in e O… CONTINUE READING

From This Paper

Figures, tables, and topics from this paper.

References

Publications referenced by this paper.
Showing 1-10 of 22 references

Some aspects of the sequential design of experiments

  • H. Robbins
  • Bull. Amer. Math. Soc. 58(5), 527–535.
  • 1952
Highly Influential
4 Excerpts

Similar Papers

Loading similar papers…