Improved sleeping bandits with stochastic action sets and adversarial rewards - Archive ouverte HAL Access content directly
Conference Papers Year :

## Improved sleeping bandits with stochastic action sets and adversarial rewards

(1) , (2) , (3)
1
2
3
• Function : Author
Pierre Gaillard
Michal Valko

#### Abstract

In this paper, we consider the problem of sleeping bandits with stochastic action sets and adversarial rewards. In this setting, in contrast to most work in bandits, the actions may not be available at all times. For instance, some products might be out of stock in item recommendation. The best existing efficient (i.e., polynomial-time) algorithms for this problem only guarantee an $O(T 2/3)$ upper-bound on the regret. Yet, inefficient algorithms based on EXP4 can achieve $O(√ T)$. In this paper , we provide a new computationally efficient algorithm inspired by EXP3 satisfying a regret of order $O(√ T)$ when the availabilities of each action $i ∈ A$ are independent. We then study the most general version of the problem where at each round available sets are generated from some unknown arbitrary distribution (i.e., without the independence assumption) and propose an efficient algorithm with $O(√ 2 K T)$ regret guarantee. Our theoretical results are corroborated with experimental evaluations.

#### Domains

Statistics [stat] Machine Learning [stat.ML]

### Dates and versions

hal-02950106 , version 1 (27-09-2020)

### Identifiers

• HAL Id : hal-02950106 , version 1

### Cite

Aadirupa Saha, Pierre Gaillard, Michal Valko. Improved sleeping bandits with stochastic action sets and adversarial rewards. International Conference on Machine Learning, 2020, Vienna, Austria. ⟨hal-02950106⟩

### Export

BibTeX TEI Dublin Core DC Terms EndNote Datacite

986 View