Bandit Pareto Set Identification in a Multi-Output Linear Model
Keywords: bandit, pure exploration, bandit pareto set identification, best arm identification
TL;DR: In this paper we introduce an optimal design algorithm to solve bandit Pareto set identification when linear relation between arms means (unknown) and some known features is assumed.
Abstract: We study the Pareto Set Identification (PSI) problem in a structured multi-output linear bandit model. In this setting each arm is associated a feature vector belonging to $\mathbb{R}^h$ and its mean vector in $\mathbb{R}^d$ linearly depends on this feature vector through a common unknown matrix $\Theta \in \mathbb{R}^{h \times d}$. The goal is to identity the set of non-dominated arms by adaptively collecting samples from the arms. We introduce and analyze the first optimal design-based algorithms for PSI, providing nearly optimal guarantees in both the fixed-budget and the fixed-confidence settings. Notably, we show that the difficulty of these tasks mainly depends on the sub-optimality gaps of $h$ arms only.
Our theoretical results are supported by an extensive benchmark on synthetic and real-world datasets.
Submission Number: 55
Loading