Active Learning with Safety Constraints

TL;DR

This paper introduces a novel adaptive algorithm for safe active learning in linear bandit settings, efficiently identifying optimal safe decisions while respecting unknown safety constraints, with promising results on synthetic and real datasets.

Contribution

It presents the first method for best-arm identification in linear bandits with safety constraints, combining adaptive experimental design with safety considerations.

Findings

Efficiently balances safety verification and optimality in decision-making.

Performs well on both synthetic and real-world datasets.

First to address safe best-arm identification in linear bandits.

Abstract

Active learning methods have shown great promise in reducing the number of samples necessary for learning. As automated learning systems are adopted into real-time, real-world decision-making pipelines, it is increasingly important that such algorithms are designed with safety in mind. In this work we investigate the complexity of learning the best safe decision in interactive environments. We reduce this problem to a constrained linear bandits problem, where our goal is to find the best arm satisfying certain (unknown) safety constraints. We propose an adaptive experimental design-based algorithm, which we show efficiently trades off between the difficulty of showing an arm is unsafe vs suboptimal. To our knowledge, our results are the first on best-arm identification in linear bandits with safety constraints. In practice, we demonstrate that this approach performs well on synthetic and real world datasets.