Near Optimal Pure Exploration in Logistic Bandits
This addresses a gap in bandit theory for researchers and practitioners, providing a near-optimal solution for logistic bandits, though it is incremental as it builds on existing track-and-stop methods.
The paper tackles the lack of optimal algorithms for pure exploration problems in generalized linear model bandits, specifically logistic bandits, by developing the first track-and-stop algorithm called Log-TS, which asymptotically matches an approximation for the instance-specific lower bound of expected sample complexity up to a logarithmic factor.
Bandit algorithms have garnered significant attention due to their practical applications in real-world scenarios. However, beyond simple settings such as multi-arm or linear bandits, optimal algorithms remain scarce. Notably, no optimal solution exists for pure exploration problems in the context of generalized linear model (GLM) bandits. In this paper, we narrow this gap and develop the first track-and-stop algorithm for general pure exploration problems under the logistic bandit called logistic track-and-stop (Log-TS). Log-TS is an efficient algorithm that asymptotically matches an approximation for the instance-specific lower bound of the expected sample complexity up to a logarithmic factor.