On reducing the order of arm-passes bandit streaming algorithms under memory bottleneck

TL;DR

This paper improves multi-arm bandit streaming algorithms under memory constraints by reducing the number of passes needed to achieve near-optimal regret, offering more efficient solutions in resource-limited settings.

Contribution

It introduces algorithms that decrease the number of streaming passes required for low-regret bandit performance under memory bottlenecks, including 2-pass algorithms with initial conditions.

Findings

Reduced the number of passes for near-optimal regret by a logarithmic factor.

Developed 2-pass algorithms achieving similar regret bounds under initial conditions.

Enhanced efficiency of bandit algorithms in memory-constrained streaming environments.

Abstract

In this work we explore multi-arm bandit streaming model, especially in cases where the model faces resource bottleneck. We build over existing algorithms conditioned by limited arm memory at any instance of time. Specifically, we improve the amount of streaming passes it takes for a bandit algorithm to incur a $O(\sqrt{T\log(T)})$ regret by a logarithmic factor, and also provide 2-pass algorithms with some initial conditions to incur a similar order of regret.