Combining Diverse Information for Coordinated Action: Stochastic Bandit Algorithms for Heterogeneous Agents

TL;DR

This paper introduces Min-Width, a UCB-style algorithm for multi-agent stochastic bandits with heterogeneous rewards, enabling efficient coordination and information aggregation among diverse agents with different sensitivities.

Contribution

The paper proposes a novel algorithm, Min-Width, that effectively aggregates heterogeneous agent rewards and coordinates actions in stochastic bandit problems, addressing a gap in existing methods.

Findings

Modeling agent heterogeneity improves performance when sensitivities vary widely.

More information sharing does not always lead to better results.

Min-Width achieves favorable regret bounds in synthetic experiments.

Abstract

Stochastic multi-agent multi-armed bandits typically assume that the rewards from each arm follow a fixed distribution, regardless of which agent pulls the arm. However, in many real-world settings, rewards can depend on the sensitivity of each agent to their environment. In medical screening, disease detection rates can vary by test type; in preference matching, rewards can depend on user preferences; and in environmental sensing, observation quality can vary across sensors. Since past work does not specify how to allocate agents of heterogeneous but known sensitivity of these types in a stochastic bandit setting, we introduce a UCB-style algorithm, Min-Width, which aggregates information from diverse agents. In doing so, we address the joint challenges of (i) aggregating the rewards, which follow different distributions for each agent-arm pair, and (ii) coordinating the assignments of agents to arms. Min-Width facilitates efficient collaboration among heterogeneous agents, exploiting the known structure in the agents' reward functions to weight their rewards accordingly. We analyze the regret of Min-Width and conduct pseudo-synthetic and fully synthetic experiments to study the performance of different levels of information sharing. Our results confirm that the gains to modeling agent heterogeneity tend to be greater when the sensitivities are more varied across agents, while combining more information does not always improve performance.