Strategic Arms with Side Communication Prevail Over Low-Regret MAB Algorithms
Artificial Intelligence
2024-09-02 v1
Abstract
In the strategic multi-armed bandit setting, when arms possess perfect information about the player's behavior, they can establish an equilibrium where: 1. they retain almost all of their value, 2. they leave the player with a substantial (linear) regret. This study illustrates that, even if complete information is not publicly available to all arms but is shared among them, it is possible to achieve a similar equilibrium. The primary challenge lies in designing a communication protocol that incentivizes the arms to communicate truthfully.
Cite
@article{arxiv.2408.17101,
title = {Strategic Arms with Side Communication Prevail Over Low-Regret MAB Algorithms},
author = {Ahmed Ben Yahmed and Clément Calauzènes and Vianney Perchet},
journal= {arXiv preprint arXiv:2408.17101},
year = {2024}
}