The Sample Complexity of Online Strategic Decision Making with Information Asymmetry and Knowledge Transportability

Information asymmetry is a pervasive feature of multi-agent systems, especially evident in economics and social sciences. In these settings, agents tailor their actions based on private information to maximize their rewards. These strategic behaviors often introduce complexities due to confounding variables. Simultaneously, knowledge transportability poses another significant challenge, arising from the difficulties of conducting experiments in target environments. It requires transferring knowledge from environments where empirical data is more readily available. Against these backdrops, this paper explores a fundamental question in online learning: Can we employ non-i.i.d. actions to learn about confounders even when requiring knowledge transfer? We present a sample-efficient algorithm designed to accurately identify system dynamics under information asymmetry and to navigate the challenges of knowledge transfer effectively in reinforcement learning, framed within an online strategic interaction model. Our method provably achieves learning of an $\epsilon$-optimal policy with a tight sample complexity of $\tilde{O}(1/\epsilon^2)$.

In many real‑world multi‑agent settings—such as economic markets or social networks—individuals make decisions based on private information, creating “information asymmetry”. In such complicated environments, people often need to transfer knowledge from one domain to another where experiments are hard to run, a challenge known as knowledge transportability.We address these intertwined challenges by designing an online learning algorithm that deliberately uses non‑identically distributed actions to tease apart private factors, while also supporting efficient transfer of what’s learned across different environments.This work provides the first sample‑efficient learning approach in multi‑agent systems under information asymmetry and knowledge transportability. By reducing the experimental burden and improving robustness, our results open the door to better predictive models and decision‑support tools in economics, social science, and beyond.