M. Chacón Falcón, T. Guy, M. Karny, D. Ríos Insua
Adversarial risk analysis (ARA) informs decision-making when facing intelligent opponents and uncertain outcomes, enabling an analyst to model beliefs about an opponent's utilities, capabilities, probabilities, and the strategic calculations that an opponent uses. ARA is applied to defend-attack games mitigating standard common knowledge and common prior assumptions in classical game settings. Yet it entails complicated modeling and computations. To mitigate this, we explore the combination of ARA and fully probabilistic designs (FPD), which supports an agent's decisions in choosing the decision distribution that is closest to achieve ideal outcomes. Specifically, we use FPD to model the opponent's decisions within an ARA setting and present how to handle simple defend-attack games within that framework. We also discuss how to merge ARA and FPD in sequential games to handle reinforcement learning problems under threats.
Palabras clave: Defend-Attack games, Adversarial Risk Analysis, Fully Probabilistic Design
Programado
Teoria de Juegos
12 de junio de 2025 19:00
MR 2