Forschung

The explore–exploit dilemma describes the challenge of choosing between trying new options and exploiting known options to maximize reward. My research examines how recurrent neural networks trained with reinforcement learning resolve this trade-off, with a particular focus on the exploration mechanisms and internal dynamics that emerge. By comparing these models to human behavior, I investigate whether similar computational mechanisms govern how artificial and biological agents balance exploration and exploitation.