Strategizing against Q-learners: A Control-theoretical Approach
Published in arXiv, 2024
In this paper, we explore the susceptibility of the Q-learning algorithm (a classical and widely used reinforcement learning method) to strategic manipulation of sophisticated opponents in games.