AAI_2025_Capstone_Chronicles_Combined

Evaluating Deep Learning Model Convergence in Chess via Nash Equilibria

17 that a reinforcement-learning algorithm that utilizes the Maxent Nash to efficiently sample from the state space could prove to be fruitful. Recomputing the Maxent Nash with the onset of new models may be an adaptive way to ensure adversarial and instructive examples are being provided to models that explore the state space.

​

91