Mathematical Finance Colloquium: A BSDE Approach to control and games on Polish spaces with applications to continuous-time RL

Filippo Beretta, ETH Zurich [in-person] Title: A BSDE Approach to control and games on Polish spaces with applications to continuous-time RL Abstract: The exploratory, or entropy-regularized, formulation of continuous-time reinforcement learning (RL) has recently drawn significant attention in the literature. In this stream of works, we recast the exploratory formulation within a general stochastic control framework over Polish action spaces, thereby unifying classical and exploratory control problems in a single setting. While standard Shannon entropy regularization emerges as a special case…