On September 14th, 2020 Edoardo Vittori will take a talk on “Reinforcement Learning via Convex Duality” as a part of his Summer School “Deep Learning and Reinforcement Learning Summer School” (DLRLSS).

The duality principle states that optimization problems may be viewed from either of two perspectives, the primal problem or the dual problem. The solution to the dual problem provides a lower bound to the solution of the primal problem. In Reinforcement Learning (RL) most existing techniques are based on dynamic programming. In this talk, we will see how to approach RL with convex programming, by first framing RL as a convex problem and then applying the tools of duality and convex optimization.

The event will be hold online via Microsoft Teams.