Deep reinforcement learning (DRL) has proven effective in recognizing patterns and finding solutions to problems that are difficult for humans. One problem in knot theory involves finding slice surfaces for knots with minimal genus. It is easy to find large genus slice surfaces bounded by a given knot, but in order to show that the slice genus of a knot is equal to a specified value, you must also prove that the knot does not bound a slice surface of a lesser genus. In this talk, I will outline an approach using DRL and braid notation of a knot to find small genus slice surfaces for a given knot, through a series of unknotted component addition/deletion, crossing addition/deletion, and relations in the braid group.