Deep-Optimal-Stopping Pricing American style option by estimating optimal stopping time using deep learning. The upper bound computation needs to be improve.