Deep Learning for Derivatives Pricing: From Theory to Practice

Deep Learning and its application to derivatives pricing and risk analytics draws increasing attention in the form of conference talks and a growing body of literature. This novel approach to pricing is at the verge of breaking into the mainstream and yet remains a mystery to many. In this talk Tim will first develop a foundation explaining Deep Learning before moving into its efficient application in pricing using NVIDIA accelerated computing technologies and illustrated with several increasingly demanding examples. In so doing you will see how Deep Learning can be applied and build intuition for the types of pricing problems that may be solved using this technique and how. To make the session as practical as possible operational considerations will also be discussed including model validation; production deployment and the regulatory response to this approach.