Deeply Learning Derivatives

14 Pages Posted: 9 Oct 2018 Last revised: 20 Oct 2018

Date Written: October 14, 2018


This paper uses deep learning to value derivatives. The approach is broadly applicable, and we use a call option on a basket of stocks as an example. We show that the deep learning model is accurate and very fast, capable of producing valuations a million times faster than traditional models. We develop a methodology to randomly generate appropriate training data and explore the impact of several parameters including layer width and depth, training data quality and quantity on model speed and accuracy.

Keywords: Deep Learning, Neural Networks, Monte Carlo, Basket Options, GPU, Quantitative Finance, XVA, Valuation

JEL Classification: C13, C15, C44, C51, C52, C63, D40, G12, G13, G15, G21, G28, K22, M40

Suggested Citation

Ferguson, Ryan and Green, Andrew David, Deeply Learning Derivatives (October 14, 2018). Available at SSRN: or

Ryan Ferguson

Riskfuel ( email )

140 Yonge Street, Suite 316
Toronto, Ontario M5C1X6


Andrew David Green (Contact Author)

Scotiabank ( email )

201 Bishopsgate
London, London EC2M 3NS
United Kingdom

Do you have negative results from your research you’d like to share?

Paper statistics

Abstract Views
PlumX Metrics