top of page

Your certificate is now private

CertificateBackground.png

Certificate of Completion

ErdosHorizontal.png

THIS ACKNOWLEDGES THAT

HAS COMPLETED THE SUMMER 2026 QUANT FINANCE BOOT CAMP

Ahmadreza Khazaeipoul

Roman Holowinsky, PhD

July 20, 2026

DIRECTOR

DATE

clear.png

TEAM

Deep Hedging

Francis White, Ahmadreza Khazaeipoul, Jude Pereira, Antareep Gogoi, Gabriella Torres Nothaft

clear.png

Our project studies deep hedging, a machine-learning approach to option hedging in which neural networks learn trading strategies directly from simulated market data. We investigate how deep hedgers perform under increasingly realistic market models, moving from Black–Scholes to Markov-switching volatility and finally to the Heston stochastic-volatility model. A central goal is to understand how deep hedging handles transaction costs, discrete rebalancing, and hidden volatility dynamics, all of which are difficult to incorporate into classical analytic hedging rules. We compare simple feedforward, recurrent, and LSTM-based architectures, and evaluate both delta-only and delta-vega hedging strategies using risk measures such as CVaR and the entropic risk measure. Our results show that deep hedging can recover classical hedges in frictionless settings and can outperform analytic benchmarks when market frictions are present.

Screen Shot 2022-06-03 at 11.31.35 AM.png
github URL

©2017-2026 by The Erdős Institute.

bottom of page