ER-RAMMACH, M. Ilyes et MIKAEL, M. Joseph et NGUYEN, M. Pascal (2021) Automated research of neural networks and application to stochastic control problem PFE - Project Graduation, ENSTA.
This is the latest version of this item.
![[img]](/style/images/fileicons/application_pdf.png)
| PDF 1653Kb |
Abstract
AutoML is the attempt to automate the tedious task of developing an end-to-end machine learning model for a specific task. In the literature, many techniques have been developed, but most of its applications only concern vision task or natural language processing. Applications on stoschastic control problem sush a risk hedging have not been reported to our knowledge. Among the numerous AutoML techniques, we have adapted one of them based on gradient descent for the Black-Scholes options hedging problem. In this report, we study its performances and also propose some original attepts or some methods based on previous works to improve its stability.
| Item Type: | Thesis (PFE - Project Graduation) |
|---|---|
| Subjects: | Mathematics and Applications |
| ID Code: | 8941 |
| Deposited By: | Ilyes ER-RAMMACH |
| Deposited On: | 19 oct. 2021 16:18 |
| Dernière modification: | 19 oct. 2021 16:18 |
Available Versions of this Item
- Automated research of neural networks and application to stochastic control problem (deposited 19 oct. 2021 16:18) [Currently Displayed]
Repository Staff Only: item control page
