Event Agenda

The training course agenda

Day 1 - Wednesday 20th March 2019

09:00

Registration and refreshments

09:30

Getting started with AAD  

  • Principles of AAD and first examples
  • Applications in Finance and discussion of their computational complexity and their rationale

11:00

Morning break

11:30

Adjoints in Finance I

  • Monte Carlo and pathwise derivative method
  • Correlation Greeks and Binning

13:00

Lunch

14:00

Writing First-Order Adjoint Code by Hand – Presentation and Live-Coding

  • While our example are written in C/C++ the derivative code generation rules and underlying theory are applicable to arbitrary imperative programming languages.
  • Driving first derivative models
  • First derivative (tangent and adjoint) code generation rules

15:30

Afternoon break

16:00

Writing First-Order Adjoint Code by Hand – Supervised Exercise

  • First-order adjoint version of Monte Carlo solver for SDE 
  • First-order adjoint version of finite difference solver for PDE

17:30

End of day one

Day 2 - Thursday 21th March 2019

09:00

Refreshments

09:30

Adjoints in Finance II

  • Market price and model parameter sensitivities. Calibration and the Implicit Function Theorem
  • Partial Differential Equations applications

11:00

Morning break

11:30

Adjoints in Finance III

  • Bermudan options and Least Squares Monte Carlo
  • Advanced XVA applications

13:00

Lunch

14:00

Introduction to AAD Software Tool dco/c++ - Presentation and Live Coding

  • The AAD software tool dco/c++ targets C and C++ explicitly. AAD tools for other languages exist. Their discussion is beyond the scope of this course.
  • First-order adjoints by operator and function overloading
  • Higher-order adjoints by template metaprogramming

15:30

Afternoon break

16:00

Hands-on First-Order AAD with dco/++ - Supervised Exercise

  • First-order adjoint version of Monte Carlo solver for SDE by dco/c++ 
  • First-order adjoint version of finite difference solver for PDE by dco/c++

17:30

End of course