applications in finance

Question 1

3 (/4) areas of progress of math finance due to ML

Answer 1

1) speed and generality of solutions
2) extend scope of problems that could be solved
3) generalisation of financial market models to MARKET GENERATORS (data sims)
4) extended scope further due to (3), eg synthetic generation of market data

Question 3

Deep hedging ?

Strengths? Weaknesses ?

Answer 3

Unsupervised learning based approach to determine optimal hedging strategies

Optimising a performance measure eg: hedging error, risk measure, utility function

Pro: in theory model free, easy to include market imperfections (such as transaction costs) whereas would be difficult to do so analytically

Con: in practise training data usually needs to be (partially) simulated to include every price path

Question 4

BS model

Answer 4

Go through Jupyter notebook