VP, Head of Quantitative Research
Windhaven Investment Management
- The issues of overfitting: how can you accurately predict in an undefined universe without overfitting?
- Interpretability: Can you ever really trust a black box to make the right decisions without human understanding, and will the regulators agree?
- Examining the success of your algorithm in different market conditions: have you correctly identified the relevant factors?
- Coping with tail/extreme events: are you prepared for the unexpected?