Bias and variance · Question 11 of 12

A recruitment model suggests women less often for technical positions. The training data consists of hiring decisions from the last ten years. Which causes are likely?

2 answers are correct. Hard

1. A The model has too few parameters for fairness
2. B Fairness cannot be measured technically for applications
3. C Features such as hobbies act as proxies for gender
4. D The data is too recent for a fair model
5. E Historical discrimination is contained in the labels