Lecture 28 Quiz
What is the problem with removing a protected feature from a model in an effort to create a fair model?
The protected feature may have been useful for classification, although unfair
The protected feature may be correlated with other features, which still induce bias in the model
It will make the model less accurate
It will reduce the size of the dataset
Equalized odds tries to ensure model fairness by stating that
The probability of an outcome is the same regardless of the value of the protected attribute
The probability of an outcome is the same when conditioned on both the value of the protected attribute and the true outcome
According to Zico, the future of data science is one that
is dominated by deep learning
has data science as the leading trend in technology
will have data science continue being known as "data science"
will see data science being applied to numerous other fields
Never submit passwords through Google Forms.
This form was created inside of Carnegie Mellon University.
Terms of Service