Beta. Content is under active construction and has not been peer-reviewed. Report errors on
GitHub
.
Disclaimer
Theorem
Path
Curriculum
Paths
Demos
Diagnostic
Search
Quiz Hub
/
Ethics and Fairness in ML
Ethics and Fairness in ML
3 questions
Difficulty 4-5
View topic
Intermediate
0 / 3
3 intermediate
Adapts to your performance
1 / 3
intermediate (4/10)
conceptual
Demographic parity requires that the positive classification rate is equal across protected groups:
P
(
y
^
=
1
∣
g
=
A
)
=
P
(
y
^
=
1
∣
g
=
B
)
. What is a common critique?
Hide and think first
A.
Demographic parity is incompatible with any notion of ground truth and can never be computed in practice
B.
It ignores the ground truth: if the true positive rates differ across groups, demographic parity forces equal errors at the cost of accuracy
C.
Demographic parity is equivalent to maximizing accuracy, so using both metrics is redundant
D.
Demographic parity ignores within-group variation, making it biased against majority groups
Submit Answer