Do not use a protected attribute directly.
Another feature may strongly predict it.
Choose two evaluation practices.
Inspect correlations and predictability for proxy features, and use protected attributes under proper controls to compare group outcomes.
Detailed explanation
Indirect disparate impact can be found.
Indirect disparate impact can be found.
Fairness measurement remains possible.
Fairness measurement remains possible.
Proxy influence can remain.
Proxy influence can remain.
Averages can hide disparate impact.
Averages can hide disparate impact.
Try it yourself
An example you can run in a temporary verification environment.
AWS公式AIF-C01 Domain 4.1の公平性、代理属性、グループ評価を確認する。Expected result
直接・間接の属性影響を評価し、全体平均だけに依存しない。Key points
- Proxy
- Correlation
- Group metrics
Notes
- Environment: AWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
- Command output formatting can vary slightly by distribution or tool version.
- Run the example in a temporary directory or process when possible.
Foundation review
Read the scope first
Check whether the command acts on the current shell, a new process, an existing process, or a file.
Verify the observable result
Use the supplied command and compare the output with the expected result.