An image classifier has higher error rates for one condition.
Investigate and improve it before release.
Choose two appropriate actions.
Analyze data, labels, and model causes, then measure group-level error rates again after mitigation.
Detailed explanation
Choose a mitigation based on the cause and measure its effect.
Choose a mitigation based on the cause and measure its effect.
Hiding the condition does not improve the user impact.
Hiding the condition does not improve the user impact.
Disparate errors become visible and can be tracked over time.
Disparate errors become visible and can be tracked over time.
High-impact or uncertain decisions may need human oversight.
High-impact or uncertain decisions may need human oversight.
Try it yourself
An example you can run in a temporary verification environment.
AWS公式AIF-C01ガイドのDomain 4.1でバイアス検出・軽減・継続評価を確認する。Expected result
バイアスを測り、原因に対処し、改善後に再測定する流れを説明できる。Key points
- Cause analysis
- Mitigation
- Re-evaluation
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.