Record the datasets: Which practice best meets the objective of documenting the training and testing process?
Documenting training and testing means recording datasets, procedures, metrics and results in enough detail for an independent reviewer to reproduce them.
The question
Six months after release, an external auditor asks a team to demonstrate that the model's reported test results are genuine and reproducible. The team wants documentation practices that will satisfy such requests going forward. Which practice best meets the objective of documenting the training and testing process?
Preparing for AIGP? Take the free 5-min readiness quiz →
- Keep a running list of open engineering tickets and their assignees so the team can show that reported defects were eventually triaged and closed out.Plausible because issue tracking supports accountability, but a ticket list does not capture the datasets, procedures and metrics needed to reproduce and validate results.
- Publish a short customer-facing release note summarizing the model's headline accuracy figure so buyers understand the general quality of the product.Plausible because release notes aid transparency, but a headline figure lacks the procedural and data detail an auditor needs to reproduce the outcomes.
- Archive the production inference logs from the live system so the team can show which predictions were served to users after the model went live.Plausible because inference logs support monitoring, but they record post-release operation rather than the training-and-testing evidence required to validate reported results.
- Record the datasets, test procedures, metrics and results with enough detail that an independent reviewer could reproduce and validate the outcomes. ✓Correct because documenting training and testing to validate results means capturing datasets, procedures, metrics and outcomes so an independent party can reproduce them.
The trap
Assuming issue trackers, release notes, or production logs document the training-and-testing process, when reproducibility requires datasets, procedures, metrics, and results. How to remember it
Documenting training and testing means recording datasets, procedures, metrics and results in enough detail for an independent reviewer to reproduce them.
How many of these would you get right?
One of 1581 AIGP questions on Certsqill. Take a free five-minute check and see your score per domain — not one number, but which section to open tonight.
Test your AIGP readiness — freeMore Understanding How to Govern AI Development questions
- Document the disparity as a risk: Which action best fulfills the objective of managing issues and risks during →
- A model card summarizing the model's intended use: Which artifact should they produce? →
- Track live performance and input distributions: Which practice best fulfills this objective? →
- All 426 Understanding How to Govern AI Development questions →
Part of the Certsqill AIGP question bank · Understanding How to Govern AI Development ·
Every answer, right and wrong, comes with its own explanation.
