Knowledge check

Unit

In Progress

Test Your Knowledge

Evaluate your understanding of inclusive collaboration and reporting in ML teams by completing this knowledge check. This will help you identify areas where you may need to review the material.

During a project debrief, executives insist on a cost-benefit breakdown, while data scientists argue the model's precision matters most. What's the most effective inclusive strategy to resolve this issue?

A) Present the cost-benefit breakdown and precision data separately in two decks.B) Begin with cost-benefit framing, then contextualise with precision as the technical driver behind the gains.C) Prioritise the executives' needs — business decisions outweigh technical detail.D) Use only visual dashboards with key performance indicators and exclude detailed metrics.

In a cross-functional meeting, product managers dominate the conversation, while engineers grow disengaged. How can you apply inclusive collaboration strategies here?

A) Rotate facilitation roles across meetings so different voices guide the discussion.B) Create a follow-up survey where engineers can anonymously submit ideas.C) Shorten meetings so technical staff don't lose interest.D) Ask engineers to limit their input to written documentation.

Your organisation mandates anonymised résumé screening. A senior manager argues that they can spot strong candidates by names or schools. What's the best way to uphold EDI principles here?

A) Respect the manager's intuition — experience often beats policy.B) Allow the manager to use their preferred method but track outcomes separately.C) Apply anonymisation only for junior roles.D) Emphasise that anonymisation mitigates unconscious bias and strengthens long-term team diversity.

A fairness audit reveals your sentiment analysis model misclassifies comments written in non-native English. Which EDI-aligned step best addresses the issue?

A) Remove all non-native English comments from training data to avoid errors.B) Present overall accuracy to stakeholders without subgroup breakdowns.C) Accept the limitation, since the majority of users are native speakers.D) Fine-tune the model with more diverse language samples and document fairness checks.

You're asked to prepare results for both regulators and senior leadership. Which reporting approach demonstrates inclusive communication?

A) Use a single detailed technical report to satisfy both groups.B) Begin with impact-focused storytelling, then add appendices for metrics and compliance.C) Remove all compliance detail, since leadership doesn't need it.D) Create entirely different reports, written independently for each group.

Your executive summary states the following: 'Model precision improved by 8%.' The chief operating officer (COO) responds: 'Why should I care?' How do you revise?

A) Reframe: 'This improvement means 2,000 fewer customers incorrectly flagged — saving £1.5 million annually.'B) Add recall and F1 score to show broader improvements.C) Leave it unchanged and trust the COO to ask follow-up questions.D) Provide a glossary of technical terms with definitions.

A fairness policy requires subgroup testing of model outcomes. Your colleague says, 'That slows us down, and the averages look fine.' What's the best response?

A) Skip subgroup testing for speed — averages are usually enough.B) Run subgroup tests informally without documenting them.C) Apply subgroup testing after deployment.D) Stress that subgroup testing ensures equitable performance and prevents hidden harms.

Your churn model achieved strong accuracy, but adoption stalled because marketing teams didn't use it. What likely went wrong?

A) Metrics weren't benchmarked against industry standards.B) The deployment pipeline wasn't automated.C) The team failed to connect model outputs to actionable marketing strategies.D) Documentation included too much technical jargon.

When sharing results, you highlight fairness testing and accuracy. A stakeholder asks, 'How does this affect our customers?' What's missing from your storytelling?

A) More technical depth on feature engineeringB) A customer-focused story that turns metrics into real impactC) Broader use of statistical significance testingD) Additional visualisations of subgroup performance

You present your results to leadership. They acknowledge the technical success but say, 'We're not sure this aligns with company values.' What inclusive communication step did you likely miss?

A) Linking project outcomes to company values and wider impactB) Adding more ROC curves and confusion matricesC) Showing cost savings as the sole success measureD) Providing longer documentation with compliance references

Action item: Knowledge check

Complete the assessment below to verify your mastery of this unit's concepts.

In progress