Responsible AI practicesFull notesSummaryRingkasanStoriesPracticeTransparent and Explainable Models8 exam-style questions on this lesson.Question 1 of 8What is the difference between transparency and explainability?AThey mean exactly the same thingBTransparency answers how a model makes decisions; explainability answers why it made a particular decisionCTransparency answers why a decision was made; explainability answers how the model was builtDTransparency applies only to data; explainability applies only to hardwareCheck answerQuestion 2 of 8What is a black box model?AA model that was trained without any dataBA linear regression model with readable coefficientsCA model whose workings and decisions are hard to understandDA model stored in an encrypted, locked containerCheck answerQuestion 3 of 8Which statement about transparent models is correct?AThey can't be used in regulated fields like healthcareBThey always outperform black box models on accuracyCThey remove the need for any human oversightDThey build trust but don't always perform betterCheck answerQuestion 4 of 8. Select two.Which TWO are explainability frameworks named in the course? (Select TWO.)AROUGEBBLEUCSHAP (SHapley Additive exPlanations)DLIME (Local Interpretable Model-agnostic Explanations)ESMOTECheck answer0 of 2 selectedQuestion 5 of 8A bank tells a rejected applicant, "If your annual income had been $5,000 higher, your loan would have been approved." What kind of explanation is this?ACounterfactual explanationBModel cardCFeature importance chartDConfusion matrixCheck answerQuestion 6 of 8Which is a risk of making a model more transparent?AIt always lowers the model's accuracy sharplyBIt can expose weaknesses that attackers can exploitCIt makes the model impossible to deploy on AWSDIt removes all bias from the model automaticallyCheck answerQuestion 7 of 8A team uses SageMaker Autopilot to build a model automatically and needs explanations for its predictions. Which tool does Autopilot use for this?ASageMaker Model MonitorBAmazon A2ICSageMaker ClarifyDAWS AI Service CardsCheck answerQuestion 8 of 8. Select two.Which TWO practices add transparency or explainability to an AI system? (Select TWO.)AHuman oversight of high-stakes decisionsBRaising the temperatureCKeeping model limitations secretDRemoving all logs to protect performanceEDocumenting architecture, data sources, training, and assumptionsCheck answer0 of 2 selectedResponsible Preparation for Datasets7 exam-style questions on this lesson.Model Trade-Offs9 exam-style questions on this lesson.