Responsible AI practicesFull notesSummaryRingkasanStoriesPracticeResponsible Preparation for Datasets7 exam-style questions on this lesson.Question 1 of 7Why are balanced datasets important for responsible AI?AThey make training faster and cheaperBThey remove the need for model evaluationCThey reduce the cost of storing training dataDThey represent every relevant group fairlyCheck answerQuestion 2 of 7A team wants to find imbalance in its dataset and then fix it. Which pair of tools fits, in order?AClarify to find it, Data Wrangler to fix itBGround Truth to find it, Model Cards to fix itCData Wrangler to find it, Clarify to fix itDModel Monitor to find it, A2I to fix itCheck answerQuestion 3 of 7A health model was trained mostly on data from middle-aged adults. What is the likely result?AIt will refuse to make predictionsBIt will be less accurate for younger and older peopleCIt will become fasterDIt will be equally accurate for every age groupCheck answerQuestion 4 of 7A team doesn't have enough real examples from an underrepresented group. Which curation technique helps?AEarly stopping, ending training soonerBNormalization, rescaling feature valuesCAugmentation, generating new examples of that groupDDeleting that group from the datasetCheck answerQuestion 5 of 7. Select two.Which TWO are part of data curation? (Select TWO.)APublishing the model's weightsBRegular auditing to check the data stays balanced and fairCPreprocessing: cleaning, normalization, and feature selectionDRaising the model's temperatureEChoosing a GPU instance typeCheck answer0 of 2 selectedQuestion 6 of 7A team builds a model to detect cancer in children. Which data should it primarily use?AAn equal mix of all ages to keep the data balancedBSynthetic data onlyCMostly adult data, because there is more of itDData about children, because balance depends on the intended use caseCheck answerQuestion 7 of 7What is the first step to a balanced dataset?AInclusive and diverse data collectionBFine-tuning a larger model on the existing dataCRemoving all demographic information from the dataDDeploying the model and waiting for user complaintsCheck answerResponsible Considerations to Select a Model10 exam-style questions on this lesson.Transparent and Explainable Models8 exam-style questions on this lesson.