A synthetic data report shows high fidelity but low utility.
What does this combination reveal about the synthetic dataset?
Fidelity ≠ utility — they measure different things:
Scenario
What it means
High fidelity, high utility
Ideal — data looks real and trains good models
High fidelity, low utility
Marginal distributions match but decision-relevant structure (e.g. class boundaries) is lost
Low fidelity, high utility
Rare — model still learns useful patterns despite imperfect statistics
Key vocabulary: class boundary preservation, rare event representation, conditional distribution fidelity, task-specific utility evaluation.
6 / 10
John from the Data Science team sent a Slack message: 'The synthetic customer data's KS statistic is consistently above 0.8 – that's pretty good, right?'. Which of the following best describes what John is likely referring to when discussing the KS statistic in this context?
The KS statistic measures the maximum difference between two cumulative distribution functions. A value above 0.8 suggests a high degree of similarity between the synthetic and real datasets, indicating strong fidelity. A common misconception is to view it as simply 'size' or 'noise'; its primary purpose is distributional comparison.
7 / 10
Sarah, during a code review comment on a PR introducing synthetic data for fraud detection, writes: 'I'm concerned the utility score is low. We need to ensure this data accurately represents fraudulent transactions in production to be effective.' What is Sarah primarily highlighting?
Utility in synthetic data evaluation refers to how well the synthetic data reflects the characteristics of the real-world data it aims to mimic. A low utility score indicates a lack of resemblance between distributions, meaning the synthetic data isn't truly useful for tasks like training models or identifying patterns that exist in production.
8 / 10
David is drafting a PR description for a new synthetic dataset used to train a model for predicting user churn. The report states: 'The fidelity score is 95%, indicating the synthetic data closely resembles the real data in terms of feature distributions.' What does a 'fidelity score' of 95% primarily signify?
A fidelity score quantifies how closely the synthetic data mirrors the statistical properties of the original real data. Specifically, it assesses similarities in feature distributions – things like the range of values for age, income, or the correlations between different variables. A high score suggests a good representation.
9 / 10
Maria, during a standup update, says: 'We're running a membership inference attack on our synthetic patient data to see if it can be used to identify individuals in the real dataset.' What is Maria describing?
Membership inference attacks (MIAs) are designed to assess whether a model can accurately predict whether a specific record was part of the original training dataset—even if that dataset is now synthetic. This directly tests the privacy risk associated with using synthetic data.
10 / 10
A synthetic dataset for credit card fraud detection has a high fidelity score (98%) – meaning its feature distributions closely match real fraud transactions. However, the utility score is low, indicating poor performance when used to train a fraud detection model. What does this combination of high fidelity and low utility primarily suggest?
High fidelity simply means the synthetic data *looks* similar to the real data. However, utility depends on whether that resemblance translates into effective performance. This combination indicates that while the synthetic dataset mimics the statistical properties of fraud, it's missing key relationships or nuances that are vital for accurately identifying fraudulent transactions – perhaps due to limitations in the synthetic generation process.
What will I learn from the "Synthetic Data Evaluation Vocabulary Exercises" exercise?
Practice English vocabulary for synthetic data evaluation: fidelity, utility, privacy metrics, TSTR evaluation, KS test, Jensen-Shannon divergence, membership inference attacks.
Is this exercise free to use?
Yes. Every exercise on CoderSlingo, including this one, is free to use with no account, sign-up, or paywall required.
How many questions are in this exercise?
This set contains 10 multiple-choice questions, each with a detailed explanation shown after you answer.
Do I need to create an account to track my progress?
No account is required. Your progress bar and score reset each time you reload the page, but you can retry the exercise as many times as you like.
Who is this Synthetic Data Vocabulary exercise for?
This exercise is built for IT professionals and non-native English speakers who need to read, write, and discuss synthetic data vocabulary topics confidently at work.
What happens if I answer a question incorrectly?
You will see the correct answer highlighted along with a detailed explanation of why it is correct -- so every wrong answer becomes a learning moment, not just a lost point.
Can I retry this exercise?
Yes -- click "Try again" on the results screen at any time to reset your score and go through all the questions again.
How long does this exercise take to complete?
Most learners finish all 10 questions in under 10 minutes, since each question is answered by clicking a single option.
Where can I find more Synthetic Data Vocabulary exercises?
See the full Synthetic Data Vocabulary exercises hub for more vocabulary drills on this topic.
Is this exercise mobile-friendly?
Yes -- the exercise works on any device with a modern browser, including phones and tablets, with no app download required.