ISTQB AI Testing Training Data Quality Quiz

Submit

3. True or False: Documentation of data sources, collection methods, and known limitations is unnecessary for training datasets.

True

False

Submit

5. Which approach helps identify if training data contains hidden biases or systemic errors?

Blind testing on held-out data

Statistical analysis and fairness audits

Increasing dataset size indefinitely

Using only automated labeling

Submit

7. True or False: Synthetic data can be used to supplement real training data but cannot fully replace it in AI testing.

True

False

Submit

9. Which of the following is a best practice for ensuring training data quality?

Use only the cheapest data sources

Implement data validation checks and quality metrics

Skip annotation quality reviews

Avoid documenting data provenance

10. True or False: All outliers in training data should be automatically removed before model training.

True

False

11. What is the primary purpose of validating training data quality in AI testing?

To reduce the size of the dataset

To ensure the model learns from accurate, representative, and unbiased data

To speed up model training

To eliminate all minority classes

12. What is a key concern when using crowdsourced labeling for training data?

Increased computational cost

Variability in label quality and potential annotator bias

Reduced dataset size

Improved model accuracy

13. Which technique is used to detect inconsistencies or errors in labeled data?

Data augmentation

Inter-annotator agreement analysis

Model overfitting

Feature engineering

14. True or False: Class imbalance in training data can lead to biased model predictions toward the majority class.

True

False

Submit

16. Which metric measures the proportion of correct labels in a training dataset?

Precision

Data accuracy

Recall

F1-score

17. What does 'representativeness' mean in the context of training data?

The dataset contains only majority class samples

The dataset reflects the distribution and characteristics of the target population

The dataset is as small as possible

The dataset uses only synthetic data

18. True or False: A training dataset with perfect accuracy on one domain will always generalize well to different domains.

True

False

Submit

20. Which of the following is a common source of bias in training data?

Imbalanced class distribution

High-quality labeled examples

Diverse geographic representation

Accurate metadata

ISTQB AI Testing Training Data Quality Quiz

1. Data ____ is the process of identifying and correcting errors or anomalies in training datasets.

2.

What first name or nickname would you like us to use?

2. In AI testing, why is it important to validate that training data is free from ____ ?

3. True or False: Documentation of data sources, collection methods, and known limitations is unnecessary for training datasets.

4. What is the relationship between training data quality and model ____ ?

5. Which approach helps identify if training data contains hidden biases or systemic errors?

6. How does data ____ affect the ability of a model to perform accurately on unseen examples?

7. True or False: Synthetic data can be used to supplement real training data but cannot fully replace it in AI testing.

8. What does 'data ____ ' mean in the context of testing AI systems?

9. Which of the following is a best practice for ensuring training data quality?

10. True or False: All outliers in training data should be automatically removed before model training.

11. What is the primary purpose of validating training data quality in AI testing?

12. What is a key concern when using crowdsourced labeling for training data?

13. Which technique is used to detect inconsistencies or errors in labeled data?

14. True or False: Class imbalance in training data can lead to biased model predictions toward the majority class.

15. Missing or ____ values in training data can significantly impact model performance.

16. Which metric measures the proportion of correct labels in a training dataset?

17. What does 'representativeness' mean in the context of training data?

18. True or False: A training dataset with perfect accuracy on one domain will always generalize well to different domains.

19. Data ____ refers to the correctness and consistency of labels assigned to training examples.

20. Which of the following is a common source of bias in training data?