The Role of Data in Effective AI Systems AIE Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   What does 'data variety' refer to in the context of AI systems?

   • AThe total amount of data stored in a system
   • BThe degree to which data is accurate and complete
   • CThe different formats and sources of data, such as text, images, and sensor readings
   • DThe speed at which data is generated and processed
   Reveal Solution Discussion Select an answer first
2. 2foundation · easy

   Why is labeled data essential for supervised learning?

   • AIt reduces the need for feature engineering
   • BIt provides the ground truth that the model learns to predict from input features
   • CIt ensures the model can be deployed without further validation
   • DIt automatically removes bias from the training data
   Reveal Solution Discussion Select an answer first
4. 3application · medium

   A logistics company is building a model to predict delivery delays. The dataset includes features like distance (in miles), package weight (in pounds), and weather conditions (categorical). The data scientist notices that distance and weight have very different scales. What preprocessing step is most important before training a distance-based model like k-nearest neighbors?

   • AIncrease the volume of data by adding more historical records to reduce the impact of scale differences.
   • BEncode the categorical weather feature as integers (e.g., 1, 2, 3) and leave the numerical features as-is.
   • CNormalize or standardize the numerical features so that no single feature dominates the distance calculation.
   • DRemove the weather feature because it is categorical and cannot be used in distance-based models.
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A bank is building a credit-risk model. The data contains a feature 'annual income' with some values that are clearly incorrect (e.g., negative values). The team wants to ensure the model is accurate. What should they do?

   • AIncrease the dataset size to dilute the impact of the errors.
   • BNormalize the income feature before training to fix the errors.
   • CIdentify and correct or remove the erroneous values during data cleaning.
   • DUse a model that is robust to outliers, such as a tree-based model.
   Reveal Solution Discussion Select an answer first
6. 5application · medium

   A company is building a sentiment-analysis model for product reviews. They have 10,000 reviews from their own website and 1 million reviews scraped from social media. The social media data is noisy and contains many irrelevant posts. What is the best strategy to build a robust model?

   • AUse only the 10,000 clean reviews to ensure high data quality.
   • BCombine both datasets without any preprocessing and train the model.
   • CTrain two models separately and use the one with higher accuracy on a validation set.
   • DUse the large social media dataset but apply cleaning and filtering to remove irrelevant posts.
   Reveal Solution Discussion Select an answer first