1. When processing numerical data, what is the benefit of applying feature scaling?
2. In machine learning, what is the effect of using a small but high-quality dataset?
3. An AI-driven recommendation system for an e-commerce platform is showing biased results. What is the best approach to mitigate this bias?
4. In reinforcement learning, what is the purpose of the reward function?
5. Which of the following is a key consideration when selecting features for a machine learning model aimed at predicting customer churn?
6. Which strategy helps to ensure that the model's performance is not overly dependent on a particular data subset?
7. Which of the following is a key advantage of using reinforcement learning in AI?
8. What is the purpose of using a hardware security module (HSM) in a machine learning pipeline?
9. Which approach ensures robustness of a machine learning model in handling rare events in production?
10. How can AI developers mitigate the risk of bias in algorithms used for credit scoring?
11. How would you explain the importance of feedback loops in AI systems to a stakeholder?
12. An AI model is used to predict customer behavior for a marketing campaign. Which challenge is key, and how to mitigate it?
13. Which strategy is most effective for securing the CI/CD pipeline for ML models?
14. What is the primary purpose of implementing role-based access control (RBAC) in a machine learning pipeline?
15. Which type of learning is characterized by an agent interacting with an environment and learning from the consequences of its actions?
16. What is the primary goal of using early stopping during training?
17. What is a key ethical issue related to AI in the legal system?
18. What is the role of model explainability tools in the post-deployment phase of a machine learning model?
19. What is the main advantage of using binary encoding over one-hot encoding for categorical variables?
20. Which business case would justify implementing a discovery and diagnostic system in a pharmaceutical company?
21. Why is it important to balance data quality with data size in machine learning?
22. Which approach is most effective for securing the ML pipeline against data poisoning attacks?
23. Which of the following best describes the purpose of supervised learning?
24. What is the effect of standardizing data to mean zero and unit variance?
25. What is the primary purpose of using containers (e.g., Docker) in ML model deployment?
26. A stakeholder is interested in the ROI of an AI project. Which metric is most relevant?
27. In multi-class classification, which metric aggregates performance across classes by giving equal importance to each class?
28. Which approach addresses ethical concerns of using AI in hiring practices?
29. A stakeholder asks why the model’s predictions differ from an expert’s opinion. Best explanation?
30. Why is model interpretability crucial in regulated industries?
31. In AI model evaluation, when is cross-validation particularly necessary?
32. Which is a common method for scaling ML models in production?
33. Which technique is most commonly used to handle class imbalance?
34. What is the role of a model registry in ML deployment?
35. Which is a common consequence of high variance in a model?
36. When dealing with imbalanced numerical data, which technique can help balance the dataset?
37. Which method is most effective in mitigating the risk of overfitting due to training on biased data?
38. Which approach can help mitigate ethical risks in AI-based hiring?
39. What is the main difference between supervised and unsupervised learning?
40. Why apply a square-root transformation to a numerical feature?
41. Which strategy is most effective in preventing ethical issues in automated decision-making?
42. When using AI in healthcare, what is a major ethical consideration?
43. What is the best practice for addressing ethical implications of AI in criminal justice?
44. Which business case justifies deploying speech recognition technology in an automotive company?
45. Which metric is most effective for evaluating predictive uncertainty of a probabilistic classification model?
46. A financial institution uses ML for loan approvals. How to address bias to ensure fair outcomes?
47. Why is handling missing data crucial in numerical datasets?
48. Which method is primary for mitigating bias in a model before deployment?
49. Which is a common approach to handling data drift in deployed ML models?
50. Which method often optimizes neural network architecture by searching configurations automatically?
51. Which evaluation metric is best suited for imbalanced classification problems?
52. In which situation prefer gradient clipping during deep learning training?
53. Primary advantage of using a larger training set in ML?
54. What is a drawback of using accuracy alone as an evaluation metric?
55. Key difference between L1 and L2 regularization?
56. In AI for government policy-making, what is a key risk and how to manage it?
57. Which transformation helps model non-linear relationships more effectively?
58. Which approach best manages ethical risks of AI in healthcare?
59. Effect of high cardinality in categorical features on ML models?
60. Primary benefit of using a model inference server in ML deployment?
61. Most effective approach for addressing privacy concerns in healthcare AI?
62. How to explain a “black-box” model to a concerned stakeholder?
63. Primary reason for using feature selection in ML?
64. A retailer wants to segment customers by purchasing behavior. Which algorithm is most suitable?
65. Which approach ensures fair and unbiased predictions in a deployed ML model?
66. Most effective method for addressing ethical risks of AI in financial markets?
67. What is “model degradation” and how to detect it?
68. How can over-reliance on AI models in decision-making pose risk?
69. Main goal of feature scaling in data preprocessing?
70. In NLP, what is the purpose of tokenization?
71. Purpose of A/B testing in deploying ML models?
72. Effect of poor data labeling on ML models, regardless of data size?
73. Which type of neural network is primarily used for sequential data?
74. An appropriate split ratio for training/validation/test sets?
75. Main challenge with textual data in NLP?
76. Effective strategy for detecting data exfiltration in ML pipelines?
77. Why monitor latency of model predictions in real-time systems?
78. Which transformation technique stabilizes variance and normalizes data distribution?
79. Which technique is most effective for detecting data drift in deployed models?
80. Primary advantage of a rolling deployment strategy for ML models?