1. For a particular classification problem, you are tasked with determining the best algorithm among SVM, random forest, K-nearest neighbors, and a deep neural network. Each algorithm has similar accuracy on your data. Stakeholders need a model that conveys
2. A student has grades of 76, 81, 78, 87, 75, and 72. The final test grade needed to achieve an 80% average is:
3. Which three security measures can defend ML workflows against malicious activities? (Select three.)
4. When classifying text into languages, which feature representation makes the most sense?
5. Which is the correct approach for scheduling model retraining in a weather prediction application?
6. Which tool is most suitable for creating a natural language processing application?
7. Considering that only 1% of the population in a dataset has a particular disease, which evaluation metrics are best for this model?
8. Which is a typical use case for video tracking?
9. Which algorithm is ideal for a fast, low-memory, low-processing power task if all options have similar accuracy?
10. Which assumption of linear regression is violated when using time-series data?
11. When should you use semi-supervised learning? (Select two.)
12. Which benefit is achieved by deploying a deep learning model on edge devices?
13. What is the definition of accuracy in classification?
14. Which cloud service model provides AI practitioner tools such as Jupyter notebooks?
15. What type of learning is optimal for developing dynamic pathing in self-driving cars?
16. What does R-squared represent in regression?
17. Which equation best represents an L1 norm?
18. Which is TRUE about highly interpretable models? (Select two.)
19. What is a neural network without an activation function equivalent to?
20. A confusion matrix shows 37 true positives and 8 false positives. How precise is the classifier?
21. What is the best way to handle missing continuous values in normally distributed data?
22. What does continuous monitoring of bias and variance in ML systems help with?
23. What causes emergent bias in a sales prediction model trained in winter and applied in summer?
24. What will NOT fix underfitting in a neural network?
25. When should a model be retrained in an ML pipeline?
26. What is Word2Vec?
27. Which two techniques are used to build personas in the ML development lifecycle?
28. Which of the following describes a benefit of machine learning for solving business problems?
29. Which algorithm is suitable for a human face detection model with a large number of features?
30. What should be included in a handover to end users for running a trained model? (Select three.)
31. Which text vectorization method is best for a translation machine?
32. Which of the following describes distributed artificial intelligence?
33. Which is a primary concern for a disease diagnostic tool prioritizing low false negatives?
34. What is TRUE about model evaluation and model validation in ML pipelines?
35. Which regression technique addresses collinearity among independent variables?
36. Which method helps rebalance a dataset?
37. Which type of regression is represented by y=c+b⋅x?
38. What is TRUE about SVM models?
39. Which is a Type 1 error in hypothesis testing?
40. Which approach is best for analyzing patients with similar unmet needs?
41. Which unsupervised learning model is suitable for fraud detection?
42. What is normalization?
43. Which AI technology generates fake videos?
44. When is lasso regression preferable to ridge regression?
45. Which open framework helps detect, respond to, and remediate threats in ML systems?
46. What retraining strategy is BEST for an AI system recommending New Year’s resolutions?
47. Which privacy-focused law should an AI practitioner follow when processing personal data?
48. What should you do before log-transforming dependent variable
49. What is the best approach when only a small portion of training data is labeled?
50. Which text vectorization methods are appropriate? (Select two.)
51. Which statements about the Beta value in an A/B test are accurate? (Select two.)
52. Which of the following is NOT an activation function?
53. Which AI technology can take a natural language question return precise answers?
54. Which regression methods are suitable for high-dimensional categorical data? (Select two.)
55. Which algorithm is ideal for preventing overfitting with a small dataset and many features?
56. Which method can help address imbalanced data with low true negative rates?
57. Which modeling technique evaluates potential negative impacts along vectors like severity and frequency?
58. Which metric does principal component analysis (PCA) capture?
59. What is the final stage of the data management lifecycle to implement data privacy?
60. What is the primary purpose of hyperparameter optimization?
61. Which algorithm is an example of unsupervised learning?
62. Which definition describes data completeness in the context of quality criteria?
63. What is it called when a sample excludes some members of the intended statistical population?
64. What is a concern with logit transformation of proportion data close to 0 and 1?
65. Which approach is best when a limited portion of your training data is labeled?
66. Which of the following metrics are text vectorization methods? (Select two.)
67. Which type of regression is represented by
68. Which two of the following statements about the beta value in an A/B test are accurate? (Select two.)
69. Which type of regression technique is ideal when the data includes many irrelevant features?
70. Which privacy-focused law must AI practitioners adhere to when processing personal data?
71. You and your team need to process large datasets of images as fast as possible for a machine learning task. The project will also use a modular framework with extensible code and an active developer community.
72. Which principle supports building an ML system with a Privacy by Design methodology?
73. A data scientist is tasked to extract business intelligence from primary data captured from the public. Which aspect is most critical?