Viva Questions Rev 2021 -November 2023.

Artificial Intelligence and Machine Learning

Module 1: Fundamentals of AI

1. Q: What is Artificial Intelligence?

A: AI is the simulation of human intelligence in machines that can perform tasks requiring human-like thinking.

2. Q: Why is it necessary to learn AI?

A: To automate tasks, enhance decision-making, and create intelligent systems for real-world applications.

3. Q: What are the types of AI learning?

A: Supervised learning, unsupervised learning, and reinforcement learning.

4. Q: Name a few fields of AI.

A: Natural Language Processing (NLP), Computer Vision, Robotics, Expert Systems, and Machine Learning.

5. Q: What are some applications of AI?

A: Healthcare diagnostics, autonomous vehicles, recommendation systems, chatbots, and gaming.

6. Q: What are some tools used in AI?

A: Python, TensorFlow, Keras, and Weka.

Module 2: Python Basics for AI

7. Q: Why is Python popular in AI?

A: Due to its simplicity, vast libraries, and community support.

8. Q: What are Python’s key features?

A: Easy syntax, extensive libraries, portability, and support for object-oriented programming.

9. Q: Name some Python libraries used in AI.

A: NumPy, pandas, scikit-learn, TensorFlow, and Keras.

10. Q: How is Python installed?

A: By downloading from the official Python website and setting it up using an installer.

11. Q: What are Python data types?

A: int, float, string, list, tuple, dictionary, and set.

12. Q: What is a conditional statement in Python?

A: A structure like if, else, and elif used to perform different actions based on conditions.

13. Q: What is the purpose of a loop in Python?

A: To repeat a block of code until a condition is met.

14. Q: What is a Python module?

A: A file containing Python code, which can include functions, classes, and variables.

15. Q: How are regular expressions used in Python?

A: For pattern matching and text manipulation.

16. Q: What is object-oriented programming in Python?

A: A paradigm where data is represented as objects with attributes and methods.

Module 3: Machine Learning Basics

17. Q: What is Machine Learning?

A: A subset of AI that enables systems to learn from data and improve performance.
18. Q: What are the types of Machine Learning?

A: Supervised, unsupervised, and reinforcement learning.

19. Q: What is supervised learning?

A: Learning from labeled data to make predictions.

20. Q: What is unsupervised learning?

A: Learning patterns from unlabeled data.

21. Q: What is classification in ML?

A: A supervised learning task to categorize data into predefined labels.

22. Q: What is regression in ML?

A: Predicting continuous values based on input data.

23. Q: Name some supervised learning algorithms.

A: Linear Regression, Support Vector Machines (SVM), and Decision Trees.

24. Q: What is clustering?

A: An unsupervised technique to group similar data points.

25. Q: What is K-Nearest Neighbors (KNN)?

A: A supervised algorithm that classifies data based on the closest k neighbors.

26. Q: What is the role of data preprocessing in ML?

A: To clean and prepare data for better model performance.

27. Q: What is feature scaling?

A: Normalizing data to improve algorithm efficiency.

28. Q: What is the purpose of the numpy library?

A: To perform numerical computations and data preprocessing.

29. Q: What is a classifier?

A: An algorithm used to categorize data into different classes.

30. Q: What is the difference between supervised and unsupervised learning?

A: Supervised learning uses labeled data, while unsupervised learning works with unlabeled data.

Module 4: AI in Gaming

31. Q: What is the role of AI in gaming?

A: To develop intelligent game characters and automate decision-making in games.

32. Q: What is a search algorithm?

A: A technique to find solutions in problem spaces.

33. Q: Name a search algorithm used in AI.

A: Minimax algorithm.

34. Q: What is a game bot?

A: A program that plays games autonomously.

35. Q: How is Python used to develop game bots?

A: By implementing decision-making algorithms and search techniques.

36. Q: What is the “Last Coin Standing” game?

A: A game where players take coins, and the player forced to take the last coin loses.

37. Q: How is a bot developed for “Tic Tac Toe”?

A: Using algorithms like Minimax to optimize moves.

38. Q: What are combinational search algorithms?

A: Algorithms that explore all possible moves to find an optimal solution.

Miscellaneous Questions on AI/ML Concepts

39. Q: What is a Naïve Bayes classifier?

A: A probabilistic model based on Bayes’ theorem.

40. Q: What is a Decision Tree?

A: A model that splits data into branches to make decisions.

41. Q: What is Random Forest?

A: An ensemble method combining multiple decision trees.

42. Q: What is the difference between AI and Machine Learning?

A: AI is the broader field; ML is a subset of AI.

43. Q: What is TensorFlow?

A: A framework for building and deploying ML models.

44. Q: What is supervised learning’s main limitation?

A: It requires labeled data, which can be costly and time-consuming to obtain.

45. Q: What is feature engineering?

A: The process of transforming raw data into meaningful features for ML.

46. Q: What is overfitting in ML?

A: A model performs well on training data but poorly on unseen data.

47. Q: What is a learning curve in ML?

A: A graph showing model performance over training iterations.

48. Q: What is the role of scikit-learn?

A: A Python library for ML algorithms.

49. Q: Define reinforcement learning.

A: Learning through trial and error using feedback from the environment.

50. Q: What is hyperparameter tuning?

A: Optimizing the parameters of an ML model for better performance.

Advanced Python Programming for AI

51. Q: What are Python packages?

A: Collections of modules grouped together for specific tasks, like NumPy or pandas.

52. Q: How do you create a Python function?

A: Using the def keyword followed by the function name and parameters.

53. Q: What is exception handling in Python?

A: A mechanism to handle runtime errors using try, except, and finally.

54. Q: What is the difference between deepcopy and shallow copy?

A: Shallow copies copy references, while deep copies duplicate objects entirely.

55. Q: How do you use regular expressions in Python?

A: By importing the re module and using functions like search(), match(), and findall().

56. Q: What is the difference between a Python list and tuple?

A: Lists are mutable; tuples are immutable.

57. Q: How can you import a specific function from a Python module?

A: Using from module_name import function_name.

58. Q: What is the role of the __init__() method in Python classes?

A: It initializes object attributes during object creation.

59. Q: How do you create a Python package?

A: By organizing modules into a directory with an __init__.py file.

60. Q: What is a lambda function in Python?

A: An anonymous, single-expression function created using the lambda keyword.

Programming with AI Tools and Libraries

61. Q: How is NumPy used in AI?

A: For numerical computations, matrix operations, and data manipulation.

62. Q: What is the purpose of pandas in AI?

A: To manipulate and analyze structured data in DataFrames.

63. Q: What is the difference between TensorFlow and Keras?

A: TensorFlow is a framework, and Keras is a high-level API for TensorFlow.

64. Q: How does Matplotlib help in ML?

A: For data visualization, including plotting graphs and charts.

65. Q: What is PyTorch used for?

A: Building and training deep learning models.

66. Q: How is scikit-learn used in ML?

A: For implementing algorithms like regression, classification, and clustering.

67. Q: What is Weka, and how is it used?

A: A software for ML tasks, such as data preprocessing and building classifiers.

68. Q: What are data pipelines in ML?

A: Sequences of data processing steps used before feeding data into a model.

69. Q: How do you save and load models in Python?

A: Using libraries like joblib or pickle.

70. Q: What is OpenAI Gym?

A: A toolkit for developing and testing reinforcement learning algorithms.

Advanced ML Programming Concepts

71. Q: What is one-hot encoding?

A: A method to convert categorical data into binary vectors.

72. Q: How do you handle missing data in ML?

A: By imputing values, such as mean, median, or using advanced methods like k-NN.

73. Q: What is cross-validation in ML?

A: A technique to evaluate model performance by splitting data into training and testing sets.

74. Q: What is the purpose of a confusion matrix?

A: To evaluate the performance of classification algorithms.

75. Q: What is a learning rate in ML?

A: A hyperparameter that controls how much to adjust the weights during training.

76. Q: How does gradient descent work?

A: It optimizes the cost function by iteratively updating model parameters.

77. Q: What is the backpropagation algorithm?

A: A process to calculate and propagate error gradients in neural networks.

78. Q: How is L2 regularization used in ML?

A: To reduce overfitting by adding a penalty term to the cost function.

79. Q: What is early stopping in ML?

A: A technique to stop training when performance on validation data ceases to improve.

80. Q: What is the difference between bagging and boosting?

A: Bagging reduces variance, while boosting reduces bias by focusing on errors.

Building and Deploying ML Models

81. Q: What are the steps in building an ML model?

A: Data collection, preprocessing, feature engineering, model selection, training, evaluation, and deployment.

82. Q: How do you evaluate an ML model?

A: Using metrics like accuracy, precision, recall, F1 score, and ROC-AUC.

83. Q: What is the pipeline for preprocessing in Python?

A: Using tools like scikit-learn’s Pipeline class for transformations and model training.

84. Q: How is K-Fold cross-validation implemented?

A: By splitting the dataset into k subsets, training on k-1, and validating on the remaining subset.

85. Q: What is feature selection?

A: Selecting the most relevant features for model building to reduce dimensionality.

86. Q: How is Grid Search used in ML?

A: To tune hyperparameters by exhaustively searching through a parameter grid.

87. Q: What is an ensemble model?

A: A combination of multiple models to improve prediction accuracy.

88. Q: What is transfer learning?

A: Reusing a pre-trained model for a new but related task.

89. Q: How do you implement data augmentation?

A: By creating variations of training data, such as rotating or flipping images.

90. Q: What are the steps to deploy an ML model?

A: Save the model, build a REST API using Flask/Django, and host it on a cloud platform.

AI in Gaming and Problem Solving

91. Q: How do search algorithms help in AI gaming?

A: By finding optimal moves or strategies in games.

92. Q: What is the Minimax algorithm?

A: A decision-making algorithm to minimize loss in adversarial games.

93. Q: How is a bot developed for the “Last Coin Standing” game?

A: By implementing logic to force the opponent into losing positions.

94. Q: What is reinforcement learning’s role in gaming AI?

A: To train bots that learn optimal actions by interacting with the environment.

95. Q: How does the Alpha-Beta pruning algorithm work?

A: It optimizes Minimax by pruning branches that do not affect the outcome.

96. Q: How do you implement Tic Tac Toe in Python?

A: Using a game board represented by arrays and logic for valid moves and winning conditions.

97. Q: What is Monte Carlo Tree Search?

A: A heuristic algorithm to make decisions by simulating all possible outcomes.

98. Q: What is a dice-rolling simulator in Python?

A: A Python program that generates random numbers to simulate rolling a die.

99. Q: How is a number-guessing game implemented in Python?

A: Using loops and conditional statements to compare user guesses to a random target number.