M.Tech AI/ML Machine Learning Exam
2026/2027 – Complete Practice Question
Bank with Rationales
SECTION A — MACHINE LEARNING FUNDAMENTALS
1. What is the primary objective of machine learning?
A. To manually program every possible rule
B. To enable systems to learn patterns from data and make predictions
or decisions
C. To eliminate the need for data
D. To replace computer hardware
Answer: B
Rationale: Machine learning develops models that learn useful patterns
from data rather than relying exclusively on explicitly programmed
rules.
2. Which type of learning uses labeled input-output pairs?
A. Unsupervised learning
B. Reinforcement learning
C. Supervised learning
D. Self-organizing learning
Answer: C
Rationale: Supervised learning learns a mapping from inputs to known
target outputs.
,3. Which is an example of unsupervised learning?
A. Spam classification
B. House-price prediction
C. Customer clustering
D. Disease diagnosis using labeled records
Answer: C
Rationale: Clustering discovers groups in data without predefined
target labels.
4. In reinforcement learning, an agent primarily learns through:
A. Labeled examples only
B. Rewards and penalties
C. Database normalization
D. Feature scaling
Answer: B
Rationale: Reinforcement learning optimizes behavior through
interactions with an environment and feedback in the form of rewards.
5. Which task is a regression problem?
A. Detecting spam
B. Predicting tomorrow's temperature
C. Grouping customers
D. Recognizing handwritten digits
Answer: B
Rationale: Regression predicts a continuous numerical quantity.
6. Which task is classification?
A. Predicting salary
B. Predicting rainfall amount
,C. Predicting whether an email is spam
D. Estimating house area
Answer: C
Rationale: Classification assigns observations to discrete categories.
7. What is a feature?
A. The model's final prediction
B. An input variable used by a model
C. The loss function
D. The optimizer
Answer: B
Rationale: Features represent measurable properties supplied to a
machine-learning model.
8. What is the target variable?
A. Input feature
B. Predicted output of interest
C. Learning rate
D. Training batch
Answer: B
Rationale: The target, or response variable, is what a supervised model
attempts to predict.
9. What does training a model mean?
A. Deleting the dataset
B. Estimating model parameters from training data
C. Increasing storage capacity
D. Converting all data to images
, Answer: B
Rationale: Training involves learning parameters that minimize an
objective or loss on training data.
10. What is generalization?
A. Memorizing training examples
B. Performing well on previously unseen data
C. Increasing model complexity indefinitely
D. Removing validation data
Answer: B
Rationale: Generalization is the ability of a trained model to perform
well beyond the examples used for training.
11. Overfitting occurs when a model:
A. Is too simple to learn patterns
B. Performs well on training data but poorly on unseen data
C. Has no parameters
D. Uses a validation set
Answer: B
Rationale: An overfit model captures training-specific noise or details
that do not generalize.
12. Underfitting occurs when:
A. The model is excessively complex
B. The model cannot adequately capture the underlying pattern
C. Training accuracy is extremely high
D. The dataset contains duplicate observations
Answer: B
Rationale: Underfitting generally results from insufficient model
complexity, poor features, or excessive regularization.
2026/2027 – Complete Practice Question
Bank with Rationales
SECTION A — MACHINE LEARNING FUNDAMENTALS
1. What is the primary objective of machine learning?
A. To manually program every possible rule
B. To enable systems to learn patterns from data and make predictions
or decisions
C. To eliminate the need for data
D. To replace computer hardware
Answer: B
Rationale: Machine learning develops models that learn useful patterns
from data rather than relying exclusively on explicitly programmed
rules.
2. Which type of learning uses labeled input-output pairs?
A. Unsupervised learning
B. Reinforcement learning
C. Supervised learning
D. Self-organizing learning
Answer: C
Rationale: Supervised learning learns a mapping from inputs to known
target outputs.
,3. Which is an example of unsupervised learning?
A. Spam classification
B. House-price prediction
C. Customer clustering
D. Disease diagnosis using labeled records
Answer: C
Rationale: Clustering discovers groups in data without predefined
target labels.
4. In reinforcement learning, an agent primarily learns through:
A. Labeled examples only
B. Rewards and penalties
C. Database normalization
D. Feature scaling
Answer: B
Rationale: Reinforcement learning optimizes behavior through
interactions with an environment and feedback in the form of rewards.
5. Which task is a regression problem?
A. Detecting spam
B. Predicting tomorrow's temperature
C. Grouping customers
D. Recognizing handwritten digits
Answer: B
Rationale: Regression predicts a continuous numerical quantity.
6. Which task is classification?
A. Predicting salary
B. Predicting rainfall amount
,C. Predicting whether an email is spam
D. Estimating house area
Answer: C
Rationale: Classification assigns observations to discrete categories.
7. What is a feature?
A. The model's final prediction
B. An input variable used by a model
C. The loss function
D. The optimizer
Answer: B
Rationale: Features represent measurable properties supplied to a
machine-learning model.
8. What is the target variable?
A. Input feature
B. Predicted output of interest
C. Learning rate
D. Training batch
Answer: B
Rationale: The target, or response variable, is what a supervised model
attempts to predict.
9. What does training a model mean?
A. Deleting the dataset
B. Estimating model parameters from training data
C. Increasing storage capacity
D. Converting all data to images
, Answer: B
Rationale: Training involves learning parameters that minimize an
objective or loss on training data.
10. What is generalization?
A. Memorizing training examples
B. Performing well on previously unseen data
C. Increasing model complexity indefinitely
D. Removing validation data
Answer: B
Rationale: Generalization is the ability of a trained model to perform
well beyond the examples used for training.
11. Overfitting occurs when a model:
A. Is too simple to learn patterns
B. Performs well on training data but poorly on unseen data
C. Has no parameters
D. Uses a validation set
Answer: B
Rationale: An overfit model captures training-specific noise or details
that do not generalize.
12. Underfitting occurs when:
A. The model is excessively complex
B. The model cannot adequately capture the underlying pattern
C. Training accuracy is extremely high
D. The dataset contains duplicate observations
Answer: B
Rationale: Underfitting generally results from insufficient model
complexity, poor features, or excessive regularization.
