1. What is machine learning?

o A. A type of computer

o B. A technique for teaching computers to learn from data

o C. A programming language

o D. A hardware device

o Answer: B

2. What is the primary goal of supervised learning?

o A. Minimize errors in predictions

o B. Maximize computational efficiency

o C. Predict future events

o D. Learn from unlabeled data

o Answer: A

3. Which of the following is an example of an unsupervised learning algorithm?

o A. Linear Regression

o B. K-Means Clustering

o C. Decision Trees

o D. Support Vector Machines

o Answer: B

4. What is the purpose of feature engineering in machine learning?

o A. Building better hardware

o B. Selecting the most relevant features

o C. Engineering new features using deep learning

o D. Extracting features from images

o Answer: B

5. In classification, what does the term “class label” refer to?

o A. The name of the model

o B. The output of a regression model

o C. The predicted category of an input

o D. The input features of a model

o Answer: C

6. What is cross-validation used for in machine learning?

 o A. Cross-training different models

o B. Evaluating model performance on multiple datasets

o C. Selecting hyperparameters

o D. Testing a model’s generalization ability

o Answer: B

7. What is the purpose of regularization in machine learning models?

o A. To increase model complexity

o B. To decrease model complexity

o C. To speed up model training

o D. To increase bias

o Answer: B

8. What is the role of a confusion matrix in classification?

o A. Visualizing decision boundaries

o B. Evaluating model performance

o C. Selecting hyperparameters

o D. Handling missing data

o Answer: B

9. What is the difference between precision and recall?

o A. Both measure the same thing

o B. Precision focuses on false positives, recall focuses on false negatives

o C. Precision focuses on false negatives, recall focuses on false positives

o D. Precision and recall are unrelated metrics

o Answer: B

10. What is the purpose of feature scaling in machine learning?

o A. To remove outliers from the data

o B. To standardize the range of features

o C. To increase the complexity of models

o D. To decrease the dimensionality of features

o Answer: B

1. What is the difference between supervised learning and unsupervised learning?

 o A. Supervised learning uses labeled data, while unsupervised learning uses
unlabeled data.

o B. Supervised learning focuses on regression tasks, while unsupervised learning

focuses on classification tasks.

o C. Supervised learning requires a neural network architecture, while unsupervised

learning does not.

o D. Supervised learning is only applicable to image data, while unsupervised learning

works with text data.

o Answer: A

2. What is the bias-variance tradeoff in machine learning?

o A. The balance between model complexity and model performance

o B. The tradeoff between underfitting and overfitting

o C. The choice between linear and nonlinear models

o D. The tradeoff between precision and recall

o Answer: B

3. What is the purpose of gradient descent in neural networks?

o A. To compute feature importance

o B. To optimize model parameters during training

o C. To visualize decision boundaries

o D. To handle missing data

o Answer: B

4. What is the difference between deep learning and traditional machine learning?

o A. Deep learning uses neural networks with multiple hidden layers, while traditional
ML uses linear models.

o B. Deep learning is only applicable to image data, while traditional ML works with
tabular data.

o C. Deep learning requires labeled data, while traditional ML can work with unlabeled
data.

o D. Deep learning focuses on unsupervised learning, while traditional ML focuses on

supervised learning.

o Answer: A

5. What is the purpose of dropout in neural networks?

o A. To prevent overfitting by randomly dropping neurons during training

o B. To increase the number of hidden layers

 o C. To improve model interpretability

o D. To speed up model training

o Answer: A

6. What is the primary goal of reinforcement learning?

A. Minimize errors in predictions

B. Maximize computational efficiency

C. Learn from labeled data

D. Optimize actions based on rewards

Answer: D

7. Which AI technique is commonly used for natural language processing (NLP)?

A. Decision Trees

B. Convolutional Neural Networks (CNNs)

C. Recurrent Neural Networks (RNNs)

D. K-Means Clustering

Answer: C

8. What is the purpose of transfer learning in deep learning models?

A. To transfer data between different devices

B. To transfer knowledge from one task to another

C. To transfer features from images to text

D. To transfer model weights during training

Answer: B

9. Which ML algorithm is commonly used for recommendation systems?

A. Linear Regression

B. Support Vector Machines (SVMs)

C. Collaborative Filtering

D. K-Nearest Neighbors (KNN)

Answer: C

10. What is the purpose of dimensionality reduction in ML?

A. To increase model complexity

B. To decrease model performance

C. To reduce the number of features

 D. To improve overfitting

Answer: C

11. Which AI application involves teaching computers to understand and generate human-like
text?

A. Computer vision

B. Speech recognition

C. Natural language generation

D. Reinforcement learning

Answer: C

12. What is the main challenge in deploying AI models in real-world scenarios?

A. Lack of computational power

B. Data privacy concerns

C. Model interpretability

D. Overfitting

Answer: C

13. Which ML technique is used for anomaly detection in network security?

A. Clustering

B. Regression

C. Classification

D. Isolation Forest

Answer: D

14. What is the difference between AI and ML?

A. AI is a subset of ML

B. ML is a subset of AI

C. AI and ML are the same

D. AI focuses on hardware, ML focuses on software

Answer: B

15. What is the role of ethics in AI development?

A. To maximize profits

B. To ensure fairness, transparency, and accountability

C. To prioritize speed over accuracy

D. To avoid collaboration with other industries

Answer: B