IBM AI Engineering Deep Learning Fundamentals Quiz

1. Which layer type is fundamental to feature extraction in image processing?

Fully connected layer

Convolutional layer

Embedding layer

Attention layer

Submit

3. Which architecture is most commonly used for natural language processing tasks?

Convolutional Neural Network

Transformer

Autoencoder

Restricted Boltzmann Machine

Submit

5. What is the primary difference between supervised and unsupervised learning?

Supervised uses labeled data; unsupervised does not

Supervised is faster than unsupervised

Unsupervised requires more computational power

Supervised cannot handle large datasets

Submit

7. What is the purpose of weight initialization in neural networks?

Prevent neurons from becoming inactive

Ensure faster convergence and avoid dead neurons

Reduce memory usage during training

Eliminate the need for hyperparameter tuning

8. Which of the following is a common loss function for multi-class classification?

Mean Squared Error (MSE)

Cross-entropy loss

Huber loss

Hinge loss

Submit

10. What does the term 'epoch' refer to in neural network training?

One forward and backward pass through the entire dataset

A single weight update iteration

The time it takes to process one batch

A hyperparameter controlling learning rate

11. What is the primary purpose of an activation function in a neural network?

Introduce non-linearity to enable learning of complex patterns

Reduce computational cost during backpropagation

Normalize input data before processing

Initialize network weights randomly

Submit

13. What is the key advantage of LSTM (Long Short-Term Memory) over standard RNNs?

Faster computation on GPU hardware

Better ability to capture long-term dependencies

Requires less training data

Simpler architecture with fewer parameters

14. In recurrent neural networks (RNNs), what is the vanishing gradient problem?

Gradients become extremely small during backpropagation through time

The network loses all learned information

Training data becomes corrupted over time

Memory requirements exceed available resources

Submit

16. What is overfitting in the context of deep learning?

Model performs well on training data but poorly on unseen data

Model fails to converge during training

Model uses too few parameters to learn patterns

Model requires excessive computational resources

17. Which optimization algorithm adapts learning rates individually for each parameter?

Stochastic Gradient Descent (SGD)

Adam optimizer

Momentum

Batch Gradient Descent

18. In convolutional neural networks (CNNs), what is the primary role of pooling layers?

Increase the number of feature maps

Reduce spatial dimensions and retain important features

Apply non-linear transformations to inputs

Normalize batch data for faster convergence

19. What does ReLU stand for in deep learning?

Recurrent Linear Unit

Rectified Linear Unit

Recursive Learning Unit

Regularized Linear Unit

20. Which of the following best describes backpropagation?

A method to distribute data across multiple GPUs

An algorithm for computing gradients and updating weights

A technique for data augmentation during training

A process for regularizing overfitting

IBM AI Engineering Deep Learning Fundamentals Quiz

1. Which layer type is fundamental to feature extraction in image processing?

2.

What first name or nickname would you like us to use?

2. Hyperparameter tuning involves adjusting ____.

3. Which architecture is most commonly used for natural language processing tasks?

4. Gradient descent is an optimization algorithm that ____.

5. What is the primary difference between supervised and unsupervised learning?

6. Attention mechanisms in transformers allow models to ____.

7. What is the purpose of weight initialization in neural networks?

8. Which of the following is a common loss function for multi-class classification?

9. Transfer learning involves ____.

10. What does the term 'epoch' refer to in neural network training?

11. What is the primary purpose of an activation function in a neural network?

12. Batch normalization helps deep networks by ____.

13. What is the key advantage of LSTM (Long Short-Term Memory) over standard RNNs?

14. In recurrent neural networks (RNNs), what is the vanishing gradient problem?

15. Dropout is a regularization technique that ____.

16. What is overfitting in the context of deep learning?

17. Which optimization algorithm adapts learning rates individually for each parameter?

18. In convolutional neural networks (CNNs), what is the primary role of pooling layers?

19. What does ReLU stand for in deep learning?

20. Which of the following best describes backpropagation?