GANs Generative Adversarial Networks Quiz

1. In a GAN training loop, what is the typical sequence?

Train Generator, then Discriminator, repeat

Train Discriminator, then Generator, repeat

Train both simultaneously with shared weights

Train only the Generator until convergence

2. Which metric is used to measure the distance between real and generated data distributions?

Frechet Inception Distance (FID)

Root Mean Squared Error

Accuracy Score

Precision and Recall

3. True or False: GANs require labeled data to function effectively.

True

False

Submit

5. What is the purpose of the Inception Score metric in GANs?

To measure image quality and diversity of generated samples

To count the number of training iterations

To determine the learning rate

To evaluate Discriminator accuracy only

6. Which of the following is a common challenge in GAN training?

Vanishing gradients in the Generator

Training instability

Difficulty achieving convergence

All of the above

Submit

8. What is the Nash equilibrium in GAN training?

When both networks achieve zero loss

When Generator produces realistic data and Discriminator cannot distinguish it

When the learning rate becomes zero

When all generated samples are identical

9. GANs are commonly applied to which of the following tasks? (Select all that apply)

Image synthesis

Text generation

Data augmentation

Time series forecasting

Submit

10. What does StyleGAN primarily improve upon previous GAN architectures?

Faster training time

Better control over generated image features and style

Lower memory consumption

Elimination of hyperparameter tuning

11. A GAN consists of two main neural networks. What are they called?

Encoder and Decoder

Generator and Discriminator

Classifier and Regressor

Convolutional and Recurrent

Submit

13. What is a primary advantage of Deep Convolutional GANs (DCGANs)?

They eliminate the need for labeled data

They use convolutional layers for better image generation

They guarantee convergence in all cases

They reduce computational requirements significantly

14. Which GAN variant uses Wasserstein distance as its loss metric?

StyleGAN

CycleGAN

WGAN

DCGAN

Submit

16. What is mode collapse in GANs?

When the Generator produces identical outputs regardless of input

When the Discriminator overfits to training data

When the network fails to converge

When batch normalization is not applied

17. Which loss function is commonly used in standard GAN training?

Mean Squared Error (MSE)

Binary Cross-Entropy (BCE)

Kullback-Leibler Divergence

Wasserstein Distance

18. In GAN training, what does the Generator try to minimize?

The Discriminator's accuracy

The probability that fake data is labeled as fake

The reconstruction error

The learning rate

Submit

20. What is the primary role of the Generator in a GAN?

Classify real vs. fake data

Create synthetic data from random noise

Extract features from raw data

Optimize loss functions

DataAI GANs Generative Adversarial Networks Quiz

1. In a GAN training loop, what is the typical sequence?

2.

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

2. Which metric is used to measure the distance between real and generated data distributions?

3. True or False: GANs require labeled data to function effectively.

4. Progressive GANs train the network by gradually adding ____.

5. What is the purpose of the Inception Score metric in GANs?

6. Which of the following is a common challenge in GAN training?

7. Spectral normalization in GANs helps to ____.

8. What is the Nash equilibrium in GAN training?

9. GANs are commonly applied to which of the following tasks? (Select all that apply)

10. What does StyleGAN primarily improve upon previous GAN architectures?

11. A GAN consists of two main neural networks. What are they called?

12. CycleGAN is primarily used for ____.

13. What is a primary advantage of Deep Convolutional GANs (DCGANs)?

14. Which GAN variant uses Wasserstein distance as its loss metric?

15. Conditional GANs (cGANs) differ from standard GANs by incorporating ____ information.

16. What is mode collapse in GANs?

17. Which loss function is commonly used in standard GAN training?

18. In GAN training, what does the Generator try to minimize?

19. The Discriminator's main objective is to ____.

20. What is the primary role of the Generator in a GAN?