Generative Model

Question 1

Question 1: What is the primary purpose of a Generative Adversarial Network (GAN)?

Answer 1

A) To improve the performance of classification algorithms.
B) To generate new data similar to the training set.
C) To reduce the dimensionality of the input data.
D) To predict future trends based on past data

ANSWER: B

Question 2

Question 2: What are the main components of a GAN? (Select all that apply)

Answer 2

A) Encoder
B) Generator
C) Discriminator
D) Decoder

ANSWER: B, C

Question 3

Question 3: How does the discriminator component of a GAN function?

Answer 3

A) It generates new data.
B) It identifies whether data is real or generated by the generator.
C) It classifies different types of real-world objects.
D) It helps in compressing data for efficient storage.

ANSWER: B

Question 4

Question 4: What role does the generator play in a GAN setup?

Answer 4

A) It learns to make real data indistinguishable from fake data.
B) It directly learns from real data to improve its accuracy.
C) It uses labeled data to learn data distributions.
D) It compresses data into a lower-dimensional space.

ANSWER: A

Question 5

Question 5: In the context of GANs, what does the training process involve? (Select all that apply)

Answer 5

A) An adversarial game between the generator and the discriminator.
B) Minimizing the error in classification tasks.
C) The generator and discriminator improving through feedback from each other.
D) The discriminator minimizing its own prediction error while maximizing the generator’s error.

ANSWER: A, C, D

Question 6

Question 6: Which applications are GANs known for? (Select all that apply)

Answer 6

A) Image denoising
B) Creating realistic images from text descriptions
C) Enhancing low-resolution images
D) Speech recognition

ANSWER: B, C

Question 7

Question 7: What is the geometric intuition behind GANs often compared to?

Answer 7

A) A cat and mouse game
B) A balancing act
C) A thief and police game
D) A teacher-student relationship

Question 8

Question 8: When does the training of a GAN typically stop?

Answer 8

A) When the discriminator can no longer differentiate real data from fake.
B) When the generator produces the maximum data possible.
C) When the discriminator achieves perfect accuracy.
D) When the dataset is fully used.

Question 9

Question 1: What is the primary objective of a generative model in the context of GANs?

Answer 9

A) To classify images as real or fake.
b) To generate realistic samples indistinguishable from real data.
c) To recognize patterns in textual data.
d) To enhance the resolution of images.

Question 10

Question 2: Which are the main

Answer 10

components of a GAN?
Encoder
Generator
Discriminator
Autoencoder

Question 11

Question 3: In GAN architecture, what role does the discriminator play?

Answer 11

Generates new data samples.
Classifies samples as real or generated.
Translates images from one domain to another.
Compresses data into a latent space.

Question 12

Question 4: How does the generator improve its performance during training?

Answer 12

A)] By correctly classifying real and fake images.
Through feedback from the discriminator on its output.
By optimizing classification accuracy.
By reducing its dimensionality of the input data.