Week 11 Generative Models

Question 1

What are generative models?

Answer 1

Generative models create new data samples that resemble a given dataset.

Question 2

What are GANs in generative modeling?

Answer 2

GANs (Generative Adversarial Networks) consist of a generator and a discriminator.

Question 3

How do GANs work?

Answer 3

The generator creates data, while the discriminator distinguishes real from generated data.

Question 4

What do Variational Autoencoders (VAEs) do?

Answer 4

Variational autoencoders learn compact latent representations for data generation.

Question 5

What are diffusion models used for?

Answer 5

Diffusion models simulate data by gradually adding and removing noise.

Question 6

What are some applications of generative models?

Answer 6

Applications include image synthesis, style transfer, and super-resolution.

Question 7

What are Conditional GANs?

Answer 7

Conditional GANs generate data conditioned on specific inputs, like text-to-image models.

Question 8

What challenges do GANs face?

Answer 8

GANs face challenges like mode collapse and training instability.

Question 9

How does Wasserstein GAN improve GANs?

Answer 9

Wasserstein GAN improves GAN training stability by addressing instability issues.

Question 10

How are generative models applied beyond images?

Answer 10

Generative models extend to video, text, and multimodal data for diverse applications.