[Google ML Study Jam] 생성형 AI(Gen AI) - Introduction to Image Generation: Quiz

1.What is the process of forward diffusion?

 

Start with a clean image and add noise iteratively 정답

 

Start with a noisy image and remove noise randomly

 

Start with a noisy image and remove noise iteratively

 

Start with a clean image and add noise randomly

 

2.What is the goal of diffusion models?

 

To generate images by treating an image as a sequence of vectors

 

To pit two neural networks against each other

 

To encode images to a compressed size, then decode back to the original size

 

To learn the latent structure of a dataset by modeling the way in which data points diffuse through the latent space 정답

 

3. What is the name of the model family that draws inspiration from physics and thermodynamics?

 

Variational autoencoders

 

Diffusion models 정답

Autoregressive models

 

Generative adversarial networks

 

4.Which process involves a model learning to remove noise from images?

 

Sampling

 

Forward diffusion

 

Reverse diffusion 정답

 

GANs

5.What are some challenges of diffusion models?

 

They can be difficult to control.

 

All of the above 정답

 

They can be computationally expensive to train.

 

They can generate images that are not realistic.