Beta. Content is under active construction and has not been peer-reviewed. Report errors on
GitHub
.
Disclaimer
Theorem
Path
Curriculum
Paths
Demos
Diagnostic
Search
Quiz Hub
/
Diffusion Models
Diffusion Models
4 questions
Difficulty 4-6
View topic
Intermediate
0 / 4
4 intermediate
Adapts to your performance
1 / 4
intermediate (4/10)
state theorem
Diffusion models consist of a fixed forward process that adds noise and a learned reverse process that removes noise. What is the training objective?
Hide and think first
A.
Adversarial training between the forward noising and reverse denoising processes
B.
Reconstruction loss between a learned latent code and the original image, similar to VAE training
C.
Predict the noise (or equivalently, the clean image) from a noisy input at a random timestep
D.
Maximum likelihood estimation of the data distribution directly, using gradient descent on log-probability
Submit Answer