Discovering Interpretable Directions in the Semantic Latent Space of Diffusion Models.
René HaasInbar Huberman-SpiegelglasRotem MulayoffStella GraßhofSami S. BrandtTomer MichaeliPublished in: FG (2024)
Keyphrases
- diffusion models
- latent space
- diffusion model
- information diffusion
- latent variables
- low dimensional
- social networks
- gaussian process
- lower dimensional
- generative model
- feature space
- influence maximization
- transfer learning
- high dimensional
- matrix factorization
- anisotropic diffusion
- machine learning
- computer vision
- image processing
- gaussian processes
- gaussian mixture
- nearest neighbor
- manifold learning
- parameter space
- random variables
- probabilistic model
- dimensionality reduction