DiGress: Discrete Denoising diffusion for graph generation.
Clément VignacIgor KrawczukAntoine SiraudinBohan WangVolkan CevherPascal FrossardPublished in: ICLR (2023)
Keyphrases
- denoising
- diffusion processes
- diffusion process
- nonlinear diffusion
- translation invariant
- image denoising
- graph representation
- image processing
- natural images
- noisy images
- total variation
- graph theoretic
- graph structure
- wavelet domain
- graph theory
- random walk
- spanning tree
- directed acyclic graph
- bipartite graph
- information propagation
- wavelet packet
- diffusion model
- information diffusion
- denoising algorithm
- graph databases
- graph model
- weighted graph
- anisotropic diffusion
- directed graph
- structured data
- neural network
- ridgelet transform
- denoising methods
- generation process
- graph partitioning
- undirected graph
- partial differential equations
- object recognition
- computer vision