The critical slowing down in training diffusion models
cond-mat.dis-nn, cond-mat.stat-mech, cs.AI, cs.LG, physics.comp-ph
Submitted: 2026-05-12
Updated: 2026-09-18
Comments: 17 pages, 8 figures
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Percolation and criticality of systems with competing interactions on Bethe lattices: limitations and potential strengths of cluster schemes
- Efficient Monte Carlo sampling of metastable systems using non-local collective variable updates
- Scaling Up Machine Learning For Quantum Field Theory with Equivariant Continuous Flows
- Scale-Adaptive Generative Flows for Multiscale Scientific Data
- A Review on Generative AI For Text-To-Image and Image-To-Image Generation and Implications To Scientific Images
- Controllable Video Generation: A Survey
- Annealing in variational inference mitigates mode collapse: A theoretical study on Gaussian mixtures
- Generating from Discrete Distributions Using Diffusions: Insights from Random Constraint Satisfaction Problems
- Renormalizing Diffusion Models
- Multilevel Generative Samplers for Investigating Critical Phenomena
- Generative diffusion model with inverse renormalization group flows
Related papers
- Few-Shot Neuromorphic Vision in a Nonlinear Photonic Network Laser
- Hyperbolic lattices with mass disorder: Phases and phase transitions
- Adaptive Neural Quantum States: A Recurrent Neural Network Perspective
- Machine learning Majorana topology using unsupervised and supervised learning
- Quenched fluctuation-induced force arising from polarization disorder