score-based diffusion models
Generative models that utilize score functions to guide data generation, often leveraging Diffusion Processes to progressively denoise random samples into structured outputs, widely used in high-dimensional data generation tasks.
- Advancing Wasserstein Convergence Analysis of Score-Based Models: Insights from Discretization and Second-Order Acceleration
- Cross-fluctuation phase transitions reveal sampling dynamics in diffusion models
- FlowDAS: A Stochastic Interpolant-based Framework for Data Assimilation
- Whitened Score Diffusion: A Structured Prior for Imaging Inverse Problems