// ✦ Diffusion, Flow Matching & Generative Models

Score-Based Models via SDEs

Unifies diffusion as SDEs; generation solves the reverse-time equation.

分数匹配与随机微分方程 2020 Research frontier extended

IN PLAIN WORDS

Learn the score (gradient of log density), view noising as a continuous SDE, and generate by solving the reverse SDE or probability-flow ODE. It connects diffusion, score matching and flow matching.

Used in: Diffusion theory and advanced samplers

TIMELINE · prerequisites → this → what builds on it

2020 Denoising Dif…2020 Score-Based Model…2022 Flow Matching2023 Consistency M…

01Original paper

Score-Based Generative Modeling through Stochastic Differential Equations

Yang Song, Jascha Sohl-Dickstein, Diederik P. Kingma, et al. · 2020

arXiv ↗ ✓ Verified · arXiv API

02Where it sits

Learn the left column first; the right column builds on this one.

Learn first

Denoising Diffusion Probabilistic Models (DDPM)

Score-Based Models via SDEs

Then you can learn

Left: prerequisites · right: builds on this · dashed: related. Full map →

03Learning resources

Official docs, explainers, courses and reference code. All links checked.

Explainers

Yang Song：Generative Modeling by Estimating Gradients ↗EN

04Where the course teaches it

U72 Diffusion, Flow Matching & Image Generation VAE/GAN, DDPM/DDIM, score SDEs, latent diffusion, CFG, DiT, flow matching and ControlNet. 14 h15 algorithms6 papers

For AI / data

Verification: verified, checked 2026-10-21.
JSON: /api/learn/algorithms.json (id: score-sde)