Research map / Image generation / Fast diffusion sampling
Score-based Generative Modeling through Stochastic Evolution Equations in Hilbert Spaces
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace DDPM (VP-SDE) with HDM-SDE | function generation · Quadratic | MMD test power (%), lower is better: ≥99.0 -> 4.2±0.3 | Helped |
| replace DDPM (VP-SDE) with HDM-SDE | function generation · Melbourne | MMD test power (%), lower is better: 3.3±0.2 -> 0.2±0.1 | Helped |
| replace DDPM (VP-SDE) with HDM-SDE | function generation · Gridwatch | MMD test power (%), lower is better: 16.6±1.9 -> 3.6±0.0 | Helped |
| replace white-noise NDP with trace-class Hilbert-space noise (HDM-SDE) | function generation · Quadratic | MMD test power (%), lower is better: ≥99.0 -> 4.2±0.3 | Helped |
| replace NDP with HDM-SDE | function generation · Melbourne | MMD test power (%), lower is better: 12.8±0.4 -> 0.2±0.1 | Helped |
| replace SP-SGM with HDM-SDE | function generation · Quadratic | MMD test power (%), lower is better: 5.4±0.7 -> 4.2±0.3 | Helped |
| replace SP-SGM with HDM-SDE | function generation · Gridwatch | MMD test power (%), lower is better: 4.7±0.5 -> 3.6±0.0 | Helped |
| replace trace-class Langevin diffusion model with HDM-SDE | function generation · Melbourne | MMD test power (%), lower is better: 4.0±0.3 -> 0.2±0.1 | Helped |
| use Hilbert-space-valued Wiener process noise instead of standard DDPM noise | HDM-SDE · HumanML3D motion synthesis | qualitative motion quality | Helped |
| replace IHDM discrete-time with continuous-time HDM-SPDE | image generation · CIFAR10 | clean FID: 18.96 -> 17.84 | Helped |
| replace IHDM with HDM-SPDE | image generation · LSUN-church | clean FID: 45.06 -> 29.09 | Helped |
| replace IHDM with HDM-SPDE | image generation · AFHQ | clean FID: 41.39 -> 17.94 | Helped |
Closest papers by meaning
- Generative Fractional Diffusion Models (NeurIPS 2024 · similarity 0.90)
- Generative Modeling with Phase Stochastic Bridge (ICLR 2024 · similarity 0.89)
- Towards Non-Asymptotic Convergence for Diffusion-Based Generative Models (ICLR 2024 · similarity 0.89)
- Denoising Diffusion Bridge Models (ICLR 2024 · similarity 0.89)
- On the Trajectory Regularity of ODE-based Diffusion Sampling (ICML 2024 · similarity 0.88)
- Distributional Diffusion Models with Scoring Rules (ICML 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.