Research map / Learning theory and algorithms / Sample complexity
The Accumulation of Score Estimation Error in Diffusion Models
Baoxiang He, Valentio Iverson, Shuai Li, Cheng Chen, Bo Jiang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| use Uniform-logSNR step schedule instead of Linear (ablation) | DDPM sampler · 4-component Gaussian mixture in R^2 | sliced Wasserstein-2 distance: 0.577 -> 0.545 | Improved |
| use Uniform-logSNR step schedule instead of Quadratic (ablation) | DDPM sampler · 4-component Gaussian mixture in R^2 | sliced Wasserstein-2 distance: 0.677 -> 0.545 | Improved |
| use Cosine schedule instead of Uniform-logSNR (ablation) | DDPM sampler · 4-component Gaussian mixture in R^2 | sliced Wasserstein-2 distance: 0.545 -> 0.736 | Hurt |
| use Quadratic schedule instead of Uniform-logSNR (lower score error) (ablation) | DDPM sampler · 4-component Gaussian mixture in R^2 | score estimation L2 error: 0.121 -> 0.113 | Improved |
| use Quadratic schedule instead of Uniform-logSNR (ablation) | DDPM sampler · 4-component Gaussian mixture in R^2 | predicted error accumulation: 1.160 -> 1.390 | Hurt |
| use Cosine schedule instead of Linear (ablation) | DDPM sampler · 4-component Gaussian mixture in R^2 | score estimation L2 error: 0.123 -> 0.175 | Hurt |
| use Cosine schedule instead of Linear (ablation) | DDPM sampler · 4-component Gaussian mixture in R^2 | predicted error accumulation: 1.210 -> 2.480 | Hurt |
Closest papers by meaning
- Assessing the quality of denoising diffusion models in Wasserstein distance: noisy score and optimal bounds (NeurIPS 2025 · similarity 0.91)
- Advancing Wasserstein Convergence Analysis of Score-Based Models: Insights from Discretization and Second-Order Acceleration (NeurIPS 2025 · similarity 0.91)
- O(d/T) Convergence Theory for Diffusion Probabilistic Models under Minimal Assumptions (ICLR 2025 · similarity 0.91)
- Minimax Optimality of Score-based Diffusion Models: Beyond the Density Lower Bound Assumptions (ICML 2024 · similarity 0.91)
- Diffusion models for Gaussian distributions: Exact solutions and Wasserstein errors (ICML 2025 · similarity 0.91)
- Towards Non-Asymptotic Convergence for Diffusion-Based Generative Models (ICLR 2024 · similarity 0.90)
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.
