AI Paper Atlas

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

ICML 2026 · Learning theory and algorithms › Sample complexity · approach: Density and property estimation

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 changedWhereMeasuredEffect
use Uniform-logSNR step schedule instead of Linear (ablation)DDPM sampler · 4-component Gaussian mixture in R^2sliced Wasserstein-2 distance: 0.577 -> 0.545Improved
use Uniform-logSNR step schedule instead of Quadratic (ablation)DDPM sampler · 4-component Gaussian mixture in R^2sliced Wasserstein-2 distance: 0.677 -> 0.545Improved
use Cosine schedule instead of Uniform-logSNR (ablation)DDPM sampler · 4-component Gaussian mixture in R^2sliced Wasserstein-2 distance: 0.545 -> 0.736Hurt
use Quadratic schedule instead of Uniform-logSNR (lower score error) (ablation)DDPM sampler · 4-component Gaussian mixture in R^2score estimation L2 error: 0.121 -> 0.113Improved
use Quadratic schedule instead of Uniform-logSNR (ablation)DDPM sampler · 4-component Gaussian mixture in R^2predicted error accumulation: 1.160 -> 1.390Hurt
use Cosine schedule instead of Linear (ablation)DDPM sampler · 4-component Gaussian mixture in R^2score estimation L2 error: 0.123 -> 0.175Hurt
use Cosine schedule instead of Linear (ablation)DDPM sampler · 4-component Gaussian mixture in R^2predicted error accumulation: 1.210 -> 2.480Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy