AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Provable Separations between Memorization and Generalization in Diffusion Models

Zeqi Ye, Qijie Zhu, Molei Tao, Minshuo Chen

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Diffusion theory and design space · 8 citations (OpenAlex)

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
add importance-based attention-head pruning (20%) plus fine-tuningtrained DiT · CIFAR-10 (5000-sample subset)memorization ratio (%): 73.82 -> 68.58Improved
add importance-based head pruningtrained DiT · CIFAR-10 (5000-sample subset)recall: 0.08 -> 0.12Improved
add importance-based head pruningtrained DiT · CIFAR-10 (5000-sample subset)FID: 15.47 -> 15.07Improved
add importance-based head pruningtrained DiT · CIFAR-10 (5000-sample subset)precision: 0.39 -> 0.33Hurt
use importance-based pruning instead of random pruning (ablation)pruned DiT · CIFAR-10 (5000-sample subset)FID: 17.14 -> 15.07Improved
use importance-based pruning instead of random pruning (ablation)pruned DiT · CIFAR-10 (5000-sample subset)recall: 0.09 -> 0.12Improved
use importance-based pruning instead of random pruning (ablation)pruned DiT · CIFAR-10 (5000-sample subset)memorization ratio (%): 66.87 -> 68.58Hurt
increase training sample size (ablation)diffusion model on Gaussian mixture · Gaussian mixture (d=32)memorization ratioImproved
increase data dimension (ablation)diffusion model on Gaussian mixture · Gaussian mixture, 12M-parameter networkmemorization ratioImproved
add strong weight decay with ample data (ablation)diffusion model on Gaussian mixture · Gaussian mixture (n=10K)mean log-likelihoodHurt
add proper weight decay and moderate width with few samples (ablation)diffusion model on Gaussian mixture · Gaussian mixture (n=3.2K)memorization ratioImproved

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