AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Adaptive Moments are Surprisingly Effective for Plug-and-Play Diffusion Sampling

Christian Belardi, Justin Lovelace, Kilian Q Weinberger, Carla P Gomes

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Diffusion posterior sampling · 0 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 adaptive moment estimation (Adam) to guidance (ablation)DPS · ImageNet SR 16xLPIPS: 0.30 -> 0.27Improved
add adaptive moment estimation (Adam) to guidance (ablation)DPS · ImageNet inpainting 90% maskLPIPS: 0.16 -> 0.12Improved
add adaptive moment estimation (Adam) to guidance (ablation)DPS · CIFAR-10 class-conditionalaccuracy (%): 42.8 -> 52.6Improved
add adaptive moment estimation (Adam) to guidance (ablation)DPS · CIFAR-10 class-conditionalFID: 32.3 -> 58.0Hurt
add adaptive moment estimation (Adam) to guidance (ablation)DPS · ImageNet class-conditional (standard classifier)top-10 accuracy (%): 0.7 -> 10.5Improved
add adaptive moment estimation (Adam) to guidance (ablation)CG · ImageNet time-aware classifieraccuracy (%): 61.0 -> 82.5Improved
add adaptive moment estimation (Adam) to guidance (ablation)MPGD · ImageNet reconstructionLPIPS: SR16x 0.35 -> 0.35; inpainting 0.45 -> 0.48No clear effect
add adaptive moment estimation (Adam) to guidance (ablation)PiGDM · ImageNet inpainting 90% maskLPIPS: 0.30 -> 0.11Improved
remove first moment (beta1=0) or second moment (beta2=0) (ablation)AdamDPS · SR16x, deblur 12, inpainting 90% (Cats)relative LPIPS improvement over DPSImproved
increase task difficulty (SR from 4x to 16x) (ablation)TFG vs DPS · ImageNetLPIPS: TFG better than DPS at 4x (0.13 vs 0.19) but worse at 16x (0.32 vs 0.30)Hurt
use TFG (recurrence, many components) instead of DPS under sparse conditioning (ablation)DPS · ImageNet SR 16xLPIPS: 0.30 -> 0.32Hurt
reduce sampling step budget (ablation)AdamDPS vs DPS · ImageNet SR 16x (DDPM and DDIM)relative LPIPS improvement over DPSImproved

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