AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Simple Denoising Diffusion Language Models

Huaisheng Zhu, Zhengyu Chen, Shijie Zhou, Zhihui Xie, Yige Yuan, Shiqi Chen, Zhimeng Guo, Siyuan Xu, Hangfan Zhang, Vasant G. Honavar, Teng Xiao

ICML 2026 · Image generation › Fast diffusion sampling · approach: Masked diffusion language model decoding

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 anti-uniform sharpening regularization (negative gradient with random tokens) (ablation)SDDLM · OWT (170M)Gen PPL: 77.07 -> 45.18Improved
add anti-uniform sharpening regularization (ablation)SDDLM · LM1B (170M)Gen PPL: 173.04 -> 116.84Improved
add anti-uniform sharpening regularization (ablation)SDDLM · OWT (170M)sample entropy: 5.53 -> 5.31Hurt
replace random negatives with the noisy sequence itself (V2 vs V1) (ablation)SDDLM-V1 · LM1B (170M)Gen PPL: 116.84 -> 101.32Improved
replace random negatives with the noisy sequence itself (V2 vs V1) (ablation)SDDLM-V1 · OWT (170M)Gen PPL: 45.18 -> 50.05Hurt
use simplified denoising loss on corrupted positions instead of Duo ELBO lossSDDLM vs Duo · OWT (170M)Gen PPL: 80.43 -> 77.07No clear effect
use simplified denoising loss instead of Duo ELBO lossSDDLM vs Duo · Zero-shot PPL, Wikitext (trained on OWT)perplexity: 33.91 -> 35.02Hurt
use simplified denoising loss instead of Duo ELBO lossSDDLM vs Duo · Zero-shot PPL, Pubmed (trained on OWT)perplexity: 45.34 -> 44.75Improved
use simplified denoising loss instead of Duo ELBO lossSDDLM vs Duo · LAMBADA (1.1B, zero-shot)accuracy: 17.02 -> 26.53Improved
add anti-uniform sharpening regularization (ablation)SDDLM · LAMBADA (1.1B, zero-shot)accuracy: 26.53 -> 31.75Improved
add anti-uniform sharpening regularization (ablation)SDDLM · BoolQ (1.1B, zero-shot)accuracy: 60.34 -> 59.66No clear effect
add unsupervised classifier-free guidance with random-token unconditional branchSDDLM · LAMBADA (1.1B, zero-shot)accuracy: 26.53 -> 32.60Improved

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