AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Breaking the Factorization Barrier in Diffusion Language Models

Ian Li, Zilei Shao, Benjie Wang, Rose Yu, Guy Van den Broeck, Anji Liu

ICML 2026 · Image generation › Fast diffusion sampling · approach: Masked diffusion language model decoding · 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 CoDD to Low Confidence decodingLLaDA-Instruct-8B · MATH500 (256 steps)accuracy: 36.00 -> 41.00Improved
add CoDD to Entropy decoding at 64 stepsDream-Instruct-7B · GSM8K (full diffusion)accuracy: 33.97 -> 56.41Improved
add CoDD to Entropy decoding at 128 stepsDream-Instruct-7B · GSM8K (full diffusion)accuracy: 56.18 -> 67.02Improved
add CoDD to Entropy decoding at 128 stepsDream-Instruct-7B · GPQA (full diffusion)accuracy: 14.06 -> 18.08Improved
add CoDD to Low Confidence decodingLLaDA-Instruct-8B · MBPP (256 steps)accuracy: 34.80 -> 34.00Hurt
add CoDD to Margin decodingLLaDA-Instruct-8B · GSM8K (64 steps)accuracy: 40.41 -> 40.41No clear effect
add CoDD to Entropy decoding (block)Dream-Instruct-7B · GPQA (block, 64 steps)accuracy: 5.13 -> 4.91Hurt
add any-order autoregressive sampling to CoDD (ablation)CoDD (LLaDA) · MATH500 (64 steps)accuracy: 15.80 -> 17.0Improved
add CoDD to DreamDream (low-confidence) · WikiText-103 (16 steps)perplexity: 3.32 -> 2.92Improved
add CoDD to DreamDream (low-confidence) · WikiText-103 (16 steps)Rep-2: 0.783 -> 0.812Hurt
add CoDD to Dream backboneDream-Instruct-7B · inference latencyseconds per sample overhead: 4-5%No clear effect
raise PC fraction gamma from 0.3 to 0.6 (tau 0.2) (ablation)CoDD (LLaDA, Random) · GPQA (256 steps)accuracy: 21.65 -> 26.34Improved

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