Parallelism, critical windows, and separations among diffusion language models

原文：Parallelism, critical windows, and separations among diffusion language models

作者：Sitan Chen, Liye Wang

来源：arXiv cs.LG（机器学习）

正文

Computer Science > Machine Learning

arXiv:2609.20539v1 (cs)

[Submitted on 17 Sep 2026]

Title:Parallelism, critical windows, and separations among diffusion language models

Authors:Sitan Chen, Liye Wang

View a PDF of the paper titled Parallelism, critical windows, and separations among diffusion language models, by Sitan Chen and 1 other authors

View PDF

HTML (experimental)

Abstract:A popular selling point of diffusion large language models (dLLMs) is their capacity for parallelism: the ability to generate sequences of text far more efficiently than autoregressive models, which require one forward pass per token. Yet among the many competing paradigms for dLLMs, from masked to uniform to Gaussian diffusion, principled understanding of how these different proposals compare in parallelism remains limited. In this work, we initiate a fine-grained comparison of the capacity for parallelism among these three leading approaches and prove the following:

• Uniform and Gaussian diffusion can sample in a number of forward passes which scales with the dual total correlation of the underlying distribution, a measure of intrinsic complexity which can be much smaller than the context length. Previously, it was only known how to achieve this using masked diffusion.

• For a certain family of random empirical measures, we show that $\widetilde{\Theta}(\sqrt{d})$ forward passes are necessary and sufficient to sample using uniform or Gaussian diffusion, yet there exist approximate score oracles for which $\widetilde{\Omega}(d)$ forward passes are needed for masked diffusion. This establishes the first provable separation in parallelism between the three prevailing dLLM paradigms.

Contrary to popular intuition that masked diffusions are harder to parallelize because they must commit to token values, the latter separation instead comes from the fact that the critical windows in masked diffusion sampling are asymptotically narrower than those in uniform and Gaussian diffusion sampling.

Comments:

90 pages

Subjects:

Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS); Statistics Theory (math.ST); Machine Learning (stat.ML)

Cite as:

arXiv:2609.20539 [cs.LG]

(or

arXiv:2609.20539v1 [cs.LG] for this version)

https://doi.org/10.48550/arXiv.2609.20539

Focus to learn more

arXiv-issued DOI via DataCite (pending registration)

主题

机器学习 · cs.DS · math.ST · 统计机器学习

由「前沿雷达」于 2026-09-20 采集。正文取自原文页面，已保留出处链接。标题与正文版权归原作者所有。