从 Mask 到 Flow，
读懂扩散语言模型。

Parallelism, critical windows, and separations among diffusion language models

Sitan Chen, Liye Wang

为什么读首次对 masked、uniform 与 Gaussian 三类 dLLM 的并行采样能力给出细粒度理论比较，并证明不同范式之间可出现可证分离。

Discrete Beckmann Transport Models for One-Step Language Modeling and Reasoning

Sophia Tang, Shiyi Wang

为什么读以时间无关的离散 Beckmann flow 直接学习固定点生成映射，无需教师模型蒸馏即可实现一到少步语言生成，并将额外函数评估转化为迭代精炼。

Diffusion Language Models for Mobile Edge Agentic AI: Foundations, Applications, and Challenges

Chenqi Li, Minghui Min, Dusit Niyato, Wei Ni

为什么读从延迟、能耗、带宽和可靠性出发梳理 dLLM 在移动边缘智能体中的系统机会与开放问题。

Simplex Relaxation for Discrete Diffusion

Jinya Sakurai, Patrick Pynadath, Satoshi Hayakawa, Jaehong Yoon +3

Reducing Pretraining-Generation Mismatch in Diffusion Language Models

Xiaocheng Lu, Huabin Liu, Song Guo, Jianguo Li

Retrofitting Linear Attention into Diffusion Language Models

Jinha Kim, Younghun Roh, Jaeyeon Kim

LLaDA MoE v2: Scaling Mixture-of-Experts Diffusion Language Models

Fengqi Zhu, Shaoxuan Xu, Jingyang Ou, Zebin You +10

MDLMPE: Distribution Aware Positional Encoding for Masked Diffusion Language Models

Tong Ling, Hang Lei, Feng Xiao, Changhui Sun +4

Masked Diffusion Language Models are Strong and Steerable Text-Based World Models for Agentic RL

Darshan Deshpande

为什么读将 masked dLLM 用作可控文本世界模型，研究其为工具型智能体生成训练环境与轨迹的潜力。

dLLM: Simple Diffusion Language Modeling

Zhanhui Zhou, Lingjie Chen, Hanghang Tong, Dawn Song

DLLM Agent: See Farther, Run Faster

Huiling Zhen, Weizhe Lin, Renxi Liu, Kai Han +14

为什么读在同一智能体流程与监督条件下比较 dLLM 和 AR 骨干，连接规划质量、工具调用与端到端效率。

The Bitter Lesson of Diffusion Language Models for Agentic Workflows: A Comprehensive Reality Check

Qingyu Lu, Liang Ding, Kanjian Zhang, Jinxia Zhang +1

为什么读系统评测 dLLM 在具身规划与工具调用中的失败模式，为判断其智能体能力提供重要基线。

Agents of Diffusion: Enhancing Diffusion Language Models with Multi-Agent Reinforcement Learning for Structured Data Generation (Extended Version)

Aja Khanal, Kaushik T. Ranade, Rishabh Agrawal, Kalyan S. Basu +1

为什么读展示提示优化智能体与评审智能体如何通过语言反馈协同约束 dLLM 的结构化生成。

Continuous Diffusion Scales Competitively with Discrete Diffusion for Language

Yang, Guo, Zhang, Sahoo +4

为什么读用现代训练与评测重新审视连续扩散的规模化结论。