arXiv cs.AI· Amir Mann, Jonathan Fhima, Shaked Brody, Or Avitan, Asif Zamir, Shahar Tsiper, Niv Nayman·· 3 小时前AI 评分32
B3：重叠块去噪加速扩散语言模型推理的研究
Blurring the Block Boundaries: Overlapping Block Denoising for Efficient Diffusion Language Models
AI 导读
一项名为 B3（Blurring the Block Boundaries）的方法让扩散语言模型（DLM）在共享前向传递中并行精炼两个块，当前一块快就绪时即打开下一块。该方法无需训练、无需调超参数，可集成到任何自编辑块扩散模型中。在 10 个数据集上的评测显示，B3 在生成质量持平的情况下相对基础模型加速 11.5-14.5%，改进了 DLM 推理的时延-质量权衡。
正文 · 原文
Abstract:Diffusion Language Models (DLMs) have emerged as effective alternatives to autoregressive models for latency-critical tasks, enabling faster token generation. Recent advances employ block diffusion with self-editing capabilities, where each block refines tokens iteratively before advancing to the next one. Yet each block is only as fast as its slowest token, and the next block cannot start before it. We suggest Blurring the Block Boundaries (B3), a novel method with which two blocks are refined in parallel within a shared forward pass. B3 calibrates itself on the fly, opening the next block only once the model is ready to start generating it. It is a training-free and hyperparameter-free acceleration method which seamlessly integrates within any self-editing block diffusion model. Evaluated on state-of-the-art DLMs across 10 diverse datasets spanning varied tasks and difficulty levels, B3 results in on par generation quality and 11.5-14.5% speedup compared to the base model, pushing the Pareto frontier for latency-quality trade-offs in DLM inference.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| ACM classes: | I.2.7 |
| Cite as: | arXiv:2611.14387 [cs.CL] |
| (or arXiv:2611.14387v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.14387 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Niv Nayman [view email]
[v1]
Tue, 10 Nov 2026 17:40:24 UTC (768 KB)
来源：arXiv cs.AI · arxiv.org