arXiv cs.CL· Chanhyuk Lee, Wonjung Kim, Patrick Pynadath, Jaehoon Yoo, Satoshi Hayakawa, Ruqi Zhang, Jinwoo Kim, Seunghoon Hong·· 11 小时前AI 评分38
异步流映射语言模型 aFMLM：将预测顺序引入一步并行生成
Asynchronous Flow Map Language Models
AI 导读
研究者提出异步流映射语言模型（aFMLMs），通过 tokenwise 时间调度把预测顺序编码进流映射，使模型单次前向即可生成全部 token。该框架支持预设与可学习的状态自适应调度，并给出新的流映射蒸馏方法。在数独求解与 GSM8K 数学应用题代码生成任务上，aFMLM 刷新了一步生成的最优性能。
正文 · 原文
Abstract:The order in which tokens are predicted in language modeling has a profound impact on the quality of generated sequences. Autoregressive and masked diffusion models iteratively fill in tokens and can exploit such orderings across generation steps, but how to leverage an ordering for one-step parallel language modeling is unclear. We introduce asynchronous flow map language models (aFMLMs), which incorporate prediction order into one-step modeling by learning the flow maps of asynchronous flows. These flows specify an ordering through a tokenwise time schedule that controls the relative denoising progress of tokens. We formulate a general class of asynchronous flows that admit both prescribed and learned, state-adaptive schedules, show that they define valid flow maps, and develop new methods for their flow map distillation. The resulting models can generate all tokens in a single forward pass by internalizing the ordering encoded in the flow into their hidden computations. On Sudoku solving and code generation for GSM8K math word problems, aFMLM establishes new state-of-the-art one-step performance.
| Comments: | Code available at: this https URL |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.42018 [cs.CL] |
| (or arXiv:2610.42018v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.42018 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Chanhyuk Lee [view email]
[v1]
Fri, 30 Oct 2026 16:51:40 UTC (1,014 KB)
来源：arXiv cs.CL · arxiv.org