提出无需训练的长度自适应解码方法，提升掩码扩散机器翻译质量。
Length-Adaptive Decoding for Masked Diffusion Machine Translation
- 基于预测熵山谷原理，动态选择最优目标长度
- 在中英、英德翻译上分别提升64.9%和65.3%的COMET-22得分
- 适合关注生成长度优化的扩散模型与机器翻译研究者
机器翻译测试掩码扩散语言模型（dLLMs）时，每个源词必须被准确还原，而固定画布解码需在去噪前确定目标长度。现有工作多关注词元去掩码顺序，忽视长度决策对覆盖率与冗余度的直接影响。本文提出熵谷（EV）方法，通过所有掩码前向传播的平均预测熵评分候选目标画布，选择模型最准备填充的长度。相比基于训练语料统计的基线，EV在英→中、中→英、英→德任务上分别恢复了64.9%、65.3%和33.0%的参考长度带来的COMET-22增益。诊断显示，利于去噪的长度无需等于参考长度。三位译者评估支持中英翻译的充分性提升，尤其在中→英任务上证据更强。与使用相同微调数据的LLaMA-3-8B自回归模型相比，该系统在英→中任务持平，在中→英任务领先；模拟最优长度的诊断表明，在此掩码扩散翻译设置下，决定揭示顺序的重要性低于目标长度供给。
原文摘要 · Abstract (English)
Machine translation tests masked diffusion language models (dLLMs) because every source token must be rendered faithfully, while fixed canvas decoding must choose target length before denoising. Existing masked diffusion decoding work mainly studies token unmasking order, leaving this length decision under-explored despite its direct effect on coverage and redundancy. We introduce Entropy-Valley (EV), a training-free length selector that scores candidate target canvases by mean predictive entropy from all-mask forward passes and selects the canvas the backbone is most prepared to fill. Relative to a baseline using training corpus length statistics, EV recovers 64.9%, 65.3%, and 33.0% of the COMET-22 gain from reference target lengths on En$\to$Zh, Zh$\to$En, and En$\to$De. Our diagnostics show that denoising-friendly lengths need not match reference lengths. Evaluation by three translation experts supports the En$\leftrightarrow$Zh adequacy gains, with stronger evidence on Zh$\to$En. Compared with a LLaMA-3-8B autoregressive (AR) model trained on the same fine-tuning data, the EV system ties on En$\to$Zh and leads on Zh$\to$En; an oracle-length diagnostic further shows that, in this masked diffusion MT setting, deciding which tokens to reveal first matters less than how the target length is supplied.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。