arXiv cs.CL· Steffen Castle, Nils Feldhus, Christopher Ebert, Leonhard Hennig, Sebastian M\"oller·· 11 小时前AI 评分31
DELTA：用扩散语言模型实现免训练摘要字符长度控制
Training-Free Character-Length Control in Summarization with Diffusion Language Models
AI 导读
研究者提出 DELTA（Diffusion Expected Length Targeting Algorithm），一种面向掩码扩散语言模型的免训练推理时方法，在每个去噪步骤估计期望字符长度并动态增删掩码位置以逼近目标长度。在三个摘要基准上，该方法将字符长度误差较基线降低一个数量级，同时自动与人工评估中的摘要质量保持竞争力。该工作已被 EMNLP 2026 主会接收。
正文 · 原文
Abstract:Summarization with a fixed character length is important for tasks such as headline and article preview generation, where text must fit an exact space. Autoregressive language models (LMs) struggle with precise character-length control: training-based methods require task-specific fine-tuning, while zero-shot prompting strategies achieve only approximate compliance. We propose DELTA (Diffusion Expected Length Targeting Algorithm), a training-free inference-time method for masked diffusion LMs that estimates expected character length at each denoising step and dynamically inserts or removes masked positions to steer toward a target length. DELTA requires no fine-tuning and is designed for off-the-shelf masked diffusion LMs. Across three summarization benchmarks, it reduces character-length errors by an order of magnitude compared to baselines while maintaining competitive summary quality in both automatic and human evaluation.
| Comments: | Accepted at EMNLP 2026 (Main Conference) |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2611.06371 [cs.CL] |
| (or arXiv:2611.06371v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.06371 arXiv-issued DOI via DataCite (pending registration) |
|
| Journal reference: | Proceedings of EMNLP 2026, pp. 2803-2813 |
| Related DOI: | https://doi.org/10.18653/2026.emnlp-main.121
DOI(s) linking to related resources |
Submission history
From: Steffen Castle [view email]
[v1]
Wed, 4 Nov 2026 11:32:20 UTC (119 KB)
来源：arXiv cs.CL · arxiv.org