AutoTerm-SST: Adaptive Terminology Memory with Zero Session-Time Setup for Streaming Speech Translation
Computer Science > Computation and Language arXiv:2610.26865 (cs) [Submitted on 13 Jul 2026] Title: AutoTerm-SST: Adaptive Terminology Memory with Zero Session-Time Setup for Streaming Speech Translation Authors: Jiaxuan Luo, Siqi Ouyang, Lei Li.
计算机科学 > 计算与语言 arXiv:2610.26865 (cs) [提交于 2026 年 7 月 13 日] 标题：AutoTerm-SST：用于流式语音翻译的零会话时间设置自适应术语记忆系统，作者：Jiaxuan Luo, Siqi Ouyang, Lei Li。
Abstract: Retrieval-augmented speech translation improves rare terminology, but systems such as RASST still assume that users provide a glossary before each session. That assumption is inconvenient for live talks, while merging every domain into one growing glossary eventually places too many irrelevant terms in the same retrieval and prompt competition.
摘要：检索增强型语音翻译改善了罕见术语的翻译效果，但诸如 RASST 之类的系统仍假设用户在每次会话前提供词汇表。这种假设对于现场演讲来说很不方便，而将所有领域合并为一个不断增长的词汇表，最终会导致过多的无关术语在检索和提示竞争中占据空间。
We present AutoTerm-SST, an interactive system that removes session-time glossary upload and domain selection. It organizes a large offline terminology catalog into topic slices, matches recent streaming translation context to bilingual slice prototypes, and activates a compact multi-slice working set under fixed retrieval and prompt budgets.
我们提出了 AutoTerm-SST，这是一个交互式系统，它取消了会话期间的词汇表上传和领域选择步骤。它将大型离线术语目录组织成主题切片，将近期的流式翻译上下文与双语切片原型进行匹配，并在固定的检索和提示预算下激活紧凑的多切片工作集。
On five ACL and five ESO oncology talks, AutoTerm matches a known-domain reference in terminology accuracy (88.3% vs. 88.1%) and mWER-resegmented BLEU (45.09 vs. 45.19), while requiring no domain input. Compared with a one-million-entry merged glossary, it improves terminology accuracy by 3.9 points and BLEU by 2.30.
在五场 ACL 和五场 ESO 肿瘤学讲座中，AutoTerm 在术语准确率（88.3% 对 88.1%）和 mWER 重分段 BLEU（45.09 对 45.19）方面与已知领域参考系统持平，且无需任何领域输入。与包含一百万条目的合并词汇表相比，其术语准确率提高了 3.9 个百分点，BLEU 提高了 2.30。
Routing plus four-slice retrieval remains lightweight: the complete system runs at 0.406 RTF, and the vLLM host sustains 32 real-time sessions. The live demo exposes active slices and retrieved terminology as the topic changes, providing zero-setup terminology-aware streaming translation.
路由加上四切片检索依然保持轻量级：整个系统以 0.406 RTF 运行，vLLM 主机可支持 32 个实时会话。现场演示随着主题的变化展示了活跃切片和检索到的术语，提供了无需设置的术语感知流式翻译。