用泰国语文本特征实现低延迟语音交互的说话结束检测
Thai Semantic End-of-Turn Detection for Real-Time Voice Agents
- 基于泰国语句末助词等语言特征，通过小模型在字词边界做二分类判断
- 微调轻量模型可实现近实时决策，延迟远低于传统静音检测方法
- 适合需要低延迟响应的本地化语音助手开发，尤其关注泰语场景
流畅的语音对话系统需要可靠且低延迟的用户说话结束检测。传统基于音频静音的端点检测会引入数百毫秒延迟，且在停顿或语言特异性现象下失效。本文首次系统研究了面向实时语音代理的泰语纯文本说话结束检测（EOT）。对比了紧凑大模型的零样本与少样本提示、以及轻量级Transformer的监督微调方法。利用YODAS语料库的转录字幕和泰语特有的语言线索（如句末助词），将EOT建模为在词元边界上的二分类任务。实验揭示了准确率与延迟之间的清晰权衡，并提供了可公开使用的实现方案。本工作建立了泰语场景下的基准，证明微调的小模型能实现接近即时的说话结束判断，适用于本地化语音代理部署。
原文摘要 · Abstract (English)
Fluid voice-to-voice interaction requires reliable and low-latency detection of when a user has finished speaking. Traditional audio-silence end-pointers add hundreds of milliseconds of delay and fail under hesitations or language-specific phenomena. We present, to our knowledge, the first systematic study of Thai text-only end-of-turn (EOT) detection for real-time agents. We compare zero-shot and few-shot prompting of compact LLMs to supervised fine-tuning of lightweight transformers. Using transcribed subtitles from the YODAS corpus and Thai-specific linguistic cues (e.g., sentence-final particles), we formulate EOT as a binary decision over token boundaries. We report a clear accuracy-latency tradeoff and provide a public-ready implementation plan. This work establishes a Thai baseline and demonstrates that small, fine-tuned models can deliver near-instant EOT decisions suitable for on-device agents.
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。