将对话拆成最小单元，实现无需训练的全双工语音交互
Unit-Based Agent for Semi-Cascaded Full-Duplex Dialogue Systems
- 把复杂对话分解为独立处理的最小单元，按需切换
- 在HumDial数据集上取得挑战赛第二名，无需训练即可部署
- 适合想快速构建全双工系统的开发者和研究者
全双工语音交互对自然人机交互至关重要。本文提出一种将复杂对话分解为最小对话单元的框架，使系统可独立处理每个单元并预测何时转入下一单元。该框架基于多模态大语言模型实现半级联全双工对话系统，辅以语音活动检测（VAD）和文本转语音（TTS）等模块。系统采用无训练、即插即用模式。在HumDial数据集上的实验表明，该框架在人类类似语音对话系统挑战赛（第2赛道：全双工交互）测试集上排名所有团队第二。代码已开源至GitHub仓库 https://github.com/yu-haoyuan/fd-badcat。
原文摘要 · Abstract (English)
Full-duplex voice interaction is crucial for natural human computer interaction. We present a framework that decomposes complex dialogue into minimal conversational units, enabling the system to process each unit independently and predict when to transit to the next. This framework is instantiated as a semi-cascaded full-duplex dialogue system built around a multimodal large language model, supported by auxiliary modules such as voice activity detection (VAD) and text-to-speech (TTS) synthesis. The resulting system operates in a train-free, plug-and-play manner. Experiments on the HumDial dataset demonstrate the effectiveness of our framework, which ranks second among all teams on the test set of the Human-like Spoken Dialogue Systems Challenge (Track 2: Full-Duplex Interaction). Code is available at the GitHub repository https://github.com/yu-haoyuan/fd-badcat.
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。