基于SURT的说话人归属研究
On Speaker Attribution with SURT
AI总结：
本研究针对多说话人语音识别的说话人归属问题，为SURT框架添加辅助说话人分支，结合HAT风格空白因子分解与说话人前缀方法，经消融实验与AMI语料库验证实现了长短录音下的说话人归属转录。
摘要：
流式解混与识别变换器（Streaming Unmixing and Recognition Transducer，SURT）是近年流行的连续流式多说话人语音识别（ASR）框架。随着架构、目标函数与混合模拟方法的进步，已有研究证明SURT可作为高效流式方法，实现真实会议的说话人无关转录。本研究进一步拓展该框架，提出适用于短混合语音和长录音的SURT说话人归属转录方法。我们为SURT添加辅助说话人分支，并通过HAT风格的空白因子分解将其标签预测与ASR词元预测同步。为保证录音中不同话语组间相对说话人标签的一致性，我们提出“说话人前缀”方法：在每个语音块前附加前序块中识别出的高置信度说话人帧，以建立相对顺序。我们在合成LibriSpeech混合数据集上开展了大量消融实验以验证设计选择，并在AMI语料库上证明了最终模型的有效性。
英文摘要：
The Streaming Unmixing and Recognition Transducer (SURT) has recently become a popular framework for continuous, streaming, multi-talker speech recognition (ASR). With advances in architecture, objectives, and mixture simulation methods, it was demonstrated that SURT can be an efficient streaming method for speaker-agnostic transcription of real meetings. In this work, we push this framework further by proposing methods to perform speaker-attributed transcription with SURT, for both short mixtures and long recordings. We achieve this by adding an auxiliary speaker branch to SURT, and synchronizing its label prediction with ASR token prediction through HAT-style blank factorization. In order to ensure consistency in relative speaker labels across different utterance groups in a recording, we propose "speaker prefixing" -- appending each chunk with high-confidence frames of speakers identified in previous chunks, to establish the relative order. We perform extensive ablation experiments on synthetic LibriSpeech mixtures to validate our design choices, and demonstrate the efficacy of our final model on the AMI corpus.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
