wav2vec 2.0 和 Whisper 在非裔美国英语辅音丛简化中的分层探测
Layer-wise Probing of wav2vec 2.0 and Whisper for Consonant Cluster Reduction in African American English
AI总结：
通过分层探测，研究自监督和监督语音模型如何编码非裔美国英语中的辅音丛简化现象，发现模型能区分简化与规范形式，且简化段保留底层塞音线索，表明该现象被编码为结构化渐变音系变异。
摘要：
自监督和监督语音模型越来越多地被用于研究其内部表示编码了哪些语言信息，以及编码的抽象层次。一个未被充分探索的现象是非裔美国英语（AAE）中的辅音丛简化（CCR），这是一种广泛的音系过程，也是自动语音识别（ASR）差异的一个来源。为了研究CCR如何被表示，我们使用wav2vec2-base和Whisper-small进行了说话人独立的分层探测，采用两个任务：段级简化检测和底层丛身份的段级恢复。两个模型都能以高准确率区分简化形式和规范形式。关键的是，简化段保留了其底层塞音的线索，表明CCR被编码为结构化的渐变音系变异，而非简单的段删除。这些结果证明了现代语音模型中对AAE CCR模式的结构化音系编码。
英文摘要：
Self-supervised and supervised speech models are increasingly used to investigate which linguistic information their internal representations encode, and at what level of abstraction they encode it. One underexplored phenomenon is consonant cluster reduction (CCR) in African American English (AAE), a widespread phonological process and a source of automatic speech recognition (ASR) disparity. To examine how CCR is represented, we conduct speaker-independent layer-wise probing of wav2vec2-base and Whisper-small using two tasks: segmental reduction detection and segmental restoration of underlying cluster identity. Both models distinguish reduced and canonical forms with high accuracy. Crucially, reduced segments retain cues to their underlying stops, indicating that CCR is encoded as structured gradient phonological variation rather than simple segmental deletion. These results demonstrate structured phonological encoding of AAE CCR patterns in modern speech models.
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。
