利用音频分类改进非裔美国人英语的语音识别
Improving Speech Recognition for African American English With Audio Classification
AI总结：
本文提出用少量域外AAE数据训练音频分类器并结合地理信息筛选未转录短格式查询子集进行半监督微调，使AAE与MAE的词错误率差异相对降低38.5%且不损害MAE质量。
摘要：
自动语音识别（ASR）系统已被证明在其预期或期望识别的语言变体之间存在较大的质量差异。缓解这一问题的一种方法是使用更具代表性的数据集来训练或微调模型。但这种方法可能受到用于训练和评估的域内数据有限的阻碍。我们提出了一种新方法，利用少量域外（长格式）非裔美国人英语（AAE）数据来提高美式英语短格式语音识别器的鲁棒性。我们使用CORAAL、YouTube和Mozilla Common Voice来训练一个音频分类器，以近似输出某段话语是AAE还是其他变体（包括主流美式英语，MAE）。通过将分类器输出与粗略的地理信息相结合，我们可以从大规模未转录的短格式查询语料库中选出一个话语子集，用于大规模的半监督学习。在该数据上进行微调后，AAE与MAE之间的相对词错误率差异降低了38.5%，且未降低MAE的质量。
英文摘要：
Automatic speech recognition (ASR) systems have been shown to have large quality disparities between the language varieties they are intended or expected to recognize. One way to mitigate this is to train or fine-tune models with more representative datasets. But this approach can be hindered by limited in-domain data for training and evaluation. We propose a new way to improve the robustness of a US English short-form speech recognizer using a small amount of out-of-domain (long-form) African American English (AAE) data. We use CORAAL, YouTube and Mozilla Common Voice to train an audio classifier to approximately output whether an utterance is AAE or some other variety including Mainstream American English (MAE). By combining the classifier output with coarse geographic information, we can select a subset of utterances from a large corpus of untranscribed short-form queries for semi-supervised learning at scale. Fine-tuning on this data results in a 38.5% relative word error rate disparity reduction between AAE and MAE without reducing MAE quality.
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。
