arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2107.04734cs.CLcs.LGeess.AS
收藏

一种自监督语音表示模型的分层分析

Layer-wise Analysis of a Self-supervised Speech Representation Model

Toyota Technological Institute at Chicago(芝加哥丰田技术学院)

Ankita Pasad, Ju-Chieh Chou, Karen Livescu

摘要:

最近提出的自监督学习方法已经成功地用于预训练语音表示模型。这些学习表征的效用已经被经验性地观察到,但是关于在预先训练的表征本身中编码的信息的类型或程度的研究并不多。开发这样的见解可以帮助理解这些模型的能力和局限性,并使研究社区能够更有效地开发其下游应用程序的使用。在这项工作中,我们开始通过使用一套分析工具,通过其中间表示向量,检查一个最近成功的预训练模型(wav2vec 2.0)来填补这一空白。我们使用典型相关性,互信息,和性能的非参数探针简单的下游任务的度量,以便(i)查询声学和语言信息内容,(ii)表征跨模型层的信息的演变,以及(iii)了解如何微调自动语音识别(ASR)的模型影响这些观察。我们的研究结果促使修改ASR的微调协议,从而在低资源环境下提高单词出错率。

英文摘要:

Recently proposed self-supervised learning approaches have been successful for pre-training speech representation models. The utility of these learned representations has been observed empirically, but not much has been studied about the type or extent of information encoded in the pre-trained representations themselves. Developing such insights can help understand the capabilities and limits of these models and enable the research community to more efficiently develop their usage for downstream applications. In this work, we begin to fill this gap by examining one recent and successful pre-trained model (wav2vec 2.0), via its intermediate representation vectors, using a suite of analysis tools. We use the metrics of canonical correlation, mutual information, and performance on simple downstream tasks with non-parametric probes, in order to (i) query for acoustic and linguistic information content, (ii) characterize the evolution of information across model layers, and (iii) understand how fine-tuning the model for automatic speech recognition (ASR) affects these observations. Our findings motivate modifying the fine-tuning protocol for ASR, which produces improved word error rates in a low-resource setting.

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。

↑