arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2003.04241eess.AScs.LGcs.SDstat.ML
收藏

用于自动语音处理的深度神经网络:从大型语料库到有限数据的调查

Deep Neural Networks for Automatic Speech Processing: A Survey from Large Corpora to Limited Data

IRIT(图卢兹计算机科学研究所);Université de Toulouse(图卢兹大学);CNRS(法国国家科学研究中心)

Vincent Roger, Jérôme Farinas, Julien Pinquier

摘要:

大多数最先进的语音系统都使用深度神经网络(DNN)。这些系统需要学习大量的数据。因此,在资源不足的语音语言/问题上学习最先进的框架是一项艰巨的任务。问题可能是语音受损的数据量有限。此外,获取更多的数据和/或专业知识既耗时又昂贵。在本文中,我们定位自己的以下语音处理任务:自动语音识别,说话人识别和情感识别。为了评估有限数据的问题,我们首先研究了最先进的自动语音识别系统,因为它代表了最困难的任务(由于每种语言的差异很大)。接下来,我们将概述需要较少数据的技术和任务。在最后一节中,我们研究了Few-Shot技术,因为我们将资源不足的语音解释为Few-Shot问题。在这个意义上,我们提出了一个概述的Few-Shot技术和观点,使用这种技术的重点语音问题在这次调查。它发生的审查技术不适合大型数据集。然而,一些有前途的结果,从文献鼓励使用这种技术的语音处理。

英文摘要:

Most state-of-the-art speech systems are using Deep Neural Networks (DNNs). Those systems require a large amount of data to be learned. Hence, learning state-of-the-art frameworks on under-resourced speech languages/problems is a difficult task. Problems could be the limited amount of data for impaired speech. Furthermore, acquiring more data and/or expertise is time-consuming and expensive. In this paper we position ourselves for the following speech processing tasks: Automatic Speech Recognition, speaker identification and emotion recognition. To assess the problem of limited data, we firstly investigate state-of-the-art Automatic Speech Recognition systems as it represents the hardest tasks (due to the large variability in each language). Next, we provide an overview of techniques and tasks requiring fewer data. In the last section we investigate few-shot techniques as we interpret under-resourced speech as a few-shot problem. In that sense we propose an overview of few-shot techniques and perspectives of using such techniques for the focused speech problems in this survey. It occurs that the reviewed techniques are not well adapted for large datasets. Nevertheless, some promising results from the literature encourage the usage of such techniques for speech processing.

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。

↑