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 1712.01541eess.AScs.SD
收藏

基于单序列到序列模型的多方言语音识别

Multi-Dialect Speech Recognition With A Single Sequence-To-Sequence Model

Google Inc.(谷歌公司)

Bo Li, Tara N. Sainath, Khe Chai Sim, Michiel Bacchiani, Eugene Weinstein, Patrick Nguyen, Zhifeng Chen, Yonghui Wu, Kanishka Rao

摘要:

序列到序列模型通过将典型系统的单独组件(即声学(AM)、发音(PM)和语言(LM)模型)折叠到单个神经网络中,为构建语音识别系统提供了简单而优雅的解决方案。在这项工作中,我们看一个这样的序列到序列模型,即听,参加和拼写(LAS),并探讨训练一个单一的模型,以服务于不同的英语方言,这简化了训练多方言系统的过程,而不需要单独的AM,PM和LM为每种方言的可能性。我们表明,简单地将所有方言的数据汇集到一个LAS模型中,落后于对每个方言进行微调的模型的性能。然后,我们研究将方言特定的信息纳入模型,通过在原始字形序列的末尾插入方言符号来修改训练目标,并将方言信息的1-hot表示送入模型的所有层。对七种英语方言的实验结果表明,我们提出的系统是有效的方言变化建模在一个单一的LAS模型,优于LAS模型单独训练的七种方言的3.1 ~ 16.5%相对。

英文摘要:

Sequence-to-sequence models provide a simple and elegant solution for building speech recognition systems by folding separate components of a typical system, namely acoustic (AM), pronunciation (PM) and language (LM) models into a single neural network. In this work, we look at one such sequence-to-sequence model, namely listen, attend and spell (LAS), and explore the possibility of training a single model to serve different English dialects, which simplifies the process of training multi-dialect systems without the need for separate AM, PM and LMs for each dialect. We show that simply pooling the data from all dialects into one LAS model falls behind the performance of a model fine-tuned on each dialect. We then look at incorporating dialect-specific information into the model, both by modifying the training targets by inserting the dialect symbol at the end of the original grapheme sequence and also feeding a 1-hot representation of the dialect information into all layers of the model. Experimental results on seven English dialects show that our proposed system is effective in modeling dialect variations within a single LAS model, outperforming a LAS model trained individually on each of the seven dialects by 3.1 ~ 16.5% relative.

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。

↑