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 1802.06941cs.CLcs.SDeess.AS
收藏

远场语音识别的并行数据知识提取

Distilling Knowledge Using Parallel Data for Far-field Speech Recognition

Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所);University of Chinese Academy of Sciences(中国科学院大学)

Jiangyan Yi, Jianhua Tao, Zhengqi Wen, Bin Liu

摘要:

为了提高远场语音识别的性能,本文提出了利用并行数据从近距离说话模型到远场模型的知识提取方法。这种近距离交谈的模式被称为教师模式。远场模型被称为学生模型。训练学生模型以模仿教师模型的输出分布。该约束可以通过最小化学生模型和教师模型的输出分布之间的Kullback-Leibler(KL)发散来实现。AMI语料库上的实验结果表明,最好的学生模型实现了高达4.7%的绝对词出错率(WER)减少时,与传统的训练基准模型。

英文摘要:

In order to improve the performance for far-field speech recognition, this paper proposes to distill knowledge from the close-talking model to the far-field model using parallel data. The close-talking model is called the teacher model. The far-field model is called the student model. The student model is trained to imitate the output distributions of the teacher model. This constraint can be realized by minimizing the Kullback-Leibler (KL) divergence between the output distribution of the student model and the teacher model. Experimental results on AMI corpus show that the best student model achieves up to 4.7% absolute word error rate (WER) reduction when compared with the conventionally-trained baseline 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。

↑