arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1704.01985cs.SDcs.LGeess.AS
收藏

基于排列不变训练的多说话人语音识别

Recognizing Multi-talker Speech with Permutation Invariant Training

Microsoft Research(微软研究院);Shanghai Jiao Tong University(上海交通大学)

Dong Yu, Xuankai Chang, Yanmin Qian

摘要:

在本文中,我们提出了一种新的技术,直接识别多个语音流的混合语音的单通道,而不首先分离它们。我们的技术是基于自动语音识别(ASR)的排列不变训练(PIT)。在PIT-ASR中,我们为每个可能的输出目标分配计算整个话语中所有帧的平均交叉熵(CE),选择具有最小CE的一个,并为该分配进行优化。PIT-ASR强制同一扬声器的所有帧与同一输出层对齐。这种策略优雅地解决了标签排列问题和说话人跟踪问题。人工混合AMI数据的实验表明,该方法是非常有前途的。

英文摘要:

In this paper, we propose a novel technique for direct recognition of multiple speech streams given the single channel of mixed speech, without first separating them. Our technique is based on permutation invariant training (PIT) for automatic speech recognition (ASR). In PIT-ASR, we compute the average cross entropy (CE) over all frames in the whole utterance for each possible output-target assignment, pick the one with the minimum CE, and optimize for that assignment. PIT-ASR forces all the frames of the same speaker to be aligned with the same output layer. This strategy elegantly solves the label permutation problem and speaker tracing problem in one shot. Our experiments on artificially mixed AMI data showed that the proposed approach is very promising.

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。

↑