arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.09394eess.AScs.CLcs.LGcs.SD
收藏

增强流媒体ASR的单调多头注意力

Enhancing Monotonic Multihead Attention for Streaming ASR

Graduate School of Informatics, Kyoto University(京都大学信息学研究生院)

Hirofumi Inaguma, Masato Mimura, Tatsuya Kawahara

摘要:

我们研究了单调多头注意力(MMA),通过扩展硬单调注意力到基于transformer的自动语音识别(ASR)的在线流媒体应用。对于流式推理,所有单调注意力(MA)头应该学习正确的对齐,因为直到所有头检测到相应的令牌边界才生成下一个令牌。然而,我们发现并不是所有的MA头都能通过简单的实现来学习对齐。为了鼓励每个头部正确学习对齐,我们提出了HeadDrop正则化,在训练过程中屏蔽了一部分头部随机对齐。此外,我们建议修剪冗余的头部,以提高边界检测的头部之间的共识,并防止延迟令牌生成所造成的这种头部。每个MA头上的分块注意力扩展到多头对应物。最后,我们提出了头部同步波束搜索解码,以保证稳定的流推理。

英文摘要:

We investigate a monotonic multihead attention (MMA) by extending hard monotonic attention to Transformer-based automatic speech recognition (ASR) for online streaming applications. For streaming inference, all monotonic attention (MA) heads should learn proper alignments because the next token is not generated until all heads detect the corresponding token boundaries. However, we found not all MA heads learn alignments with a naïve implementation. To encourage every head to learn alignments properly, we propose HeadDrop regularization by masking out a part of heads stochastically during training. Furthermore, we propose to prune redundant heads to improve consensus among heads for boundary detection and prevent delayed token generation caused by such heads. Chunkwise attention on each MA head is extended to the multihead counterpart. Finally, we propose head-synchronous beam search decoding to guarantee stable streaming inference.

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。

↑