Streaming Transformer Asr with Blockwise Synchronous Beam Search

Emiru Tsunoo, Yosuke Kashiwagi, Shinji Watanabe

研究成果: Conference contribution

4 被引用数 (Scopus)

抄録

The Transformer self-attention network has shown promising performance as an alternative to recurrent neural networks in end-to-end (E2E) automatic speech recognition (ASR) systems. However, Transformer has a drawback in that the entire input sequence is required to compute both self-attention and source-target attention. In this paper, we propose a novel blockwise synchronous beam search algorithm based on blockwise processing of encoder to perform streaming E2E Transformer ASR. In the beam search, encoded feature blocks are synchronously aligned using a block boundary detection technique, where a reliability score of each predicted hypothesis is evaluated based on the end-of-sequence and repeated tokens in the hypothesis. Evaluations of the HKUST and AISHELL-1 Mandarin, LibriSpeech English, and CSJ Japanese tasks show that the proposed streaming Transformer algorithm outperforms conventional online approaches, including monotonic chunkwise attention (MoChA), especially when using the knowledge distillation technique. An ablation study indicates that our streaming approach contributes to reducing the response time, and the repetition criterion contributes significantly in certain tasks. Our streaming ASR models achieve comparable or superior performance to batch models and other streaming-based Transformer methods in all tasks considered.

本文言語English
ホスト出版物のタイトル2021 IEEE Spoken Language Technology Workshop, SLT 2021 - Proceedings
出版社Institute of Electrical and Electronics Engineers Inc.
ページ22-29
ページ数8
ISBN(電子版)9781728170664
DOI
出版ステータスPublished - 2021 1 19
外部発表はい
イベント2021 IEEE Spoken Language Technology Workshop, SLT 2021 - Virtual, Shenzhen, China
継続期間: 2021 1 192021 1 22

出版物シリーズ

名前2021 IEEE Spoken Language Technology Workshop, SLT 2021 - Proceedings

Conference

Conference2021 IEEE Spoken Language Technology Workshop, SLT 2021
国/地域China
CityVirtual, Shenzhen
Period21/1/1921/1/22

ASJC Scopus subject areas

  • 言語学および言語
  • 言語および言語学
  • 人工知能
  • コンピュータ サイエンスの応用
  • コンピュータ ビジョンおよびパターン認識
  • ハードウェアとアーキテクチャ

フィンガープリント

「Streaming Transformer Asr with Blockwise Synchronous Beam Search」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

引用スタイル