TR2019-157

MIMO-Speech: End-to-End Multi-Channel Multi-Speaker Speech Recognition


    •  Chang, X., Zhang, W., Qian, Y., Le Roux, J., Watanabe, S., "MIMO-Speech: End-to-End Multi-Channel Multi-Speaker Speech Recognition", IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU), December 2019, pp. 237-144.
      BibTeX TR2019-157 PDF
      • @inproceedings{Chang2019dec,
      • author = {Chang, Xuankai and Zhang, Wangyou and Qian, Yanmin and Le Roux, Jonathan and Watanabe, Shinji},
      • title = {MIMO-Speech: End-to-End Multi-Channel Multi-Speaker Speech Recognition},
      • booktitle = {IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU)},
      • year = 2019,
      • pages = {237--144},
      • month = dec,
      • isbn = {978-1-7281-0305-1},
      • url = {https://www.merl.com/publications/TR2019-157}
      • }
  • MERL Contact:
  • Research Areas:

    Artificial Intelligence, Machine Learning, Speech & Audio

Abstract:

Recently, the end-to-end approach has proven its efficacy in monaural multi-speaker speech recognition. However, high word error rates (WERs) still prevent these systems from being used in practical applications. On the other hand, the spatial information in multi-channel signals has proven helpful in far-field speech recognition tasks. In this work, we propose a novel neural sequence-tosequence (seq2seq) architecture, MIMO-Speech, which extends the original seq2seq to deal with multi-channel input and multi-channel output so that it can fully model multi-channel multi-speaker speech separation and recognition. MIMO-Speech is a fully neural end-toend framework, which is optimized only via an ASR criterion. It is comprised of: 1) a monaural masking network, 2) a multi-source neural beamformer, and 3) a multi-output speech recognition model. With this processing, the input overlapped speech is directly mapped to text sequences. We further adopted a curriculum learning strategy, making the best use of the training set to improve the performance. The experiments on the spatialized wsj1-2mix corpus show that our model can achieve more than 60% WER reduction compared to the single-channel system with high quality enhanced signals (SI-SDR = 23.1 dB) obtained by the above separation function

 

  • Related News & Events

    •  AWARD    Best Paper Award at the IEEE Automatic Speech Recognition and Understanding Workshop (ASRU) 2019
      Date: December 18, 2019
      Awarded to: Xuankai Chang, Wangyou Zhang, Yanmin Qian, Jonathan Le Roux, Shinji Watanabe
      MERL Contact: Jonathan Le Roux
      Research Areas: Artificial Intelligence, Machine Learning, Speech & Audio
      Brief
      • MERL researcher Jonathan Le Roux and co-authors Xuankai Chang, Shinji Watanabe (Johns Hopkins University), Wangyou Zhang, and Yanmin Qian (Shanghai Jiao Tong University) won the Best Paper Award at the 2019 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU 2019), for the paper "MIMO-Speech: End-to-End Multi-Channel Multi-Speaker Speech Recognition". MIMO-Speech is a fully neural end-to-end framework that can transcribe the text of multiple speakers speaking simultaneously from multi-channel input. The system is comprised of a monaural masking network, a multi-source neural beamformer, and a multi-output speech recognition model, which are jointly optimized only via an automatic speech recognition (ASR) criterion. The award was received by lead author Xuankai Chang during the conference, which was held in Sentosa, Singapore from December 14-18, 2019.
    •