End-to-End Multi-speaker ASR with Independent Vector Analysis
Not on the map. Why?
Abstract (arXiv)
We develop an end-to-end system for multi-channel, multi-speaker automatic speech recognition. We propose a frontend for joint source separation and dereverberation based on the independent vector analysis (IVA) paradigm. It uses the fast and stable iterative source steering algorithm together with a neural source model. The parameters from the ASR module and the neural source model are optimized jointly from the ASR loss itself. We demonstrate competitive performance with previous systems using neural beamforming frontends. First, we explore the trade-offs when using various number of channels for training and testing. Second, we demonstrate that the proposed IVA frontend performs well on noisy data, even when trained on clean mixtures only. Furthermore, it extends without retraining to the separation of more speakers, which is demonstrated on mixtures of three and four speakers.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 26 of its 51 references that are in the map. First posted on arXiv 1 Apr 2022.
Builds on
26 of 51 references.
References · 26
- 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
- 2018ESPnet: End-to-End Speech Processing Toolkitcited by 554
- 2020MLS: A Large-Scale Multilingual Dataset for Speech Researchcited by 467
- 2016Joint CTC-Attention based End-to-End Speech Recognition using Multi-task Learningcited by 254
- 2019A Comparative Study on Transformer vs RNN in Speech Applicationscited by 210
- 2017Pyroomacoustics: A Python package for audio room simulations and array processing algorithmscited by 131
- 2021The People's Speech: A Large-Scale Diverse English Speech Recognition Dataset for Commercial Usagecited by 98
- 2017Multi-talker Speech Separation with Utterance-level Permutation Invariant Training of Deep Recurrent Neural Networkscited by 97
- 2018The fifth 'CHiME' Speech Separation and Recognition Challenge: Dataset, task and baselinescited by 90
- 2018A multi-device dataset for urban acoustic scene classificationcited by 75
- 2019MIMO-SPEECH: End-to-End Multi-Channel Multi-Speaker Speech Recognitioncited by 44
- 2020End-to-End Multi-speaker Speech Recognition with Transformercited by 42
14 more
- 2020Far-Field Automatic Speech Recognitioncited by 32
- 2018End-to-end Speech Recognition with Word-based RNN Language Modelscited by 23
- 2019Unsupervised training of a deep clustering model for multichannel blind source separationcited by 14
- 2021End-to-End Dereverberation, Beamforming, and Speech Recognition with Improved Numerical Stability and Advanced Frontendcited by 12
- 2020End-to-End Far-Field Speech Recognition with Unified Dereverberation and Beamformingcited by 9
- 2018A unified convolutional beamformer for simultaneous denoising and dereverberationcited by 5
- 2020Surrogate Source Model Learning for Determined Source Separationcited by 5
- 2019Unsupervised Training for Deep Speech Source Separation with Kullback-Leibler Divergence Based Probabilistic Loss Functioncited by 3
- 2020Jointly optimal denoising, dereverberation, and source separationcited by 3
- 2021Joint Dereverberation and Separation with Iterative Source Steeringcited by 2
- 2021A Comparison and Combination of Unsupervised Blind Source Separation Techniquescited by 2
- 2020Overdetermined independent vector analysiscited by 1
- 2020MM Algorithms for Joint Independent Subspace Analysis with Application to Blind Single and Multi-Source Extractioncited by 1
- 2021Independence-based Joint Dereverberation and Separation with Neural Source Modelcited by 1
Built on by
Citing papers per year: 2023: 1.
1 paper.