mlsanityData to 19 Oct 2026

Pseudo Labels-based Neural Speech Enhancement for the AVSR Task in the MISP-Meeting Challenge

arXiv 2505.24446 ↗First posted 30 May 2025cs.SD eess.AScited by 1

On the map: Speech and audio language models → Speech enhancement and noise suppression

Abstract (arXiv)

This paper presents our system for the MISP-Meeting Challenge Track 2. The primary difficulty lies in the dataset, which contains strong background noise, reverberation, overlapping speech, and diverse meeting topics. To address these issues, we (a) designed G-SpatialNet, a speech enhancement (SE) model to improve Guided Source Separation (GSS) signals; (b) proposed TLS, a framework comprising time alignment, level alignment, and signal-to-noise ratio filtering, to generate signal-level pseudo labels for real-recorded far-field audio data, thereby facilitating SE models' training; and (c) explored fine-tuning strategies, data augmentation, and multimodal information to enhance the performance of pre-trained Automatic Speech Recognition (ASR) models in meeting scenarios. Finally, our system achieved character error rates (CERs) of 5.44% and 9.52% on the Dev and Eval sets, respectively, with relative improvements of 64.8% and 52.6% over the baseline, securing second place.

Key facts. Cited by 1 paper in the map (data through 19 Oct 2026). Placed in Speech enhancement and noise suppression, one of 11 directions in Speech and audio language models. Builds on 23 of its 33 references that are in the map. First posted on arXiv 30 May 2025.

Builds on

23 of 33 references.

Closest ancestors · 1

Same direction, newest first

  1. 2023The CHiME-7 UDASE task: Unsupervised domain adaptation for conversational speech enhancement7 Jul

Other references · 22

  1. 2022Robust Speech Recognition via Large-Scale Weak Supervisioncited by 4,661
  2. 2017Supervised Speech Separation Based on Deep Learning: An Overviewcited by 342
  3. 2021DNSMOS P.835: A Non-Intrusive Perceptual Objective Speech Quality Metric to Evaluate Noise Suppressorscited by 292
  4. 2021WenetSpeech: A 10000+ Hours Multi-domain Mandarin Corpus for Speech Recognitioncited by 270
  5. 2020CHiME-6 Challenge:Tackling Multispeaker Speech Recognition for Unsegmented Recordingscited by 231
  6. 2022Paraformer: Fast and Accurate Parallel Transformer for Non-autoregressive End-to-End Speech Recognitioncited by 198
  7. 2018The fifth 'CHiME' Speech Separation and Recognition Challenge: Dataset, task and baselinescited by 154
  8. 2021M2MeT: The ICASSP 2022 Multi-Channel Multi-Party Meeting Transcription Challengecited by 149
  9. 2021AISHELL-4: An Open Source Dataset for Speech Enhancement, Separation, Recognition and Speaker Diarization in Conference Scenariocited by 125
  10. 2020Glancing Transformer for Non-Autoregressive Neural Machine Translationcited by 106
  11. 2019CIF: Continuous Integrate-and-Fire for End-to-End Speech Recognitioncited by 104
  12. 2023The CHiME-7 DASR Challenge: Distant Meeting Transcription with Multiple Devices in Diverse Scenarioscited by 52
10 more
  1. 2024NOTSOFAR-1 Challenge: New Datasets, Baseline, and Tasks for Distant Meeting Transcriptioncited by 52
  2. 2023SpatialNet: Extensively Learning Spatial Information for Multichannel Joint Speech Separation, Denoising and Dereverberationcited by 39
  3. 2022GPU-accelerated Guided Source Separation for Meeting Transcriptioncited by 33
  4. 2021A Simultaneous Denoising and Dereverberation Framework with Target Decouplingcited by 28
  5. 2024The CHiME-8 DASR Challenge for Generalizable and Array Agnostic Distant Automatic Speech Recognition and Diarizationcited by 27
  6. 2021Convolutive Prediction for Monaural Speech Dereverberation and Noisy-Reverberant Speaker Separationcited by 21
  7. 2023The Multimodal Information based Speech Processing (MISP) 2022 Challenge: Audio-Visual Diarization and Recognitioncited by 18
  8. 2023UNSSOR: Unsupervised Neural Speech Separation by Leveraging Over-determined Training Mixturescited by 10
  9. 2023The Multimodal Information Based Speech Processing (MISP) 2023 Challenge: Audio-Visual Target Speaker Extractioncited by 9
  10. 2024Mixture to Mixture: Leveraging Close-talk Mixtures as Weak-supervision for Speech Separationcited by 5

Built on by

Cited by year (2026 to 19 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Cross-Talk Speech Reduction, by Separation, for Separationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025SuPseudo: A Pseudo-supervised Learning Method for Neural Speech Enhancement in Far-field Speech Recognition
  2. 2024Joint Beamforming and Speaker-Attributed ASR for Real Distant-Microphone Meeting Transcription
  3. 2022Pseudo Label Is Better Than Human Label
  4. 2021Large-Scale Pre-Training of End-to-End Multi-Talker ASR for Meeting Transcription with Single Distant Microphone
  5. 2026When AVSR Meets Video Conferencing: Dataset, Degradation, and the Hidden Mechanism Behind Performance Collapse
  6. 2021A Conformer-based ASR Frontend for Joint Acoustic Echo Cancellation, Speech Enhancement and Speech Separation
  7. 2025Lightweight Front-end Enhancement for Robust ASR via Frame Resampling and Sub-Band Pruning
  8. 2021Human Listening and Live Captioning: Multi-Task Training for Speech Enhancement