mlsanityData to 14 Oct 2026

Universal Speech Enhancement with Regression and Generative Mamba

arXiv 2505.21198 ↗First posted 27 May 2025cs.SD eess.AScited by 7

On the map: State-space sequence models (S4, Mamba) → Speech and audio SSMs

Abstract (arXiv)

The Interspeech 2025 URGENT Challenge aimed to advance universal, robust, and generalizable speech enhancement by unifying speech enhancement tasks across a wide variety of conditions, including seven different distortion types and five languages. We present Universal Speech Enhancement Mamba (USEMamba), a state-space speech enhancement model designed to handle long-range sequence modeling, time-frequency structured processing, and sampling frequency-independent feature extraction. Our approach primarily relies on regression-based modeling, which performs well across most distortions. However, for packet loss and bandwidth extension, where missing content must be inferred, a generative variant of the proposed USEMamba proves more effective. Despite being trained on only a subset of the full training data, USEMamba achieved 2nd place in Track 1 during the blind test phase, demonstrating strong generalization across diverse conditions.

Key facts. Cited by 7 papers in the map (data through 14 Oct 2026). Placed in Speech and audio SSMs, one of 10 directions in State-space sequence models (S4, Mamba). Builds on 28 of its 38 references that are in the map. First posted on arXiv 27 May 2025.

Builds on

28 of 38 references.

Closest ancestors · 3

Same direction, newest first

  1. 2024An Investigation of Incorporating Mamba for Speech Enhancement10 May
  2. 2023Spiking Structured State Space Model for Monaural Speech Enhancement7 Sep
  3. 2023A Multi-dimensional Deep Structured State Space Approach to Speech Enhancement Using Small-footprint Models1 Jun

Other references · 25

  1. 2022Scalable Diffusion Models with Transformerscited by 6,054
  2. 2022Flow Matching for Generative Modelingcited by 5,702
  3. 2023Mamba: Linear-Time Sequence Modeling with Selective State Spacescited by 4,718
  4. 2022On the Parameterization and Initialization of Diagonal State Space Modelscited by 424
  5. 2018SDR - half-baked or well done?cited by 423
  6. 2022UTMOS: UTokyo-SaruLab System for VoiceMOS Challenge 2022cited by 390
  7. 2023Voicebox: Text-Guided Multilingual Universal Speech Generation at Scalecited by 379
  8. 2017Supervised Speech Separation Based on Deep Learning: An Overviewcited by 216
  9. 2021NISQA: A Deep CNN-Self-Attention Model for Multidimensional Speech Quality Prediction with Crowdsourced Datasetscited by 192
  10. 2021DNSMOS P.835: A Non-Intrusive Perceptual Objective Speech Quality Metric to Evaluate Noise Suppressorscited by 172
  11. 2022Universal Speech Enhancement with Score-based Diffusioncited by 75
  12. 2023SpeechX: Neural Codec Language Model as a Versatile Speech Transformercited by 64
13 more
  1. 2023Generative Pre-training for Speech with Flow Matchingcited by 55
  2. 2023MP-SENet: A Speech Enhancement Model with Parallel Denoising of Magnitude and Phase Spectracited by 52
  3. 2020HiFi-GAN: High-Fidelity Denoising and Dereverberation Based on Speech Deep Features in Adversarial Networkscited by 40
  4. 2024SpeechBERTScore: Reference-Aware Automatic Evaluation of Speech Generation Leveraging NLP Evaluation Metricscited by 38
  5. 2022VoiceFixer: A Unified Framework for High-Fidelity Speech Restorationcited by 37
  6. 2021DPT-FSNet: Dual-path Transformer Based Full-band and Sub-band Fusion Network for Speech Enhancementcited by 33
  7. 2024MaskSR: Masked Language Model for Full-band Speech Restorationcited by 24
  8. 2022CMGAN: Conformer-Based Metric-GAN for Monaural Speech Enhancementcited by 23
  9. 2025AnyEnhance: A Unified Generative Model with Prompt-Guidance and Self-Critic for Voice Enhancementcited by 23
  10. 2023Miipher: A Robust Speech Restoration Model Integrating Self-Supervised Speech and Text Representationscited by 22
  11. 2024URGENT Challenge: Universality, Robustness, and Generalizability For Speech Enhancementcited by 17
  12. 2024Generative Speech Foundation Model Pretraining for High-Quality Speech Extraction and Restorationcited by 16
  13. 2023Toward Universal Speech Enhancement for Diverse Input Conditionscited by 10

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 5; 2026: 2.

7 papers.

  1. 2025EDNet: A Versatile Speech Enhancement Framework with Gating Mamba Mechanism and Phase Shift-Invariant Trainingcited by 3
  2. 2025Schr"odinger Bridge Mamba for One-Step Speech Enhancementcited by 3
  3. 2025End-to-end audio-visual learning for cochlear implant sound coding simulations in noisy environmentscited by 0
  4. 2025Leveraging Mamba with Full-Face Vision for Audio-Visual Speech Enhancementcited by 0
  5. 2025An overview of neural architectures for self-supervised audio representation learning from masked spectrogramscited by 0
  6. 2026UniPASE: A Generative Model for Universal Speech Enhancement with High Fidelity and Low Hallucinationscited by 0
  7. 2026RT-SEMamba: Real-Time Speech Enhancement Mamba via Progressive Knowledge Distillationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Selective State Space Model for Monaural Speech Enhancement
  2. 2024Mamba-SEUNet: Mamba UNet for Monaural Speech Enhancement
  3. 2024MC-SEMamba: A Simple Multi-channel Extension of SEMamba
  4. 2024Leveraging Joint Spectral and Spatial Learning with MAMBA for Multichannel Speech Enhancement
  5. 2025MambAttention: Mamba with Multi-Head Attention for Generalizable Single-Channel Speech Enhancement
  6. 2025Improving Speech Enhancement by Cross- and Sub-band Processing with State Space Model
  7. 2025A Composite Predictive-Generative Approach to Monaural Universal Speech Enhancement
  8. 2026HAMMER: Harmonic-Aware Parallel Context Modeling and Discriminator-Free Perceptual Optimization for Speech Enhancement