analiza aiqbench
This paper introduces Schrödinger Bridge Mamba (SBM), a novel architecture that combines the Schrödinger Bridge training paradigm with the state-space Mamba model for efficient speech enhancement. By achieving high-quality denoising and dereverberation in just one inference step, SBM demonstrates superior performance over traditional attention-based methods like Transformers while maintaining real-time feasibility for streaming applications. The results underscore the growing synergy between advanced trajectory-based learning frameworks and linear-complexity sequence models to solve practical audio processing challenges.
Schr\"odinger Bridge Mamba for One-Step Speech Enhancement
arXiv:2510.16834v3 Announce Type: replace-cross Abstract: We present Schr\"odinger Bridge Mamba (SBM), a novel model for efficient speech enhancement by integrating the Schr\"odinger Bridge (SB) training paradigm and the Mamba architecture. Experiments of joint denoising and dereverberation tasks demonstrate SBM outperforms strong generative and discriminative methods on multiple metrics with only one step of inference while achieving a competitive real-time factor for streaming feasibility. Ablation studies reveal that the SB paradigm consistently yields improved performance across diverse architectures over conventional mapping. Furthermore, Mamba exhibits a stronger performance under the SB paradigm compared to Multi-Head Self-Attention (MHSA) and Long Short-Term Memory (LSTM) backbones. These findings highlight the synergy between the Mamba architecture and the SB trajectory-based training, providing a high-quality solution for real-world speech enhancement. Demo page: https://sbmse.github.io
Abstract:We present Schrödinger Bridge Mamba (SBM), a novel model for efficient speech enhancement by integrating the Schrödinger Bridge (SB) training paradigm and the Mamba architecture. Experiments of joint denoising and dereverberation tasks demonstrate SBM outperforms strong generative and discriminative methods on multiple metrics with only one step of inference while achieving a competitive real-time factor for streaming feasibility. Ablation studies reveal that the SB paradigm consistently yields improved performance across diverse architectures over conventional mapping. Furthermore, Mamba exhibits a stronger performance under the SB paradigm compared to Multi-Head Self-Attention (MHSA) and Long Short-Term Memory (LSTM) backbones. These findings highlight the synergy between the Mamba architecture and the SB trajectory-based training, providing a high-quality solution for real-world speech enhancement. Demo page: this https URL
Sound (cs.SD); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Audio and Speech Processing (eess.AS)
arXiv:2510.16834 [cs.SD]
(or arXiv:2510.16834v3 [cs.SD] for this version)
Text integral sub licența deschisă a publicației · 144 cuvinte
Subiecte conexe
SE-MSB: End-to-End Unpaired Speech Enhancement using Mamba Schr\"odinger Bridges
arXiv · 24 sept. 2026
Regularized Schr\"odinger Bridge via Distortion-Perception Perturbation for High-Fidelity Speech Enhancement
arXiv · 4 aug. 2026
BridgeCodec: Mamba Enhanced Neural Audio Codec with Schr\"odinger Bridge at Low Bitrate
arXiv · 15 oct. 2026
[AINews] Claude Opus 5.5, the new default model for AINews — and everybody cuts prices 40-50%
Latent Space · 23 sept. 2026