# Few-step Adversarial Schr"odinger Bridge for Generative Speech Enhancement

arXiv: https://arxiv.org/abs/2506.01460 · HTML: https://mlsanity.com/paper/2506.01460/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 2 Jun 2025.

## Built on by (most cited first, 6 of 6)

- [VoiceBridge: General Speech Restoration with One-step Latent Bridge Models](https://mlsanity.com/paper/2509.25275/) (2025)
- [Schr"odinger Bridge Mamba for One-Step Speech Enhancement](https://mlsanity.com/paper/2510.16834/) (2025)
- [Speech Enhancement Based on Drifting Models](https://mlsanity.com/paper/2604.24199/) (2026)
- [A Fast Solver for Interpolating Stochastic Differential Equation Diffusion Models for Speech Restoration](https://mlsanity.com/paper/2603.09508/) (2026)
- [HybridSB-MoE: Dual-Domain Schr"odinger Bridges with Scene-Adaptive Expert Routing for Speech Enhancement](https://mlsanity.com/paper/2608.12715/) (2026)
- [BridgeCodec: Mamba Enhanced Neural Audio Codec with Schr"odinger Bridge at Low Bitrate](https://mlsanity.com/paper/2610.15453/) (2026)

## Abstract (arXiv)

Deep generative models have recently been employed for speech enhancement to generate perceptually valid clean speech on large-scale datasets. Several diffusion models have been proposed, and more recently, a tractable Schr"odinger Bridge has been introduced to transport between the clean and noisy speech distributions. However, these models often suffer from an iterative reverse process and require a large number of sampling steps -- more than 50. Our investigation reveals that the performance of baseline models significantly degrades when the number of sampling steps is reduced, particularly under low-SNR conditions. We propose integrating Schr"odinger Bridge with GANs to effectively mitigate this issue, achieving high-quality outputs on full-band datasets while substantially reducing the required sampling steps. Experimental results demonstrate that our proposed model outperforms existing baselines, even with a single inference step, in both denoising and dereverberation tasks.