Enregistré dans:
Détails bibliographiques
Auteurs principaux: Li, Duojia, Lu, Shenghui, Pan, Hongchen, Zhan, Zongyi, Hong, Qingyang, Li, Lin
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2509.14858
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866911483279966208
author Li, Duojia
Lu, Shenghui
Pan, Hongchen
Zhan, Zongyi
Hong, Qingyang
Li, Lin
author_facet Li, Duojia
Lu, Shenghui
Pan, Hongchen
Zhan, Zongyi
Hong, Qingyang
Li, Lin
contents Multistep inference is a bottleneck for real-time generative speech enhancement because flow- and diffusion-based systems learn an instantaneous velocity field and therefore rely on iterative ordinary differential equation (ODE) solvers. We introduce MeanFlowSE, a conditional generative model that learns the average velocity over finite intervals along a trajectory. Using a Jacobian-vector product (JVP) to instantiate the MeanFlow identity, we derive a local training objective that directly supervises finite-interval displacement while remaining consistent with the instantaneous-field constraint on the diagonal. At inference, MeanFlowSE performs single-step generation via a backward-in-time displacement, removing the need for multistep solvers; an optional few-step variant offers additional refinement. On VoiceBank-DEMAND, the single-step model achieves strong intelligibility, fidelity, and perceptual quality with substantially lower computational cost than multistep baselines. The method requires no knowledge distillation or external teachers, providing an efficient, high-fidelity framework for real-time generative speech enhancement. The proposed method is open-sourced at https://github.com/liduojia1/MeanFlowSE.
format Preprint
id arxiv_https___arxiv_org_abs_2509_14858
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MeanFlowSE: one-step generative speech enhancement via conditional mean flow
Li, Duojia
Lu, Shenghui
Pan, Hongchen
Zhan, Zongyi
Hong, Qingyang
Li, Lin
Sound
Artificial Intelligence
Multistep inference is a bottleneck for real-time generative speech enhancement because flow- and diffusion-based systems learn an instantaneous velocity field and therefore rely on iterative ordinary differential equation (ODE) solvers. We introduce MeanFlowSE, a conditional generative model that learns the average velocity over finite intervals along a trajectory. Using a Jacobian-vector product (JVP) to instantiate the MeanFlow identity, we derive a local training objective that directly supervises finite-interval displacement while remaining consistent with the instantaneous-field constraint on the diagonal. At inference, MeanFlowSE performs single-step generation via a backward-in-time displacement, removing the need for multistep solvers; an optional few-step variant offers additional refinement. On VoiceBank-DEMAND, the single-step model achieves strong intelligibility, fidelity, and perceptual quality with substantially lower computational cost than multistep baselines. The method requires no knowledge distillation or external teachers, providing an efficient, high-fidelity framework for real-time generative speech enhancement. The proposed method is open-sourced at https://github.com/liduojia1/MeanFlowSE.
title MeanFlowSE: one-step generative speech enhancement via conditional mean flow
topic Sound
Artificial Intelligence
url https://arxiv.org/abs/2509.14858