Title: Posterior Transition Modeling for Unsupervised Diffusion-Based Speech Enhancement Show Chinese title

Title: 基于后验转换的无监督扩散语音增强

Abstract: We explore unsupervised speech enhancement using diffusion models as expressive generative priors for clean speech. Existing approaches guide the reverse diffusion process using noisy speech through an approximate, noise-perturbed likelihood score, combined with the unconditional score via a trade-off hyperparameter. In this work, we propose two alternative algorithms that directly model the conditional reverse transition distribution of diffusion states. The first method integrates the diffusion prior with the observation model in a principled way, removing the need for hyperparameter tuning. The second defines a diffusion process over the noisy speech itself, yielding a fully tractable and exact likelihood score. Experiments on the WSJ0-QUT and VoiceBank-DEMAND datasets demonstrate improved enhancement metrics and greater robustness to domain shifts compared to both supervised and unsupervised baselines. Show Chinese abstract

Abstract: 我们探索使用扩散模型作为干净语音的表达生成先验的无监督语音增强。 现有方法通过一个近似、噪声扰动的似然梯度，结合无条件梯度，利用一个权衡超参数来引导反向扩散过程。 在本工作中，我们提出了两种替代算法，直接建模扩散状态的条件反向转移分布。 第一种方法以一种合理的方式将扩散先验与观测模型结合，消除了对超参数调优的需求。 第二种方法在噪声语音本身上定义了一个扩散过程，产生了一个完全可处理且精确的似然梯度。 在WSJ0-QUT和VoiceBank-DEMAND数据集上的实验表明，与监督和无监督基线相比，增强了指标并具有更高的领域迁移鲁棒性。