跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2603.20387 · v2

End-to-End Multi-Task Learning for Adjustable Joint Noise Reduction and Hearing Loss Compensation

首次提交 2026-03-21 02:05
版本更新 2026-11-18 17:48
雷达观测 2026-11-19 15:04
arXiv ↗PDF ↗

摘要

A multi-task learning framework is proposed for optimizing a single deep neural network (DNN) for joint noise reduction (NR) and hearing loss compensation (HLC). A distinct training objective is defined for each task, and the DNN predicts two time-frequency masks. During inference, NR and HLC can be adjusted independently by exponentiating each mask before combining them. In contrast to recent approaches that rely on training an auditory-model emulator to define a differentiable training objective, we propose an auditory model that is inherently differentiable, thus allowing end-to-end optimization. The audiogram is provided as an input to the DNN, thereby enabling listener-specific personalization without the need for retraining. Results show that the proposed approach not only allows adjusting NR and HLC independently, but also improves objective metrics compared with optimizing a single training objective. It also outperforms a cascade of two DNNs separately trained for NR and HLC and shows competitive HLC performance against a traditional hearing-aid prescription on objective metrics. To the best of our knowledge, this is the first study that uses an auditory model to train a single DNN for both NR and HLC across a wide range of flat and high-frequency-sloping hearing-loss profiles.

作者备注:Accepted version. Models retrained with an updated auditory model (49 CFs up to 10 kHz). Added paired statistical analysis, gain analyses, related work, and a discussion on scope and limitations

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

本站首次收录为 v2,暂无更早版本可对照。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v2 · 2026-11-18 17:48 · 最新观测版本

End-to-End Multi-Task Learning for Adjustable Joint Noise Reduction and Hearing Loss Compensation

A multi-task learning framework is proposed for optimizing a single deep neural network (DNN) for joint noise reduction (NR) and hearing loss compensation (HLC). A distinct training objective is defined for each task, and the DNN predicts two time-frequency masks. During inference, NR and HLC can be adjusted independently by exponentiating each mask before combining them. In contrast to recent approaches that rely on training an auditory-model emulator to define a differentiable training objective, we propose an auditory model that is inherently differentiable, thus allowing end-to-end optimization. The audiogram is provided as an input to the DNN, thereby enabling listener-specific personalization without the need for retraining. Results show that the proposed approach not only allows adjusting NR and HLC independently, but also improves objective metrics compared with optimizing a single training objective. It also outperforms a cascade of two DNNs separately trained for NR and HLC and shows competitive HLC performance against a traditional hearing-aid prescription on objective metrics. To the best of our knowledge, this is the first study that uses an auditory model to train a single DNN for both NR and HLC across a wide range of flat and high-frequency-sloping hearing-loss profiles.

查看此版本解读 →查看此版本原文 ↗