arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2401.13611cs.SDcs.AIeess.AS
收藏

利用中间 ASR 特征和人类记忆模型为听障用户进行非侵入式语音可懂度预测

Non-Intrusive Speech Intelligibility Prediction for Hearing-Impaired Users using Intermediate ASR Features and Human Memory Models

University of Sheffield(谢菲尔德大学)

Rhiannon Mogridge, George Close, Robert Sutherland, Thomas Hain, Jon Barker, Stefan Goetze, Anton Ragni

AI总结:

本文提出结合 Whisper ASR 解码器中间特征与基于样例的人类记忆模型,用于预测听障用户的语音可懂度,并在未见增强系统和听者上显著优于 HASPI 基线。

摘要:

神经网络已成功用于非侵入式语音可懂度预测。近来,研究发现,使用来自预训练自监督和弱监督模型中间层的特征表示对此任务尤其有用。本研究将 Whisper ASR 解码器层表示作为神经网络输入特征,并结合一种基于样例、具有心理学依据的人类记忆模型,来预测助听器用户的人类可懂度评分。研究发现,相较于成熟的侵入式 HASPI 基线系统,该方法取得了显著的性能提升,包括在训练数据中未见过的增强系统和听者上;其均方根误差为 25.3,而基线为 28.7。

英文摘要:

Neural networks have been successfully used for non-intrusive speech intelligibility prediction. Recently, the use of feature representations sourced from intermediate layers of pre-trained self-supervised and weakly-supervised models has been found to be particularly useful for this task. This work combines the use of Whisper ASR decoder layer representations as neural network input features with an exemplar-based, psychologically motivated model of human memory to predict human intelligibility ratings for hearing-aid users. Substantial performance improvement over an established intrusive HASPI baseline system is found, including on enhancement systems and listeners unseen in the training data, with a root mean squared error of 25.3 compared with the baseline of 28.7.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑