arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2408.06264cs.SDcs.AIeess.AS
收藏

面向计算机听觉的音频增强——一种利用样本重要性的迭代训练范式

Audio Enhancement for Computer Audition -- An Iterative Training Paradigm Using Sample Importance

University of Augsburg(奥格斯堡大学);Technical University of Munich(慕尼黑工业大学);Munich Center for Machine Learning(慕尼黑机器学习中心);Huawei Technologies(华为技术有限公司);Munich Data Science Institute(慕尼黑数据科学研究所);Imperial College London(伦敦帝国学院);Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所)

Manuel Milling, Shuo Liu, Andreas Triantafyllopoulos, Ilhan Aslan, Björn W. Schuller

AI总结:

针对音频任务模型易受噪声干扰的问题,提出利用逐样本性能指标作为样本重要性引导的端到端联合优化训练范式,可显著提升多类计算机听觉任务模型的噪声鲁棒性,低信噪比下效果更突出。

摘要:

用于自动语音识别(ASR)、声学场景分类(ASC)等音频任务的神经网络模型,在实际应用中易受噪声干扰。为提升音频质量,目标音频应用的前端会专门设置可独立开发的增强模块。本文提出一种端到端学习方案,可联合优化音频增强(AE)模型与后续应用模型。为引导AE模块的优化朝向目标应用方向推进,尤其是攻克难样本,我们采用逐样本性能指标作为样本重要性的衡量依据。实验中,我们选取四类代表性应用评估该训练范式,分别为ASR、语音命令识别(SCR)、语音情感识别(SER)和ASC,覆盖涉及语义与非语义特征、瞬态与全局信息的语音及非语音任务。实验结果表明,在日常噪声环境下的各类计算机听觉任务中,所提方法可显著提升模型的噪声鲁棒性,在低信噪比(SNR)条件下效果尤为突出。

英文摘要:

Neural network models for audio tasks, such as automatic speech recognition (ASR) and acoustic scene classification (ASC), are susceptible to noise contamination for real-life applications. To improve audio quality, an enhancement module, which can be developed independently, is explicitly used at the front-end of the target audio applications. In this paper, we present an end-to-end learning solution to jointly optimise the models for audio enhancement (AE) and the subsequent applications. To guide the optimisation of the AE module towards a target application, and especially to overcome difficult samples, we make use of the sample-wise performance measure as an indication of sample importance. In experiments, we consider four representative applications to evaluate our training paradigm, i.e., ASR, speech command recognition (SCR), speech emotion recognition (SER), and ASC. These applications are associated with speech and non-speech tasks concerning semantic and non-semantic features, transient and global information, and the experimental results indicate that our proposed approach can considerably boost the noise robustness of the models, especially at low signal-to-noise ratios (SNRs), for a wide range of computer audition tasks in everyday-life noisy environments.

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。

↑