An Interpretable Deep Learning Model for Automatic Sound Classification

Author:

Zinemanas Pablo1ORCID,Rocamora Martín2ORCID,Miron Marius1ORCID,Font Frederic1ORCID,Serra Xavier1ORCID

Affiliation:

1. Music Technology Group, Universitat Pompeu Fabra, 08018 Barcelona, Spain

2. Facultad de Ingeniería, Universidad de la República, Montevideo 11300, Uruguay

Abstract

Deep learning models have improved cutting-edge technologies in many research areas, but their black-box structure makes it difficult to understand their inner workings and the rationale behind their predictions. This may lead to unintended effects, such as being susceptible to adversarial attacks or the reinforcement of biases. There is still a lack of research in the audio domain, despite the increasing interest in developing deep learning models that provide explanations of their decisions. To reduce this gap, we propose a novel interpretable deep learning model for automatic sound classification, which explains its predictions based on the similarity of the input to a set of learned prototypes in a latent space. We leverage domain knowledge by designing a frequency-dependent similarity measure and by considering different time-frequency resolutions in the feature space. The proposed model achieves results that are comparable to that of the state-of-the-art methods in three different sound classification tasks involving speech, music, and environmental audio. In addition, we present two automatic methods to prune the proposed model that exploit its interpretability. Our system is open source and it is accompanied by a web application for the manual editing of the model, which allows for a human-in-the-loop debugging approach.

Publisher

MDPI AG

Reference84 articles.

1. Deep learning;LeCun;Nature,2015

2. Goodfellow, I., Bengio, Y., and Courville, A. (2016). Deep Learning, MIT Press. Available online: http://www.deeplearningbook.org.

3. Pereira, F., Burges, C.J.C., Bottou, L., and Weinberger, K.Q. (2012). ImageNet Classification with Deep Convolutional Neural Networks. Advances in Neural Information Processing Systems 25, Curran Associates, Inc.

4. Peters, M., Neumann, M., Iyyer, M., Gardner, M., Clark, C., Lee, K., and Zettlemoyer, L. (2018). Deep Contextualized Word Representations. Volume 1 (Long Papers), Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, New Orleans, LA, USA, 2–7 February 2018, Association for Computational Linguistics.

5. Deep Neural Networks for Acoustic Modeling in Speech Recognition: The Shared Views of Four Research Groups;Hinton;IEEE Signal Process. Mag.,2012

Cited by 42 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

1. AudioProtoPNet: An interpretable deep learning model for bird sound classification;Ecological Informatics;2025-07

2. Robust Threat Identification through Real-Time Audio Event Detection with 1D CNNs;International Journal of Emerging Technologies and Innovative Research;2025-06-11

3. (Mis)Communicating with our AI Systems;Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems;2025-04-25

4. Robust Relatable Explanations of Machine Learning with Disentangled Cue-specific Saliency;Proceedings of the 30th International Conference on Intelligent User Interfaces;2025-03-24

5. This Sounds Like That: Explainable Audio Classification via Prototypical Parts;Lecture Notes in Computer Science;2025

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.7亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2025 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3