Deep learning, audio adversaries, and music content analysis

Corey Mose Kereliuk, Bob L. Sturm, Jan Larsen

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Abstract

We present the concept of adversarial audio in the context of deep neural networks (DNNs) for music content analysis. An adversary is an algorithm that makes minor perturbations to an input that cause major repercussions to the system response. In particular, we design an adversary for a DNN that takes as input short-time spectral magnitudes of recorded music and outputs a high-level music descriptor. We demonstrate how this adversary can make the DNN behave in any way with only extremely minor changes to the music recording signal. We show that the adversary cannot be neutralised by a simple filtering of the input. Finally, we discuss adversaries in the broader context of the evaluation of music content analysis systems.
Original languageEnglish
Title of host publicationProceedings of the IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA 2015)
PublisherIEEE
Publication date2015
Pages1-5
DOIs
Publication statusPublished - 2015
EventIEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA 2015) - Mohonk Mountain House, New Paltz, New York, United States
Duration: 18 Oct 201521 Oct 2015
http://www.waspaa.com/

Workshop

WorkshopIEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA 2015)
LocationMohonk Mountain House
CountryUnited States
CityNew Paltz, New York
Period18/10/201521/10/2015
Internet address

Keywords

  • Deep Learning
  • Music Content Analysis

Cite this

Kereliuk, C. M., Sturm, B. L., & Larsen, J. (2015). Deep learning, audio adversaries, and music content analysis. In Proceedings of the IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA 2015) (pp. 1-5). IEEE. https://doi.org/10.1109/WASPAA.2015.7336950