Skip to main content

Weighted and Multi-Task Loss for Rare Audio Event Detection

Phan, Huy, Krawczyk-Becker, Martin, Gerkmann, Timo, Mertins, Alfred (2018) Weighted and Multi-Task Loss for Rare Audio Event Detection. In: 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing: Proceedings. . pp. 336-340. IEEE, Calgary, Canada ISBN 978-1-5386-4658-8. (doi:10.1109/ICASSP.2018.8461353) (KAR id:72667)

PDF Author's Accepted Manuscript
Language: English
Download (204kB) Preview
[img]
Preview
Official URL
https://doi.org/10.1109/ICASSP.2018.8461353

Abstract

We present in this paper two loss functions tailored for rare audio event detection in audio streams. The weighted loss is designed to tackle the common issue of imbalanced data in background/foreground classification while the multi-task loss enables the networks to simultaneously model the class distribution and the temporal structures of the target events for recognition. We study the proposed loss functions with deep neural networks (DNNs) and convolutional neural networks (CNNs) coupled with state-of-the-art phase-aware signal enhancement. Experiments on the DCASE 2017 challenge’s data show that our system with the proposed losses significantly outperforms not only the DCASE 2017 baseline but also our baseline which has a similar network architecture and a standard loss function.

Item Type: Conference or workshop item (Proceeding)
DOI/Identification number: 10.1109/ICASSP.2018.8461353
Uncontrolled keywords: audio event detection, convolutional neural networks, deep neural networks, weighted loss, multi-task loss
Divisions: Faculties > Sciences > School of Computing
Faculties > Sciences > School of Computing > Data Science
Depositing User: Huy Phan
Date Deposited: 22 Feb 2019 10:58 UTC
Last Modified: 03 Jun 2019 09:27 UTC
Resource URI: https://kar.kent.ac.uk/id/eprint/72667 (The current URI for this page, for reference purposes)
Phan, Huy: https://orcid.org/0000-0003-4096-785X
  • Depositors only (login required):

Downloads

Downloads per month over past year