Limitations of weak labels for embedding and tagging - Archive ouverte HAL Access content directly
Conference Papers Year :

Limitations of weak labels for embedding and tagging

(1) , (1) , (1)
1

Abstract

Many datasets and approaches in ambient sound analysis use weakly labeled data. Weak labels are employed because annotating every data sample with a strong label is too expensive. Yet, their impact on the performance in comparison to strong labels remains unclear. Indeed, weak labels must often be dealt with at the same time as other challenges, namely multiple labels per sample, unbalanced classes and/or overlapping events. In this paper, we formulate a supervised learning problem which involves weak labels. We create a dataset that focuses on the difference between strong and weak labels as opposed to other challenges. We investigate the impact of weak labels when training an embedding or an end-to-end classifier. Different experimental scenarios are discussed to provide insights into which applications are most sensitive to weakly labeled data.
Fichier principal
Vignette du fichier
icassp2020.pdf (267.72 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-02467401 , version 1 (04-02-2020)
hal-02467401 , version 2 (07-02-2020)
hal-02467401 , version 3 (30-04-2020)
hal-02467401 , version 4 (07-12-2020)

Identifiers

Cite

Nicolas Turpault, Romain Serizel, Emmanuel Vincent. Limitations of weak labels for embedding and tagging. ICASSP 2020 - 45th International Conference on Acoustics, Speech, and Signal Processing, May 2020, Barcelona, Spain. ⟨hal-02467401v4⟩
247 View
246 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More