Addressing Missing Labels in Large-Scale Sound Event Recognition Using a Teacher-Student Framework With Loss Masking

By Eduardo Fonseca, Shawn Hershey, Manoj Plakal, D. Ellis, Aren Jansen, R. Channing Moore, Xavier Serra. 2020

27 citations recorded by the source corpus.

The study of label noise in sound event recognition has recently gained attention with the advent of larger and noisier datasets. This work addresses the problem of missing labels, one of the big weaknesses of large audio datasets, and one of the most conspicuous issues for AudioSet. We propose a simple and model-agnostic method based on a teacher-student framework with loss masking to first identify the most critical missing label candidates, and then ignore their contribution during the learning process. We find that a simple optimisation of the training label set improves recognition performance without additional computation. We discover that most of the improvement comes from ignoring a critical tiny portion of the missing labels. We also show that the damage done by missing labels is larger as the training set gets smaller, yet it can still be observed even when training with massive amounts of audio. We believe these insights can generalize to other large-scale datasets.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

FSD50K: An Open Dataset of Human-Labeled Sound Events (2020, methodology): “FSD50K is an open and stable dataset aimed at complementing AudioSet in order to foster reproducible large-vocabulary SER research.”
The Benefit of Temporally-Strong Labels in Audio Event Classification (2021, background): “Leveraging only weakly labeled data to train audio event recognizers has been extensively investigated [8, 9, 10, 11, 12].”
Improving Sound Event Classification by Increasing Shift Invariance in Convolutional Neural Networks (2021, methodology): “Convolutional Neural Networks (CNNs) have been one of the cornerstones of Sound Event Classiﬁcation or Tagging (SET) in recent years [1, 2, 3, 4].”
Unsupervised Contrastive Learning of Sound Event Representations (2020, background): “Alternatives to conventional supervised learning include the paradigms of semi-supervised learning [3], few-shot learning [4], learning from noisy labels [5, 6], or self-supervision [7].”
Self-Supervised Learning from Automatically Separated Sound Scenes (2021, background): “For example, label noise issues such as missing, incorrect or inconsistent labels are inherent to this process [1, 2, 3, 4].”