Data-Efficient Weakly Supervised Learning for Low-Resource Audio Event Detection Using Deep Learning

TL;DR

This paper introduces a data-efficient deep learning method for low-resource audio event detection that leverages weak labels and a novel loss function within a multi-instance learning framework, achieving improved performance on limited datasets.

Contribution

It presents a new training approach using a stacked CNN-RNN with a novel loss function for weakly supervised learning in low-resource audio detection.

Findings

Effective on two low-resource datasets

Improved training compared to standard weakly supervised methods

Demonstrates feasibility of weakly supervised deep learning for audio detection

Abstract

We propose a method to perform audio event detection under the common constraint that only limited training data are available. In training a deep learning system to perform audio event detection, two practical problems arise. Firstly, most datasets are "weakly labelled" having only a list of events present in each recording without any temporal information for training. Secondly, deep neural networks need a very large amount of labelled training data to achieve good quality performance, yet in practice it is difficult to collect enough samples for most classes of interest. In this paper, we propose a data-efficient training of a stacked convolutional and recurrent neural network. This neural network is trained in a multi instance learning setting for which we introduce a new loss function that leads to improved training compared to the usual approaches for weakly supervised learning. We successfully test our approach on two low-resource datasets that lack temporal labels.