Weakly Labeled Sound Event Detection Using Tri-training and Adversarial Learning
arXiv:1910.06790
Abstract
This paper considers a semi-supervised learning framework for weakly labeled polyphonic sound event detection problems for the DCASE 2019 challenge's task4 by combining both the tri-training and adversarial learning. The goal of the task4 is to detect onsets and offsets of multiple sound events in a single audio clip. The entire dataset consists of the synthetic data with a strong label (sound event labels with boundaries) and real data with weakly labeled (sound event labels) and unlabeled dataset. Given this dataset, we apply the tri-training where two different classifiers are used to obtain pseudo labels on the weakly labeled and unlabeled dataset, and the final classifier is trained using the strongly labeled dataset and weakly/unlabeled dataset with pseudo labels. Also, we apply the adversarial learning to reduce the domain gap between the real and synthetic dataset. We evaluated our learning framework using the validation set of the task4 dataset, and in the experiments, our learning framework shows a considerable performance improvement over the baseline model.
5 pages, DCASE 2019 Workshop
References in corpus (8)
- Adam: A Method for Stochastic Optimization
- Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results
- Convolutional Recurrent Neural Networks for Polyphonic Sound Event Detection
- Asymmetric Tri-training for Unsupervised Domain Adaptation
- Gated Feedback Recurrent Neural Networks
- Recurrent Neural Networks for Polyphonic Sound Event Detection in Real Life Recordings
- Audio Event Detection using Weakly Labeled Data
- Large-Scale Weakly Labeled Semi-Supervised Sound Event Detection in Domestic Environments