Self-Supervised Learning for Anomalous Sound Detection
This work addresses the need for robust ASD systems without manual annotations, offering a significant improvement in performance for audio anomaly detection tasks.
The paper tackles the problem of anomalous sound detection (ASD) by proposing a self-supervised learning method called FeatEx, which achieves state-of-the-art performance on the DCASE2023 dataset, outperforming all other published results by a large margin.
State-of-the-art anomalous sound detection (ASD) systems are often trained by using an auxiliary classification task to learn an embedding space. Doing so enables the system to learn embeddings that are robust to noise and are ignoring non-target sound events but requires manually annotated meta information to be used as class labels. However, the less difficult the classification task becomes, the less informative are the embeddings and the worse is the resulting ASD performance. A solution to this problem is to utilize self-supervised learning (SSL). In this work, feature exchange (FeatEx), a simple yet effective SSL approach for ASD, is proposed. In addition, FeatEx is compared to and combined with existing SSL approaches. As the main result, a new state-of-the-art performance for the DCASE2023 ASD dataset is obtained that outperforms all other published results on this dataset by a large margin.