Self-supervised Attention Model for Weakly Labeled Audio Event Classification
This addresses the problem of reducing labeling costs for audio event classification, particularly benefiting short audio events, though it is incremental as it builds on prior attention models.
The paper tackles weakly labeled audio event classification by proposing a self-supervised attention model to distinguish relevant parts of audio clips without expensive labeling, achieving 8.8% and 17.6% relative mean average precision improvements over state-of-the-art systems on SL-DCASE-17 and balanced AudioSet.
We describe a novel weakly labeled Audio Event Classification approach based on a self-supervised attention model. The weakly labeled framework is used to eliminate the need for expensive data labeling procedure and self-supervised attention is deployed to help a model distinguish between relevant and irrelevant parts of a weakly labeled audio clip in a more effective manner compared to prior attention models. We also propose a highly effective strongly supervised attention model when strong labels are available. This model also serves as an upper bound for the self-supervised model. The performances of the model with self-supervised attention training are comparable to the strongly supervised one which is trained using strong labels. We show that our self-supervised attention method is especially beneficial for short audio events. We achieve 8.8% and 17.6% relative mean average precision improvements over the current state-of-the-art systems for SL-DCASE-17 and balanced AudioSet.