A Sequential Self Teaching Approach for Improving Generalization in Sound Event Recognition
This addresses the challenge of sound event recognition in adverse labeling conditions, offering an incremental improvement for machine auditory perception.
The paper tackles the problem of learning sound events from weakly or noisy labeled data by proposing a sequential self-teaching approach, which achieves up to 9% performance improvement on Audioset and enhances generalization and transfer learning capabilities.
An important problem in machine auditory perception is to recognize and detect sound events. In this paper, we propose a sequential self-teaching approach to learning sounds. Our main proposition is that it is harder to learn sounds in adverse situations such as from weakly labeled and/or noisy labeled data, and in these situations a single stage of learning is not sufficient. Our proposal is a sequential stage-wise learning process that improves generalization capabilities of a given modeling system. We justify this method via technical results and on Audioset, the largest sound events dataset, our sequential learning approach can lead to up to 9% improvement in performance. A comprehensive evaluation also shows that the method leads to improved transferability of knowledge from previously trained models, thereby leading to improved generalization capabilities on transfer learning tasks.