A Sequential Self Teaching Approach for Improving Generalization in Sound Event Recognition

Anurag Kumar, Vamsi Ithapu
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:5447-5457, 2020.

Abstract

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.

Cite this Paper

BibTeX
@InProceedings{pmlr-v119-kumar20a, title = {A Sequential Self Teaching Approach for Improving Generalization in Sound Event Recognition}, author = {Kumar, Anurag and Ithapu, Vamsi}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {5447--5457}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/kumar20a/kumar20a.pdf}, url = {https://proceedings.mlr.press/v119/kumar20a.html}, abstract = {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.} }
Endnote
%0 Conference Paper %T A Sequential Self Teaching Approach for Improving Generalization in Sound Event Recognition %A Anurag Kumar %A Vamsi Ithapu %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-kumar20a %I PMLR %P 5447--5457 %U https://proceedings.mlr.press/v119/kumar20a.html %V 119 %X 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.
APA
Kumar, A. & Ithapu, V.. (2020). A Sequential Self Teaching Approach for Improving Generalization in Sound Event Recognition. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:5447-5457 Available from https://proceedings.mlr.press/v119/kumar20a.html.

Related Material