Confidence Learning for Semi-Supervised Acoustic Event Detection
Source abstract
Abstract
In recent years, the involvement of synthetic strongly labeled data, weakly labeled data, and unlabeled data has drawn much research attention in semi-supervised acoustic event detection (SAED). The classic self-training method carries out predictions for unlabeled data and then selects predictions with high probabilities as pseudo-labels for retraining. Such models have shown its effectiveness in SAED. However, probabilities are poorly calibrated confidence estimates, and samples with low probabilities are ignored. Hence, we introduce a confidence-based semi-supervised Acoustic event detection (C-SAED) framework. The C-SAED method learns confidence deliberately and retrains all data distinctly by applying confidence as weights. Additionally, we apply a power pooling function whose coefficient can be trained automatically and use weakly labeled data more efficiently. The experimental results demonstrate that the generated confidence is proportional to the accuracy of the predictions. Our C-SAED framework achieves a relative error rate reduction of 34% in contrast to the baseline model.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Music and Audio Processing · Speech and Audio Processing · Music Technology and Sound Studies
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Key Laboratory of Speech Acoustics and Content Understanding, Institute of Acoustics, Chinese Academy of Sciences, Beijing 100190, China
University of Chinese Academy of Sciences, Beijing 100049, China
National Internet Emergency Center, Beijing 100029, China
Funding
National Natural Science Foundation of China
References (26)
Show source-provided references
1. SONYC: A System for Monitoring, Analysis and Mitigation of Urban Noise Pollution10.1145/3224204 ↗
2. Lostanlen, V., Salamon, J., Farnsworth, A., Kelling, S., and Bello, J.P. (2018, January 15–20). Birdvox-Full-Night: A Dataset and Benchmark for Avian Flight Call Detection. Proceedings of the 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Calgary, AB, Canada.10.1109/ICASSP.2018.8461410 ↗
3. Audio Surveillance: A Systematic Review10.1145/2871183 ↗
4. Hershey, S., Chaudhuri, S., Ellis, D.P.W., Gemmeke, J.F., Jansen, A., Moore, R.C., Plakal, M., Platt, D., Saurous, R.A., and Seybold, B. (2017, January 5–9). CNN architectures for large-scale audio classification. Proceedings of the 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), New Orleans, LA, USA.10.1109/ICASSP.2017.7952132 ↗
5. Reliable Detection of Audio Events in Highly Noisy Environments10.1016/j.patrec.2015.06.026 ↗
6. Context-dependent sound event detection10.1186/1687-4722-2013-1 ↗
7. Cakir, E., Heittola, T., Huttunen, H., and Virtanen, T. (2015, January 12–17). Polyphonic sound event detection using multi label deep neural networks. Proceedings of the 2015 International Joint Conference on Neural Networks (IJCNN), Killarney, Ireland.10.1109/IJCNN.2015.7280624 ↗
8. Parascandolo, G., Huttunen, H., and Virtanen, T. (2016, January 20–25). Recurrent Neural Networks for Polyphonic Sound Event Detection in Real Life Recordings. Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing, Shanghai, China.10.1109/ICASSP.2016.7472917 ↗
9. Phan, H., Hertel, L., Maass, M., and Mertins, A. (2016). Robust Audio Event Recognition with 1-Max Pooling Convolutional Neural Networks. arXiv.10.21437/Interspeech.2016-123 ↗
10. Akr, E., and Virtanen, T. (2018, January 8–13). End-to-End Polyphonic Sound Event Detection Using Convolutional Recurrent Neural Networks with Learned Time-Frequency Representation Input. Proceedings of the 2018 International Joint Conference on Neural Networks (IJCNN), Rio de Janeiro, Brazil.
11. Serizel, R., Turpault, N., Eghbal-Zadeh, H., and Shah, A.P. (2018, January 19–20). Large-scale weakly labeled semi-supervised sound event detection in domestic environments. Proceedings of the Detection and Classification of Acoustic Scenes and Events 2018 Workshop (DCASE2018), Surrey, UK.
12. Turpault, N., Serizel, R., Parag Shah, A., and Salamon, J. (2019, January 25–26). Sound event detection in domestic environments with weakly labeled data and soundscape synthesis. Proceedings of the Detection and Classification of Acoustic Scenes and Events 2019 Workshop (DCASE2019), New York, NY, USA.10.33682/006b-jx26 ↗
13. Lin, L., Wang, X., Liu, H., and Qian, Y. (2020, January 4–8). Guided Learning for Weakly-Labeled Semi-Supervised Sound Event Detection. Proceedings of the 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Barcelona, Spain.10.1109/ICASSP40776.2020.9053584 ↗
14. Yan, J., Song, Y., Dai, L., and McLoughlin, I. (2020, January 4–8). Task-Aware Mean Teacher Method for Large Scale Weakly Labeled Semi-Supervised Sound Event Detection. Proceedings of the ICASSP 2020—2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Barcelona, Spain.10.1109/ICASSP40776.2020.9053073 ↗
15. Tarvainen, A., and Valpola, H. (2017). Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results. arXiv.
16. Research on Semi-Supervised Sound Event Detection Based on Mean Teacher Models Using ML-LoBCoD-NET10.1109/ACCESS.2020.2974479 ↗
17. Lu, J. (2021, September 14). Mean Teacher Convolution System for Dcase 2018 Task 4. Available online: http://dcase.community/documents/challenge2018/technical_reports/DCASE2018_Lu_19.pdf.
18. Lionel, D., and Cyril, P. (2019). Mean Teacher with Data Augmentation for Dcase 2019 Task 4, Orange Labs. DCASE2019 Challenge.
19. McClosky, D., Charniak, E., and Johnson, M. (2006, January 4–9). Effective Self-training for Parsing. Proceedings of the Human Language Technology Conference of the North American Chapter of the Association of Computational Linguistics, New York, NY, USA.10.3115/1220835.1220855 ↗
20. Liu, Y.L., Yan, J., and Song, Y. (2021, September 14). Ustc-Nelslip System for Dcase 2018 Challenge Task 4. Available online: http://dcase.community/documents/challenge2018/technical_reports/DCASE2018_Liu_69.pdf.
21. Elizalde, B., Shah, A., Dalmia, S., Min, H.L., and Lane, I. (September, January 28). An Approach for Self-Training Audio Event Detectors Using Web Data. Proceedings of the 2017 25th European Signal Processing Conference (EUSIPCO), Kos, Greece.10.23919/EUSIPCO.2017.8081532 ↗
22. Kothintiu, S., Imoto, K., and Charkrabarty, D. (2021, September 14). Joint Acoustic and Class Inference for Weakly Supervised Sound Event Detection. Available online: http://dcase.community/documents/challenge2018/technical_reports/DCASE2018_Kothinti_90.pdf.
23. Guo, C., Pleiss, G., Yu, S., and Weinberger, K.Q. (2017, January 6–11). On Calibration of Modern Neural Networks. Proceedings of the 34th International Conference on Machine Learning, Sydney, Australia.
24. DeVries, T., and Graham, W.T. (2018). Learning Confidence for Out-of-Distribution Detection in Neural Network. arXiv.
25. Liu, Y., Chen, H., Wang, Y., and Zhang, P. (2021). Power pooling: An adaptive pooling function for weakly labelled sound event detection. arXiv.10.1109/IJCNN52387.2021.9533332 ↗
26. Mesaros, A., Heittola, T., and Virtanen, T. (2016). Metrics for Polyphonic Sound Event Detection. Appl. Sci., 6.10.3390/app6060162 ↗