论文信息 - The Impact of Label Noise on a Music Tagger

The Impact of Label Noise on a Music Tagger

We explore how much can be learned from noisy labels in audio music tagging. Our experiments show that carefully annotated labels result in highest figures of merit, but even high amounts of noisy labels contain enough information for successful learning. Artificial corruption of curated data allows us to quantize this contribution of noisy labels.

[1] Rob Fergus,et al. Learning from Noisy Labels with Deep Neural Networks , 2014, ICLR.

[2] Li Fei-Fei,et al. MentorNet: Learning Data-Driven Curriculum for Very Deep Neural Networks on Corrupted Labels , 2017, ICML.

[3] Yale Song,et al. Learning from Noisy Labels with Distillation , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[4] Nagarajan Natarajan,et al. Learning with Noisy Labels , 2013, NIPS.

[5] Daniel P. W. Ellis,et al. Learning Sound Event Classifiers from Web Audio with Noisy Labels , 2019, ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[6] Daniel P. W. Ellis,et al. Audio tagging with noisy labels and minimal supervision , 2019, DCASE.

[7] Mert R. Sabuncu,et al. Generalized Cross Entropy Loss for Training Deep Neural Networks with Noisy Labels , 2018, NeurIPS.

[8] Fabian Paischer,et al. AUDIO TAGGING WITH CONVOLUTIONAL NEURAL NETWORKS TRAINED WITH NOISY DATA Technical Report , 2019 .

[10] Bhiksha Raj,et al. A Closer Look at Weak Label Learning for Audio Events , 2018, ArXiv.

[11] Xavier Serra,et al. Freesound Datasets: A Platform for the Creation of Open Audio Datasets , 2017, ISMIR.

[12] Dacheng Tao,et al. Classification with Noisy Labels by Importance Reweighting , 2014, IEEE Transactions on Pattern Analysis and Machine Intelligence.