Toward Robustness in Multi-Label Classification: A Data Augmentation Strategy Against Imbalance and Noise

Abstract

Multi-label classification poses challenges due to imbalanced and noisy labels in training data. In this paper, we propose a unified data augmentation method, named BalanceMix, to address these challenges. Our approach includes two samplers for imbalanced labels, generating minority-augmented instances with high diversity. It also refines multi-labels at the label-wise granularity, categorizing noisy labels as clean, re-labeled, or ambiguous for robust optimization. Extensive experiments on three benchmark datasets demonstrate that BalanceMix outperforms existing state-of-the-art methods. We release the code at https://github.com/DISL-Lab/BalanceMix.

Cite

Text

Song et al. "Toward Robustness in Multi-Label Classification: A Data Augmentation Strategy Against Imbalance and Noise." AAAI Conference on Artificial Intelligence, 2024. doi:10.1609/AAAI.V38I19.30157

Markdown

[Song et al. "Toward Robustness in Multi-Label Classification: A Data Augmentation Strategy Against Imbalance and Noise." AAAI Conference on Artificial Intelligence, 2024.](https://mlanthology.org/aaai/2024/song2024aaai-robustness/) doi:10.1609/AAAI.V38I19.30157

BibTeX

@inproceedings{song2024aaai-robustness,
  title     = {{Toward Robustness in Multi-Label Classification: A Data Augmentation Strategy Against Imbalance and Noise}},
  author    = {Song, Hwanjun and Kim, Minseok and Lee, Jae-Gil},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2024},
  pages     = {21592-21601},
  doi       = {10.1609/AAAI.V38I19.30157},
  url       = {https://mlanthology.org/aaai/2024/song2024aaai-robustness/}
}