Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
175 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
42 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Bias Amplification Enhances Minority Group Performance (2309.06717v2)

Published 13 Sep 2023 in cs.LG and cs.CY

Abstract: Neural networks produced by standard training are known to suffer from poor accuracy on rare subgroups despite achieving high accuracy on average, due to the correlations between certain spurious features and labels. Previous approaches based on worst-group loss minimization (e.g. Group-DRO) are effective in improving worse-group accuracy but require expensive group annotations for all the training samples. In this paper, we focus on the more challenging and realistic setting where group annotations are only available on a small validation set or are not available at all. We propose BAM, a novel two-stage training algorithm: in the first stage, the model is trained using a bias amplification scheme via introducing a learnable auxiliary variable for each training sample; in the second stage, we upweight the samples that the bias-amplified model misclassifies, and then continue training the same model on the reweighted dataset. Empirically, BAM achieves competitive performance compared with existing methods evaluated on spurious correlation benchmarks in computer vision and natural language processing. Moreover, we find a simple stopping criterion based on minimum class accuracy difference that can remove the need for group annotations, with little or no loss in worst-group accuracy. We perform extensive analyses and ablations to verify the effectiveness and robustness of our algorithm in varying class and group imbalance ratios.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (55)
  1. Invariant risk minimization. arXiv preprint arXiv:1907.02893, 2019.
  2. Recognition in terra incognita. In Proceedings of the European conference on computer vision (ECCV), pp.  456–473, 2018.
  3. Nuanced metrics for measuring unintended bias with real data for text classification. In Companion proceedings of the 2019 world wide web conference, pp.  491–500, 2019.
  4. What is the effect of importance weighting in deep learning? In International Conference on Machine Learning, pp.  872–881. PMLR, 2019.
  5. Heteroskedastic and imbalanced deep learning with adaptive regularization. arXiv preprint arXiv:2006.15766, 2020.
  6. High-performance neural networks for visual object classification. arXiv preprint arXiv:1102.0183, 2011.
  7. Don’t take the easy way out: Ensemble based methods for avoiding known dataset biases. arXiv preprint arXiv:1909.03683, 2019.
  8. Environment inference for invariant learning. In International Conference on Machine Learning, pp.  2189–2200. PMLR, 2021.
  9. BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pp.  4171–4186, Minneapolis, Minnesota, June 2019. Association for Computational Linguistics. doi: 10.18653/v1/N19-1423. URL https://aclanthology.org/N19-1423.
  10. Variance-based regularization with convex objectives. The Journal of Machine Learning Research, 20(1):2450–2504, 2019.
  11. Distributionally robust losses against mixture covariate shifts. Under review, 2, 2019.
  12. Shortcut learning in deep neural networks. Nature Machine Intelligence, 2(11):665–673, 2020. doi: 10.1038/s42256-020-00257-z.
  13. Model patching: Closing the subgroup performance gap with data augmentation. arXiv preprint arXiv:2008.06775, 2020.
  14. Annotation artifacts in natural language inference data. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 2 (Short Papers), pp.  107–112, New Orleans, Louisiana, June 2018. Association for Computational Linguistics. doi: 10.18653/v1/N18-2017. URL https://aclanthology.org/N18-2017.
  15. Learning from imbalanced data. IEEE Transactions on Knowledge and Data Engineering, 21(9):1263–1284, 2009. doi: 10.1109/TKDE.2008.239.
  16. Deep residual learning for image recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp.  770–778, 2016. doi: 10.1109/CVPR.2016.90.
  17. Simple and effective regularization methods for training on noisily labeled data with generalization guarantee. arXiv: Learning, 2020.
  18. Learning deep representation for imbalanced classification. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.  5375–5384, 2016.
  19. Selecmix: Debiased learning by contradicting-pair sampling. Advances in Neural Information Processing Systems, 35:14345–14357, 2022.
  20. Simple data balancing achieves competitive worst-group-accuracy. In Conference on Causal Learning and Reasoning, pp.  336–351. PMLR, 2022.
  21. Neural tangent kernel: Convergence and generalization in neural networks. Advances in neural information processing systems, 31, 2018.
  22. A conservative approach for unbiased learning on unknown biases. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.  16752–16760, 2022.
  23. Survey on deep learning with class imbalance. Journal of Big Data, 6(1):1–54, 2019.
  24. Cost-sensitive learning of deep feature representations from imbalanced data. IEEE transactions on neural networks and learning systems, 29(8):3573–3587, 2017.
  25. Maximum weighted loss discrepancy. arXiv preprint arXiv:1906.03518, 2019.
  26. Learning debiased classifier with biased committee. arXiv preprint arXiv:2206.10843, 2022.
  27. Last layer re-training is sufficient for robustness to spurious correlations. arXiv preprint arXiv:2204.02937, 2022.
  28. Wilds: A benchmark of in-the-wild distribution shifts. In Marina Meila and Tong Zhang (eds.), Proceedings of the 38th International Conference on Machine Learning, volume 139 of Proceedings of Machine Learning Research, pp.  5637–5664. PMLR, 18–24 Jul 2021. URL https://proceedings.mlr.press/v139/koh21a.html.
  29. Learning multiple layers of features from tiny images. 2009.
  30. Diversify and disambiguate: Learning from underspecified data. arXiv preprint arXiv:2202.03418, 2022.
  31. The surprising creativity of digital evolution: A collection of anecdotes from the evolutionary computation and artificial life research communities. Artificial life, 26(2):274–306, 2020.
  32. Large-scale methods for distributionally robust optimization. Advances in Neural Information Processing Systems, 33:8847–8860, 2020.
  33. Just train twice: Improving group robustness without training group information. In Marina Meila and Tong Zhang (eds.), Proceedings of the 38th International Conference on Machine Learning, volume 139 of Proceedings of Machine Learning Research, pp.  6781–6792. PMLR, 18–24 Jul 2021. URL https://proceedings.mlr.press/v139/liu21f.html.
  34. Deep learning face attributes in the wild. 2015 IEEE International Conference on Computer Vision (ICCV), pp.  3730–3738, 2015.
  35. Wilds: A benchmark of in-the-wild distribution shifts. arXiv preprint arXiv:2012.07421, 2020.
  36. Learning from failure: De-biasing classifier from biased classifier. Advances in Neural Information Processing Systems, 33:20673–20684, 2020.
  37. Spread spurious attribute: Improving worst-group accuracy with spurious attribute estimation. arXiv preprint arXiv:2204.02070, 2022.
  38. Distributionally robust language modeling. arXiv preprint arXiv:1909.02060, 2019.
  39. Gradient starvation: A learning proclivity in neural networks. Advances in Neural Information Processing Systems, 34:1256–1272, 2021.
  40. Distributionally robust neural networks for group shifts: On the importance of regularization for worst-case generalization. arXiv preprint arXiv:1911.08731, 2019.
  41. An investigation of why overparameterization exacerbates spurious correlations. In Hal Daumé III and Aarti Singh (eds.), Proceedings of the 37th International Conference on Machine Learning, volume 119 of Proceedings of Machine Learning Research, pp.  8346–8356. PMLR, 13–18 Jul 2020. URL https://proceedings.mlr.press/v119/sagawa20a.html.
  42. Grad-cam: Visual explanations from deep networks via gradient-based localization. In Proceedings of the IEEE international conference on computer vision, pp.  618–626, 2017.
  43. Unsupervised learning of debiased representations with pseudo-attributes. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.  16742–16751, 2022.
  44. Meta-weight-net: Learning an explicit mapping for sample weighting. Advances in neural information processing systems, 32, 2019.
  45. No subclass left behind: Fine-grained robustness in coarse-grained classification problems. Advances in Neural Information Processing Systems, 33:19339–19352, 2020.
  46. Robust representation learning via perceptual similarity metrics. In International Conference on Machine Learning, pp.  10043–10053. PMLR, 2021.
  47. Data imbalance in classification: Experimental evaluation. Information Sciences, 513:429–441, 2020.
  48. Towards debiasing nlu models from unknown biases. arXiv preprint arXiv:2009.12303, 2020.
  49. The caltech-ucsd birds-200-2011 dataset. 2011.
  50. A broad-coverage challenge corpus for sentence understanding through inference. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers), pp.  1112–1122, New Orleans, Louisiana, June 2018. Association for Computational Linguistics. doi: 10.18653/v1/N18-1101. URL https://aclanthology.org/N18-1101.
  51. Increasing robustness to spurious correlations using forgettable examples. arXiv preprint arXiv:1911.03861, 2019.
  52. Improving out-of-distribution robustness via selective augmentation. arXiv preprint arXiv:2201.00299, 2022.
  53. mixup: Beyond empirical risk minimization. arXiv preprint arXiv:1710.09412, 2017.
  54. Coping with label shift via distributionally robust optimisation. arXiv preprint arXiv:2010.12230, 2020.
  55. Correct-n-contrast: A contrastive approach for improving robustness to spurious correlations. arXiv preprint arXiv:2203.01517, 2022.
Citations (3)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now