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

A noisy elephant in the room: Is your out-of-distribution detector robust to label noise? (2404.01775v1)

Published 2 Apr 2024 in cs.CV, cs.AI, and cs.LG

Abstract: The ability to detect unfamiliar or unexpected images is essential for safe deployment of computer vision systems. In the context of classification, the task of detecting images outside of a model's training domain is known as out-of-distribution (OOD) detection. While there has been a growing research interest in developing post-hoc OOD detection methods, there has been comparably little discussion around how these methods perform when the underlying classifier is not trained on a clean, carefully curated dataset. In this work, we take a closer look at 20 state-of-the-art OOD detection methods in the (more realistic) scenario where the labels used to train the underlying classifier are unreliable (e.g. crowd-sourced or web-scraped labels). Extensive experiments across different datasets, noise types & levels, architectures and checkpointing strategies provide insights into the effect of class label noise on OOD detection, and show that poor separation between incorrectly classified ID samples vs. OOD samples is an overlooked yet important limitation of existing methods. Code: https://github.com/glhr/ood-labelnoise

Definition Search Book Streamline Icon: https://streamlinehq.com
References (71)
  1. Image classification with deep learning in the presence of noisy labels: A survey. Knowledge-Based Systems, 215:106771, 2021.
  2. The evolution of out-of-distribution robustness throughout fine-tuning. Transactions on Machine Learning Research, 2022.
  3. Towards open set deep networks. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 1563–1572, 2016.
  4. In or out? fixing imagenet out-of-distribution detection evaluation. In ICML, 2023.
  5. Food-101 – mining discriminative components with random forests. In European Conference on Computer Vision, 2014.
  6. Describing textures in the wild. In 2014 IEEE Conference on Computer Vision and Pattern Recognition, pages 3606–3613, 2014.
  7. An optimal transportation approach for assessing almost stochastic order. In The Mathematics of the Uncertain, pages 33–44. Springer, 2018.
  8. Imagenet: A large-scale hierarchical image database. In 2009 IEEE Conference on Computer Vision and Pattern Recognition, pages 248–255, 2009.
  9. Li Deng. The mnist database of handwritten digit images for machine learning research [best of the web]. IEEE Signal Processing Magazine, 29(6):141–142, 2012.
  10. Extremely simple activation shaping for out-of-distribution detection. In The Eleventh International Conference on Learning Representations, 2023.
  11. Deep dominance - how to properly compare deep neural models. In Proceedings of the 57th Conference of the Association for Computational Linguistics, ACL 2019, Florence, Italy, July 28- August 2, 2019, Volume 1: Long Papers, pages 2773–2785. Association for Computational Linguistics, 2019.
  12. Exploring the limits of out-of-distribution detection. In Advances in Neural Information Processing Systems, 2021.
  13. Classification in the presence of label noise: A survey. IEEE Transactions on Neural Networks and Learning Systems, 25(5):845–869, 2014.
  14. A framework for benchmarking class-out-of-distribution detection and its application to imagenet. In The Eleventh International Conference on Learning Representations, 2023.
  15. On calibration of modern neural networks. In Proceedings of the 34th International Conference on Machine Learning, pages 1321–1330. PMLR, 2017.
  16. Escaping the big data paradigm with compact transformers. 2021.
  17. Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016.
  18. Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2019.
  19. A baseline for detecting misclassified and out-of-distribution examples in neural networks. In 5th International Conference on Learning Representations, ICLR 2017, Toulon, France, April 24-26, 2017, Conference Track Proceedings. OpenReview.net, 2017.
  20. Using pre-training can improve model robustness and uncertainty. In Proceedings of the 36th International Conference on Machine Learning, pages 2712–2721. PMLR, 2019.
  21. Scaling out-of-distribution detection for real-world settings. In Proceedings of the 39th International Conference on Machine Learning, pages 8759–8773. PMLR, 2022a.
  22. Scaling out-of-distribution detection for real-world settings. ICML, 2022b.
  23. Mos: Towards scaling out-of-distribution detection for large semantic space. In 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 8706–8715, Los Alamitos, CA, USA, 2021. IEEE Computer Society.
  24. On the importance of gradients for detecting distributional shifts in the wild. In Advances in Neural Information Processing Systems, 2021.
  25. Beyond auroc & co. for evaluating out-of-distribution detection performance. In 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pages 3881–3890, 2023.
  26. Paul F Jaeger et al. A call to reflect on evaluation practices for failure detection in image classification. In ICLR, 2023.
  27. Beyond synthetic noise: Deep learning on controlled noisy labels. In Proceedings of the 37th International Conference on Machine Learning, pages 4804–4815. PMLR, 2020.
  28. Deep learning with noisy labels: Exploring techniques and remedies in medical image analysis. Medical Image Analysis, 65:101759, 2020.
  29. Fine samples for learning with noisy labels. In Advances in Neural Information Processing Systems, pages 24137–24149. Curran Associates, Inc., 2021.
  30. Pytorch-ood: A library for out-of-distribution detection based on pytorch. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 4351–4360, 2022.
  31. Learning multiple layers of features from tiny images. Technical Report 0, University of Toronto, Toronto, Ontario, 2009.
  32. A simple unified framework for detecting out-of-distribution samples and adversarial attacks. In Advances in Neural Information Processing Systems. Curran Associates, Inc., 2018.
  33. Rethinking out-of-distribution (ood) detection: Masked image modeling is all you need. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 11578–11589, 2023.
  34. Gradient descent with early stopping is provably robust to label noise for overparameterized neural networks. In Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics, pages 4313–4324. PMLR, 2020.
  35. Enhancing the reliability of out-of-distribution image detection in neural networks. In International Conference on Learning Representations, 2018.
  36. Early-learning regularization prevents memorization of noisy labels. Advances in Neural Information Processing Systems, 33, 2020a.
  37. Energy-based out-of-distribution detection. In Advances in Neural Information Processing Systems, pages 21464–21475. Curran Associates, Inc., 2020b.
  38. Fine-grained visual classification of aircraft. Technical report, 2013.
  39. Reading Digits in Natural Images with Unsupervised Feature Learning. In NIPS Workshop on Deep Learning and Unsupervised Feature Learning 2011, 2011.
  40. Robustness and reliability when training with noisy labels. In Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, pages 922–942. PMLR, 2022.
  41. Robustness to label noise depends on the shape of the noise distribution. In Advances in Neural Information Processing Systems, 2022.
  42. Towards robust uncertainty estimation in the presence of noisy labels. In Artificial Neural Networks and Machine Learning – ICANN 2022, pages 673–684, Cham, 2022. Springer International Publishing.
  43. Identifying mislabeled data using the area under the margin ranking. In Advances in Neural Information Processing Systems, pages 17044–17056. Curran Associates, Inc., 2020.
  44. A simple fix to mahalanobis distance for improving near-ood detection. In ICML 2021 Workshop on Uncertainty and Robustness in Deep Learning, 2021.
  45. Deep learning is robust to massive label noise, 2018.
  46. Evidentialmix: Learning with combined open-set and closed-set noisy labels. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 3607–3615, 2021.
  47. A unified survey on anomaly, novelty, open-set, and out of-distribution detection: Solutions and future challenges. Transactions on Machine Learning Research, 2022.
  48. Detecting out-of-distribution examples with Gram matrices. In Proceedings of the 37th International Conference on Machine Learning, pages 8491–8501. PMLR, 2020.
  49. Analyzing the robustness of open-world machine learning. In Proceedings of the 12th ACM Workshop on Artificial Intelligence and Security, page 105–116, New York, NY, USA, 2019. Association for Computing Machinery.
  50. SELFIE: Refurbishing unclean samples for robust deep learning. In ICML, 2019.
  51. Learning from noisy labels with deep neural networks: A survey. IEEE Transactions on Neural Networks and Learning Systems, pages 1–19, 2022a.
  52. Deep metric learning via lifted structured feature embedding. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016.
  53. Rankfeat: Rank-1 feature removal for out-of-distribution detection. In Advances in Neural Information Processing Systems, 2022b.
  54. Utility data annotation with amazon mechanical turk. In 2008 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, pages 1–8, 2008.
  55. Dice: Leveraging sparsification for out-of-distribution detection. In Computer Vision – ECCV 2022, pages 691–708, Cham, 2022. Springer Nature Switzerland.
  56. React: Out-of-distribution detection with rectified activations. In Advances in Neural Information Processing Systems, pages 144–157. Curran Associates, Inc., 2021.
  57. Out-of-distribution detection with deep nearest neighbors. In Proceedings of the 39th International Conference on Machine Learning, pages 20827–20840. PMLR, 2022.
  58. MLP-mixer: An all-MLP architecture for vision. In Advances in Neural Information Processing Systems, 2021.
  59. deep-significance-easy and meaningful statistical significance testing in the age of neural networks. arXiv preprint arXiv:2204.06815, 2022.
  60. Open-set recognition: A good closed-set classifier is all you need. In International Conference on Learning Representations, 2022.
  61. The caltech-ucsd birds-200-2011 dataset. Technical Report CNS-TR-2011-001, California Institute of Technology, 2011.
  62. Vim: Out-of-distribution with virtual-logit matching. In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 4911–4920, Los Alamitos, CA, USA, 2022. IEEE Computer Society.
  63. Mitigating memorization of noisy labels by clipping the model prediction. In Proceedings of the 40th International Conference on Machine Learning. JMLR.org, 2023.
  64. Learning with noisy labels revisited: A study using real-world human annotations. In International Conference on Learning Representations, 2022.
  65. Learning from massive noisy labeled data for image classification. In 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 2691–2699, 2015.
  66. Generalized out-of-distribution detection: A survey. arXiv preprint arXiv:2110.11334, 2021.
  67. OpenOOD: Benchmarking generalized out-of-distribution detection. In Thirty-sixth Conference on Neural Information Processing Systems Datasets and Benchmarks Track, 2022.
  68. Adversarial examples: Attacks and defenses for deep learning. IEEE Transactions on Neural Networks and Learning Systems, 30(9):2805–2824, 2019.
  69. Understanding deep learning requires rethinking generalization. In International Conference on Learning Representations, 2017.
  70. Out-of-distribution detection based on in-distribution data patterns memorization with modern hopfield energy. In The Eleventh International Conference on Learning Representations, 2023a.
  71. Openood v1.5: Enhanced benchmark for out-of-distribution detection. arXiv preprint arXiv:2306.09301, 2023b.
Citations (4)
View on Semantic Scholar

Summary

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

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