Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
102 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Pursuing Feature Separation based on Neural Collapse for Out-of-Distribution Detection (2405.17816v1)

Published 28 May 2024 in cs.CV and cs.LG

Abstract: In the open world, detecting out-of-distribution (OOD) data, whose labels are disjoint with those of in-distribution (ID) samples, is important for reliable deep neural networks (DNNs). To achieve better detection performance, one type of approach proposes to fine-tune the model with auxiliary OOD datasets to amplify the difference between ID and OOD data through a separation loss defined on model outputs. However, none of these studies consider enlarging the feature disparity, which should be more effective compared to outputs. The main difficulty lies in the diversity of OOD samples, which makes it hard to describe their feature distribution, let alone design losses to separate them from ID features. In this paper, we neatly fence off the problem based on an aggregation property of ID features named Neural Collapse (NC). NC means that the penultimate features of ID samples within a class are nearly identical to the last layer weight of the corresponding class. Based on this property, we propose a simple but effective loss called OrthLoss, which binds the features of OOD data in a subspace orthogonal to the principal subspace of ID features formed by NC. In this way, the features of ID and OOD samples are separated by different dimensions. By optimizing the feature separation loss rather than purely enlarging output differences, our detection achieves SOTA performance on CIFAR benchmarks without any additional data augmentation or sampling, demonstrating the importance of feature separation in OOD detection. The code will be published.

PDF HTML Abstract
Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Bookmark Chat Bubble Oval Streamline Icon: https://streamlinehq.com Chat (Pro)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (58)
  1. Neco: Neural collapse based out-of-distribution detection. arXiv preprint arXiv:2310.06823, 2023.
  2. Feed two birds with one scone: Exploiting wild data for both out-of-distribution generalization and detection. In International Conference on Machine Learning, pages 1454–1471. PMLR, 2023.
  3. A benchmark of medical out of distribution detection. arxiv 2020. arXiv preprint arXiv:2007.04250, 2007.
  4. Contrastive test-time adaptation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 295–305, 2022.
  5. Atom: Robustifying out-of-distribution detection using outlier mining. In Machine Learning and Knowledge Discovery in Databases. Research Track: European Conference, ECML PKDD 2021, Bilbao, Spain, September 13–17, 2021, Proceedings, Part III 21, pages 430–445. Springer, 2021.
  6. Waic, but why? generative ensembles for robust anomaly detection. arXiv preprint arXiv:1810.01392, 2018.
  7. Describing textures in the wild. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3606–3613, 2014.
  8. Outlier detection through null space analysis of neural networks. arXiv preprint arXiv:2007.01263, 2020.
  9. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009.
  10. Unifying low dimensional observations in deep learning through the deep linear unconstrained feature model. arXiv preprint arXiv:2404.06106, 2024.
  11. Knowledge distillation: A survey. International Journal of Computer Vision, 129(6):1789–1819, 2021.
  12. Linking neural collapse and l2 normalization with improved out-of-distribution detection in deep neural networks. arXiv preprint arXiv:2209.08378, 2022.
  13. A baseline for detecting misclassified and out-of-distribution examples in neural networks. arXiv preprint arXiv:1610.02136, 2016.
  14. Deep anomaly detection with outlier exposure. arXiv preprint arXiv:1812.04606, 2018.
  15. On the importance of gradients for detecting distributional shifts in the wild. Advances in Neural Information Processing Systems, 34:677–689, 2021.
  16. Limitations of neural collapse for understanding generalization in deep learning. arXiv preprint arXiv:2202.08384, 2022.
  17. How useful are gradients for ood detection really? arXiv preprint arXiv:2205.10439, 2022.
  18. Revisiting flow generative models for out-of-distribution detection. In International Conference on Learning Representations, 2021.
  19. Dos: Diverse outlier sampling for out-of-distribution detection. arXiv preprint arXiv:2306.02031, 2023.
  20. Supervised contrastive learning. Advances in neural information processing systems, 33:18661–18673, 2020.
  21. Glow: Generative flow with invertible 1x1 convolutions. Advances in neural information processing systems, 31, 2018.
  22. Normalizing flows: An introduction and review of current methods. IEEE transactions on pattern analysis and machine intelligence, 43(11):3964–3979, 2020.
  23. Vignesh Kothapalli. Neural collapse: A review on modelling principles and generalization. arXiv preprint arXiv:2206.04041, 2022.
  24. Learning multiple layers of features from tiny images. 2009.
  25. Solomon Kullback. Information theory and statistics. Courier Corporation, 1997.
  26. Contrastive representation learning: A framework and review. Ieee Access, 8:193907–193934, 2020.
  27. Probing the purview of neural networks via gradient analysis. IEEE Access, 11:32716–32732, 2023.
  28. A simple unified framework for detecting out-of-distribution samples and adversarial attacks. Advances in neural information processing systems, 31, 2018.
  29. A comprehensive survey on test-time adaptation under distribution shifts. arXiv preprint arXiv:2303.15361, 2023.
  30. Enhancing the reliability of out-of-distribution image detection in neural networks. arXiv preprint arXiv:1706.02690, 2017.
  31. Towards out-of-distribution generalization: A survey. arXiv preprint arXiv:2108.13624, 2021.
  32. Detecting out-of-distribution through the lens of neural collapse. arXiv preprint arXiv:2311.01479, 2023.
  33. Energy-based out-of-distribution detection. Advances in neural information processing systems, 33:21464–21475, 2020.
  34. Gran: An efficient gradient-norm based detector for adversarial and misclassified examples. arXiv preprint arXiv:2004.09179, 2020.
  35. Poem: Out-of-distribution detection with posterior sampling. In International Conference on Machine Learning, pages 15650–15665. PMLR, 2022.
  36. How to exploit hyperspherical embeddings for out-of-distribution detection? arXiv preprint arXiv:2203.04450, 2022.
  37. Out-of-distribution detection with subspace techniques and probabilistic modeling of features. arXiv preprint arXiv:2012.04250, 2020.
  38. Reading digits in natural images with unsupervised feature learning. In NIPS workshop on deep learning and unsupervised feature learning, volume 2011, page 7. Granada, Spain, 2011.
  39. Efficient test-time model adaptation without forgetting. In International conference on machine learning, pages 16888–16905. PMLR, 2022.
  40. Prevalence of neural collapse during the terminal phase of deep learning training. Proceedings of the National Academy of Sciences, 117(40):24652–24663, 2020.
  41. Feature learning in deep classifiers through intermediate neural collapse. In International Conference on Machine Learning, pages 28729–28745. PMLR, 2023.
  42. Gradient-based novelty detection boosted by self-supervised binary classification. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, pages 8370–8377, 2022.
  43. React: Out-of-distribution detection with rectified activations. Advances in Neural Information Processing Systems, 34:144–157, 2021.
  44. Out-of-distribution detection with deep nearest neighbors. In International Conference on Machine Learning, pages 20827–20840. PMLR, 2022.
  45. Csi: Novelty detection via contrastive learning on distributionally shifted instances. Advances in neural information processing systems, 33:11839–11852, 2020.
  46. 80 million tiny images: A large data set for nonparametric object and scene recognition. IEEE transactions on pattern analysis and machine intelligence, 30(11):1958–1970, 2008.
  47. Vim: Out-of-distribution with virtual-logit matching. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4921–4930, 2022.
  48. Learning to augment distributions for out-of-distribution detection. Advances in Neural Information Processing Systems, 36, 2024.
  49. Out-of-distribution detection with implicit outlier transformation. arXiv preprint arXiv:2303.05033, 2023.
  50. Low-dimensional gradient helps out-of-distribution detection. arXiv preprint arXiv:2310.17163, 2023.
  51. Turkergaze: Crowdsourcing saliency with webcam based eye tracking. arXiv preprint arXiv:1504.06755, 2015.
  52. Lsun: Construction of a large-scale image dataset using deep learning with humans in the loop. arXiv preprint arXiv:1506.03365, 2015.
  53. Wide residual networks. arXiv preprint arXiv:1605.07146, 2016.
  54. Epa: Neural collapse inspired robust out-of-distribution detector. arXiv preprint arXiv:2401.01710, 2024.
  55. Out-of-distribution detection learning with unreliable out-of-distribution sources. Advances in Neural Information Processing Systems, 36, 2024.
  56. Places: A 10 million image database for scene recognition. IEEE transactions on pattern analysis and machine intelligence, 40(6):1452–1464, 2017.
  57. Boosting out-of-distribution detection with typical features. arXiv preprint arXiv:2210.04200, 2022.
  58. Ev Zisselman and Aviv Tamar. Deep residual flow for out of distribution detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 13994–14003, 2020.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (5)
  1. Yingwen Wu (12 papers)
  2. Ruiji Yu (3 papers)
  3. Xinwen Cheng (9 papers)
  4. Zhengbao He (13 papers)
  5. Xiaolin Huang (101 papers)
Citations (1)
View on Semantic Scholar
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets