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

Learning Prototype Classifiers for Long-Tailed Recognition (2302.00491v3)

Published 1 Feb 2023 in cs.CV and cs.LG

Abstract: The problem of long-tailed recognition (LTR) has received attention in recent years due to the fundamental power-law distribution of objects in the real-world. Most recent works in LTR use softmax classifiers that are biased in that they correlate classifier norm with the amount of training data for a given class. In this work, we show that learning prototype classifiers addresses the biased softmax problem in LTR. Prototype classifiers can deliver promising results simply using Nearest-Class- Mean (NCM), a special case where prototypes are empirical centroids. We go one step further and propose to jointly learn prototypes by using distances to prototypes in representation space as the logit scores for classification. Further, we theoretically analyze the properties of Euclidean distance based prototype classifiers that lead to stable gradient-based optimization which is robust to outliers. To enable independent distance scales along each channel, we enhance Prototype classifiers by learning channel-dependent temperature parameters. Our analysis shows that prototypes learned by Prototype classifiers are better separated than empirical centroids. Results on four LTR benchmarks show that Prototype classifier outperforms or is comparable to state-of-the-art methods. Our code is made available at https://github.com/saurabhsharma1993/prototype-classifier-ltr.

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 (39)
  1. Long-tailed recognition via weight balancing. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 6897–6907, 2022.
  2. Learning imbalanced datasets with label-distribution-aware margin loss. In NeurIPS, 2019.
  3. Smote: synthetic minority over-sampling technique. JAIR, 2002.
  4. Smoteboost: Improving prediction of the minority class in boosting. In European conference on principles of data mining and knowledge discovery, 2003.
  5. Autoaugment: Learning augmentation policies from data. arXiv preprint arXiv:1805.09501, 2018.
  6. Randaugment: Practical automated data augmentation with a reduced search space. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops, pages 702–703, 2020.
  7. Class-balanced loss based on effective number of samples. In CVPR, 2019.
  8. Parametric contrastive learning. In Proceedings of the IEEE/CVF international conference on computer vision, pages 715–724, 2021.
  9. Imagenet: A large-scale hierarchical image database. In CVPR, 2009.
  10. Improved regularization of convolutional neural networks with cutout. arXiv preprint arXiv:1708.04552, 2017.
  11. C4. 5, class imbalance, and cost sensitivity: why under-sampling beats over-sampling. In Workshop on learning from imbalanced datasets II, 2003.
  12. A multiple resampling method for learning from imbalanced data sets. Computational intelligence, 2004.
  13. Exploring deep neural networks via layer-peeled model: Minority collapse in imbalanced training. Proceedings of the National Academy of Sciences, 118(43):e2103091118, 2021.
  14. Deepncm: Deep nearest class mean classifiers. In International Conference on Learning Representations Workshop, 2018.
  15. Learning from imbalanced data. TKDE, 2009.
  16. Momentum contrast for unsupervised visual representation learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9729–9738, 2020.
  17. Decoupling representation and classifier for long-tailed recognition. In Eighth International Conference on Learning Representations (ICLR), 2020.
  18. Focal loss for dense object detection. In ICCV, 2017.
  19. Large-scale long-tailed recognition in an open world. In CVPR, 2019.
  20. Gistnet: a geometric structure transfer network for long-tailed recognition. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 8209–8218, 2021.
  21. Sgdr: Stochastic gradient descent with warm restarts. In International Conference on Learning Representations, 2017.
  22. Long-tail learning via logit adjustment. In International Conference on Learning Representations, 2021.
  23. Distance-based image classification: Generalizing to new classes at near-zero cost. TPAMI, 2013.
  24. Fernando Nogueira et al. Bayesian optimization: Open source constrained global optimization tool for python. URL https://github. com/fmfn/BayesianOptimization, 2014.
  25. Prevalence of neural collapse during the terminal phase of deep learning training. Proceedings of the National Academy of Sciences, 117(40):24652–24663, 2020.
  26. icarl: Incremental classifier and representation learning. In Proceedings of the IEEE conference on Computer Vision and Pattern Recognition, 2017.
  27. Balanced meta-softmax for long-tailed visual recognition. Advances in Neural Information Processing Systems, 33:4175–4186, 2020.
  28. Distributional robustness loss for long-tail learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021.
  29. Long-tailed recognition using class-balanced experts. In DAGM German Conference on Pattern Recognition. Springer, 2020.
  30. Prototypical networks for few-shot learning. In NeurIPS, 2017.
  31. Meta-transfer learning for few-shot learning. In CVPR, 2019.
  32. Imbalance trouble: Revisiting neural-collapse geometry. In Advances in Neural Information Processing Systems, 2022.
  33. The devil is in the tails: Fine-grained classification in the wild. arXiv preprint arXiv:1709.01450, 2017.
  34. The inaturalist species classification and detection dataset. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 8769–8778, 2018.
  35. Long-tailed recognition by routing diverse distribution-aware experts. In International Conference on Learning Representations, 2020.
  36. f-vaegan-d2: A feature generating framework for any-shot learning. In CVPR, 2019.
  37. Learning from multiple experts: Self-paced knowledge distillation for long-tailed classification. In European Conference on Computer Vision. Springer, 2020.
  38. Distribution alignment: A unified framework for long-tail visual recognition. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 2361–2370, 2021.
  39. Bbn: Bilateral-branch network with cumulative learning for long-tailed visual recognition. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 9719–9728, 2020.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (4)
  1. Saurabh Sharma (90 papers)
  2. Yongqin Xian (33 papers)
  3. Ning Yu (78 papers)
  4. Ambuj Singh (34 papers)
Citations (11)
View on Semantic Scholar