MCC: Multi-Centroid Curriculum for Pre-defined Centroid-based Classifiers

• Nattan Gomes Silva Cortes UFRJ
• Natália Azeredo Viana da Silva UFRJ
• Rogério Ferreira de Moraes UFRJ

DOI: https://doi.org/10.5753/eniac.2026.30862

Resumo

Em problemas de classificação, métodos que fixam os centros de classe na hiperesfera, como PEDCC e POD, supõem que cada classe cabe em um único centro, mas muitas reúnem subgrupos visuais distintos. Propomos o Multi-Centroid Curriculum (MCC): um currículo adaptativo que cria e remove subcentros por classe durante o treino, guiado pela complexidade dos dados; com um único centro, ele reproduz a formulação dos métodos originais. Em CIFAR-10/100 (ResNet-18 e ShuffleNet), o MCC supera o baseline POD em todas as configurações avaliadas (com vitórias em 20 de 20 execuções). Uma análise controlada demonstra que o ganho provém especificamente do currículo adaptativo, não de reinícios da taxa de aprendizado ou de um número fixo de centros.

Referências

Allen, K. R., Shelhamer, E., Shin, H., and Tenenbaum, J. B. (2019). Infinite Mixture Prototypes for Few-Shot Learning. In Proceedings of the 36th International Conference on Machine Learning (ICML), pages 232–241. PMLR.

Deng, J., Guo, J., Liu, T., Gong, M., and Zafeiriou, S. (2020). Sub-center ArcFace: Boosting Face Recognition by Large-scale Noisy Web Faces. In European Conference on Computer Vision (ECCV), pages 741–757. Springer.

Deng, J., Guo, J., Xue, N., and Zafeiriou, S. (2019). ArcFace: Additive angular margin loss for deep face recognition. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4690–4699.

He, K., Zhang, X., Ren, S., and Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR), pages 770–778.

Holm, S. (1979). A Simple Sequentially Rejective Multiple Test Procedure. Scandinavian Journal of Statistics, 6(2):65–70.

Krizhevsky, A. and Hinton, G. (2009). Learning Multiple Layers of Features from Tiny Images. Technical report, University of Toronto.

Liu, W., Wen, Y., Yu, Z., Li, M., Raj, B., and Song, L. (2017). SphereFace: Deep hypersphere embedding for face recognition. In Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), pages 512–520.

Ma, N., Zhang, X., Zheng, H.-T., and Sun, J. (2018). ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design. In Proceedings of the European Conference on Computer Vision (ECCV), pages 116–131.

Qian, Q., Shang, L., Sun, B., Hu, J., Li, H., and Jin, R. (2019). SoftTriple Loss: Deep Metric Learning Without Triplet Sampling. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 6450–6458.

Shen, Y., Sun, X., and Wei, X.-S. (2023). Equiangular Basis Vectors. arXiv preprint arXiv:2303.11637.

Snell, J., Swersky, K., and Zemel, R. (2017). Prototypical Networks for Few-shot Learning. In Advances in Neural Information Processing Systems (NeurIPS), volume 30, pages 4077–4087.

Wang, F., Cheng, J., Liu, W., and Liu, H. (2018). Additive Margin Softmax for Face Verification. IEEE Signal Processing Letters, 25(6):926–930.

Wen, Y., Zhang, K., Li, Z., and Qiao, Y. (2016). A Discriminative Feature Learning Approach for Deep Face Recognition. In Proceedings of the European Conference on Computer Vision (ECCV), pages 499–515.

Zhu, Q., Zhang, J., Wang, R., and Zheng, J. (2020). A New Loss Function for CNN Classifier Based on Pre-Defined Evenly-Distributed Class Centroids. IEEE Access, 8:1084–1092.

Zhu, Q. and Zu, X. (2023). A Softmax-Free Loss Function Based on Predefined Optimal-Distribution of Latent Features for Deep Learning Classifier. IEEE Transactions on Circuits and Systems for Video Technology, 33(3):1386–1397.