Robust Partial Multi-Label Learning Under Dual Noise via Joint Subspace Learning

Abstract

Partial Multi-label Learning (PML) deals with the ambiguity where each instance is annotated with a set of candidate labels, and only a subset of which is valid. While existing PML methods focus primarily on label disambiguation, they often rely on the assumption of a clean feature space. However, in real-world applications, data are frequently plagued by the co-existence of label noise and feature noise, referred to as the dual noise challenge. Consequently, model robustness degrades substantially. To address this, we propose a framework named Ranking-Consistent Correntropy-based subspace learning for Partial Multi-label Learning (RCC-PML). Unlike existing dual noise PML methods that operate in the input space, our work introduces a subspace learning framework, where robust representation and semantic ranking are jointly optimized to enforce cross-space consistency. Specifically, we leverage the Maximum Correntropy Criterion (MCC) to construct robust scatter matrices, effectively suppressing heavy-tailed feature noise. To tackle label ambiguity, a ranking-consistent constraint is introduced to encourage a reasonable margin between ground-truth and false-positive labels in the projected subspace. Furthermore, we incorporate dual-graph regularization to preserve both the local manifold structure via anchor embedding and global semantic consistency. Finally,L_2,1-norm regularization is imposed on the projection matrix to perform adaptive feature selection. Extensive experiments on benchmark datasets demonstrate that the proposed method significantly outperforms state-of-the-art algorithms, particularly in heavy-tailed environments.

Keywords

feature noise / feature selection / maximum correntropy criterion / partial multi-label learning / subspace learning

Acknowledgements

This paper is supported by National Natural Science Foundation of China (Grant 62506223).

Conflicts of Interest

Duoqian Miao is an Associate Editor for the journal and Witold Pedrycz is an Editorial board member for the journal. They were not involved in peer review process or the decision to publish this article. The authors declare that they have no conflict of interest.

Data Availability Statement

The datasets analysed during the current study are publicly available. Specifically, they can be accessed from the Mulan multi-label learning library (https://mulan.sourceforge.net/datasets-mlc.html) and the SEU PALM multi-label dataset repository (https://palm.seu.edu.cn/zhangml/).

References

[1] M. Zhang and Z. Zhou, “A Review on Multi-Label Learning Algorithms,” IEEE Transactions on Knowledge and Data Engineering 26, no. 8 (2014): 1819-1837, https://doi.org/10.1109/TKDE.2013.39.

[2] W. Liu, H. Wang, X. Shen, and I. W. Tsang, “The Emerging Trends of Multi-Label Learning,” IEEE Transactions on Pattern Analysis and Machine Intelligence 44, no. 11 (2022): 7955-7974, https://doi.org/10.1109/TPAMI.2021.3119334.

[3] Y. Zhang, T. Zhao, D. Miao, and Y. Yao, “Three-Way Multi-Label Classification: A Review, a Framework, and New Challenges,” Applied Soft Computing 171 (2025): 112757, https://doi.org/10.1016/j.asoc.2025.112757.

[4] M. Xie and S. Huang, “Partial Multi-Label Learning,” in Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence, (AAAI-18), the 30th Innovative Applications of Artificial Intelligence (AAAI-18), and the 8th AAAI Symposium on Educational Advances in Artificial Intelligence (EAAI-18), eds. S. A. McIlraith and K. Q. Weinberger (AAAI Press, 2018), 4302-4309.

[5] Y. Hu, X. Fang, P. Kang, Y. Chen, Y. Fang, and S. Xie, “Dual Noise Elimination and Dynamic Label Correlation Guided Partial Multi-Label Learning,” IEEE Transactions on Multimedia 26 (2024): 5641-5656, https://doi.org/10.1109/TMM.2023.3338080.

[6] Z. Li, G. Lyu, and S. Feng, “Partial Multi-Label Learning via Multi-Subspace Representation,” in Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence, IJCAI 2020, ed. C. Bessiere (ijcai.org, 2020), 2612-2618.

[7] Y. Lin, Y. Li, S. Lin, L. Guo, and Y. Mao, “Partial Multi-Label Feature Selection Based on Label Distribution Learning,” Pattern Recognition 164 (2025): 111523, https://doi.org/10.1016/j.patcog.2025.111523.

[8] Q. Han, L. Hu, and W. Gao, “Integrating Label Confidence-Based Feature Selection for Partial Multi-Label Learning,” Pattern Recognition 161 (2025): 111281, https://doi.org/10.1016/j.patcog.2024.111281.

[9] T. Yin, H. Chen, Z. Yuan, et al. , “A Robust Multilabel Feature Selection Approach Based on Graph Structure Considering Fuzzy Dependency and Feature Interaction,” IEEE Transactions on Fuzzy Systems 31, no. 12 (2023): 4516-4528, https://doi.org/10.1109/TFUZZ.2023.3287193.

[10] Y. Wu, P. Li, and Y. Zou, “Partial Multi-Label Feature Selection With Feature Noise,” Pattern Recognition 162 (2025): 111310, https://doi.org/10.1016/j.patcog.2024.111310.

[11] H. Wang, S. Yang, G. Lyu, et al., Deep Partial Multi-Label Learning With Graph Disambiguation,” in Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence, IJCAI 2023, 19th-25th August 2023, Macao, SAR, China, ijcai.org (2023), 4308-4316, https://doi.org/10.24963/ijcai.2023/479.

[12] Y. Yang, H. Chen, Y. Mi, C. Luo, S. Horng, and T. Li, “Multi-Label Feature Selection Based on Stable Label Relevance and Label-Specific Features,” Information Sciences 648 (2023): 119525, https://doi.org/10.1016/j.ins.2023.119525.

[13] G. Lyu, B. Sun, X. Deng, and S. Feng, “Addressing Multi-Label Learning With Partial Labels: From Sample Selection to Label Selection,” in AAAI-25, Sponsored by the Association for the Advancement of Artificial Intelligence, February 25 - March 4, 2025, eds. T. Walsh, J. Shah, and Z. Kolter (AAAI Press, 2025), 19251-19259.

[14] F. Yang, Y. Jia, H. Liu, Y. Dong, and J. Hou, “Noisy Label Removal for Partial Multi-Label Learning,” in Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD, eds. R. Baeza-Yates and F. Bonchi (ACM; 2024, 2024), 3724-3735.

[15] X. Fang, X. Hu, Y. Hu, Y. Chen, S. Xie, and N. Han, “Fuzzy Bifocal Disambiguation for Partial Multi-Label Learning,” Neural Networks 185 (2025): 107137, https://doi.org/10.1016/j.neunet.2025.107137.

[16] G. Yu, X. Chen, C. Domeniconi, et al., Feature-Induced Partial Multi-Label Learning,” in IEEE International Conference on Data Mining, ICDM 2018, Singapore, November 17-20, 2018 (IEEE Computer Society, 2018), 1398-1403, https://doi.org/10.1109/ICDM.2018.00192.

[17] L. Sun, S. Feng, T. Wang, C. Lang, and Y. Jin, “Partial Multi-Label Learning by Low-Rank and Sparse Decomposition,” in Proceedings of the AAAI Conference on Artificial Intelligence vol. 33, no. 1 (AAAI Press, 2019), 5016-5023, https://doi.org/10.1609/aaai.v33i01.33015016.

[18] H. Pan, K. Liu, and W. Gao, “Reconsidering Feature Structure Information and Latent Space Alignment in Partial Multi-label Feature Selection,” in AAAI-25, Sponsored by the Association for the Advancement of Artificial Intelligence, eds. T. Walsh, J. Shah, and Z. Kolter (AAAI Press, 2025), 19786-19794.

[19] W. Qian, Y. Tu, J. Huang, W. Shu, and Y. Cheung, “Partial Multi-label Learning Using Noise-Tolerant Broad Learning System With Label Enhancement and Dimensionality Reduction,” IEEE Transactions on Neural Networks and Learning Systems 36, no. 2 (2025): 3758-3772, https://doi.org/10.1109/TNNLS.2024.3352285.

[20] K. Wang, Y. Guan, Y. Xie, et al. , “Partial Multi-Label Learning With Label and Classifier Correlations,” Information Sciences 712 (2025): 122101, https://doi.org/10.1016/j.ins.2025.122101.

[21] Y. Chen, Y. Wu, N. Han, X. Fang, B. Chen, and J. Wen, “ Partial Multi-Label Learning Based on Near-Far Neighborhood Label Enhancement and Nonlinear Guidance,” in Proceedings of the 32nd ACM International Conference on Multimedia, MM 2024 (eds), J. Cai, M. S. Kankanhalli, B. Prabhakaran, et al. (ACM, 2024), 3722-3731.

[22] N. Xu, Y. Wu, C. Qiao, Y. Ren, M. Zhang, and X. Geng, “Multi-View Partial Multi-Label Learning via Graph-Fusion-Based Label Enhancement,” IEEE Transactions on Knowledge and Data Engineering 35, no. 11 (2023): 11656-11667, https://doi.org/10.1109/TKDE.2022.3232482.

[23] J. Hang and M. Zhang, “Partial Multi-Label Learning With Probabilistic Graphical Disambiguation,” in Advances in Neural Information Processing Systems 36: Annual Conference on Neural Information Processing Systems 2023, NeurIPS, eds. A. Oh, T. Naumann, A. Globerson, K. Saenko, M. Hardt, and S. Levine (2023), 1339-1351.

[24] A. Zhou, B. Liu, J. Wang, and G. Tsoumakas, “Batch Selection for Multi-Label Classification Guided by Uncertainty and Dynamic Label Correlations,” in AAAI-25, Sponsored by the Association for the Advancement of Artificial Intelligence, eds. T. Walsh, J. Shah, and Z. Kolter (AAAI Press, 2025), 22902-22909.

[25] B. Liu, B. Jia, and M. Zhang, “Towards Enabling Binary Decomposition for Partial Multi-Label Learning,” IEEE Transactions on Pattern Analysis and Machine Intelligence 45, no. 11 (2023): 13203-13217, https://doi.org/10.1109/TPAMI.2023.3290797.

[26] Y. Fan, J. Liu, J. Tang, P. Liu, Y. Lin, and Y. Du, “Learning Correlation Information for Multi-Label Feature Selection,” Pattern Recognition 145 (2024): 109899, https://doi.org/10.1016/j.patcog.2023.109899.

[27] L. Ma, L. Hu, Y. Li, W. Ding, and W. Gao, “MI-MCF: A Mutual Information-Based Multilabel Causal Feature Selection,” IEEE Transactions on Neural Networks and Learning Systems 36, no. 6 (2025): 9864-9878, https://doi.org/10.1109/TNNLS.2025.3556128.

[28] Z. Sun, H. Xie, J. Liu, and Y. Yu, “Multi-Label Feature Selection via Adaptive Dual-Graph Optimization,” Expert Systems With Applications 243 (2024): 122884, https://doi.org/10.1016/j.eswa.2023.122884.

[29] Z. Sun, Z. Chen, J. Liu, Y. Chen, and Y. Yu, “Partial Multi-Label Feature Selection via Low-Rank and Sparse Factorization With Manifold Learning,” Knowledge-Based Systems 296 (2024): 111899, https://doi.org/10.1016/j.knosys.2024.111899.

[30] Q. Han, W. Zhang, C. Ma, J. He, and W. Gao, “Robust Multilabel Feature Selection With Label Enhancement for Fault Diagnosis,” IEEE Transactions on Systems, Man, and Cybernetics: Systems 55, no. 11 (2025): 7841-7850, https://doi.org/10.1109/TSMC.2025.3598796.

[31] P. Hao, K. Liu, and W. Gao, “Uncertainty-Aware Global-View Reconstruction for Multi-View Multi-Label Feature Selection,” in AAAI-25, Sponsored by the Association for the Advancement of Artificial Intelligence, eds. T. Walsh, J. Shah, and Z. Kolter (AAAI Press, 2025), 17068-17076.

[32] S. Zhang, Y. Li, P. Zhang, and W. Gao, “Exploring Multi-Label Feature Selection via Feature and Label Information Supplementation,” Engineering Applications of Artificial Intelligence 159 (2025): 111552, https://doi.org/10.1016/j.engappai.2025.111552.

[33] Y. Fan, P. Liu, and J. Liu, “Reconstructing Data Representation for Multi-Label Feature Selection,” Pattern Recognition 169 (2026): 111941, https://doi.org/10.1016/j.patcog.2025.111941.

[34] Q. Deng, N. Zhou, W. Luo, Y. Du, K. Shi, and B. Chen, “Correntropy Based Label Loss for Multi-Classification on Deep Neural Networks,” Neurocomputing 646 (2025): 130500, https://doi.org/10.1016/j.neucom.2025.130500.

[35] W. Wang, G. Wang, C. Hu, and K. C. Ho, “Robust Ellipse Fitting Based on Maximum Correntropy Criterion With Variable Center,” IEEE Transactions on Image Processing 32 (2023): 2520-2535, https://doi.org/10.1109/TIP.2023.3270026.

[36] S. Huang, J. Liu, G. Qian, and X. Wang, “A Proportionate Maximum Total Complex Correntropy Algorithm for Sparse Systems,” Circuits, Systems, and Signal Processing 43, no. 10 (2024): 6415-6436, https://doi.org/10.1007/s00034-024-02752-9.

[37] R. Yadav, K. Agarwal, and A. K. Gupta, “Constrained Maximum Correntropy Criterion Based Sparse Algorithm for Sparse Channel Estimation Against Noisy Input,” Signal, Image and Video Processing 19, no. 14 (2025): 1168, https://doi.org/10.1007/s11760-025-04704-5.

[38] J. Li, Q. Hu, X. Liu, and Y. Zhang, “Augmented Maximum Correntropy Criterion for Robust Geometric Perception,” IEEE Transactions on Robotics 40 (2024): 4705-4724, https://doi.org/10.1109/TRO.2024.3484608.

[39] Y. Tang, Y. Chien, and G. Qian, “Optimizing Kernel Width for New Risk-Sensitive Loss: A Generalized Algorithmic Approach,” Digital Signal Processing 154 (2024): 104655, https://doi.org/10.1016/j.dsp.2024.104655.

[40] M. Xiang and Z. Zhang, “Non-Signal Components Minimization for Sparse Signal Recovery,” Signal Processing 226 (2025): 109617, https://doi.org/10.1016/j.sigpro.2024.109617.

[41] W. Liu, J. He, and S. Chang, “Large Graph Construction for Scalable Semi-Supervised Learning,” in Proceedings of the 27th International Conference on Machine Learning (ICML-10), eds. J. Fürnkranz and T. Joachims (Omnipress, 2010), 679-686.

[42] L. Sun, S. Feng, G. Lyu, H. Zhang, and G. Dai, “Partial Multi-Label Learning With Noisy Side Information,” Knowledge and Information Systems 63, no. 2 (2021): 541-564, https://doi.org/10.1007/s10115-020-01527-3.

[43] M. Zhang and Z. Zhou, “ML-KNN: A Lazy Learning Approach to Multi-Label Learning,” Pattern Recognition 40, no. 7 (2007): 2038-2048, https://doi.org/10.1016/j.patcog.2006.12.019.

[44] J. Demsar, “Statistical Comparisons of Classifiers Over Multiple Data Sets,” Journal of Machine Learning Research 7 (2006): 1-30, https://jmlr.org/papers/v7/demsar06a.html.