噪声标签学习（Label-noise Learning）是人工智能领域的一个研究方向，主要关注在训练数据中存在标签错误或噪声的情况下，如何有效地训练准确的机器学习模型。在现实世界的数据收集过程中，由于人为因素、自动标注、不完全的数据等原因，训练数据中的标签可能会带有一定的错误，这会对模型的训练和性能产生负面影响。噪声标签学习的目标是使得模型在存在噪声标签的情况下仍然能够获得良好的泛化性能，提高模型对噪声的鲁棒性。

相关成果：

[1] Yang Lu, Yiliang Zhang, Bo Han, Yiu-ming Cheung, Hanzi Wang, “Label-Noise Learning with Intrinsically Long-Tailed Data,” in Proceedings of IEEE/CVF International Conference of Computer Vision (ICCV), Paris, France, October 2-6, 2023. (CCF A)

[2] Yan Yan, Youze Xu, Jing-Hao Xue, Yang Lu, Hanzi Wang, and Wentao Zhu, “Drop Loss for Person Attribute Recognition with Imbalanced Noisy-Labeled Samples” IEEE Transactions on Cybernetics (TCYB), 2022, accepted. (JCR 1区 / CCF B)

[3] Youze Xu, Yan Yan, Jing-hao Xue, Yang Lu, and Hanzi Wang, “Small-Vote Sample Selection for Label-Noise Learning,” in Proceedings of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, pp. 729-744, Bilbao, Spain, September 13-17, 2021. (CCF B)