最佳转移矩阵估计还是最佳标签噪声鲁棒分类器?两种可能提升 T-Revision 性能的方法
Best Transition Matrix Esitimation or Best Label Noise Robustness Classifier? Two Possible Methods to Enhance the Performance of T-revision
AI总结:
本文研究标签噪声下的转移矩阵估计与鲁棒分类器构建,提出 T-Revision-Alpha 和 T-Revision-Softmax 改进方法,并在 FashionMINIST 和 CIFAR-10 上验证其性能。
摘要:
标签噪声是指数据集中由于人为错误或收集缺陷导致的错误标签,这在实际应用中很常见,并且会显著降低模型的准确率。本报告探讨了如何估计噪声转移矩阵,以及如何构建对标签噪声具有鲁棒性的深度学习分类器。在转移矩阵已知的情况下,我们应用前向校正和重要性重加权方法,利用转移矩阵来校正标签噪声的影响。当转移矩阵未知或不准确时,我们使用锚点假设和 T-Revision 系列方法来估计或校正噪声矩阵。在本研究中,我们通过开发 T-Revision-Alpha 和 T-Revision-Softmax 进一步改进了 T-Revision 方法,以增强稳定性和鲁棒性。此外,我们基于交叉熵损失函数设计并实现了两个基线分类器:多层感知机(MLP)和 ResNet-18。我们使用具有已知噪声转移矩阵的 FashionMINIST 数据集,比较了这些方法在预测干净标签和估计转移矩阵方面的性能。对于噪声转移矩阵未知的 CIFAR-10 数据集,我们估计了噪声矩阵,并评估了这些方法预测干净标签的能力。
英文摘要:
Label noise refers to incorrect labels in a dataset caused by human errors or collection defects, which is common in real-world applications and can significantly reduce the accuracy of models. This report explores how to estimate noise transition matrices and construct deep learning classifiers that are robust against label noise. In cases where the transition matrix is known, we apply forward correction and importance reweighting methods to correct the impact of label noise using the transition matrix. When the transition matrix is unknown or inaccurate, we use the anchor point assumption and T-Revision series methods to estimate or correct the noise matrix. In this study, we further improved the T-Revision method by developing T-Revision-Alpha and T-Revision-Softmax to enhance stability and robustness. Additionally, we designed and implemented two baseline classifiers, a Multi-Layer Perceptron (MLP) and ResNet-18, based on the cross-entropy loss function. We compared the performance of these methods on predicting clean labels and estimating transition matrices using the FashionMINIST dataset with known noise transition matrices. For the CIFAR-10 dataset, where the noise transition matrix is unknown, we estimated the noise matrix and evaluated the ability of the methods to predict clean labels.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。