论文标题
矩阵平滑：在嘈杂标签下具有过渡矩阵的DNN的正则化
Matrix Smoothing: A Regularization for DNN with Transition Matrix under Noisy Labels
论文作者
论文摘要
在嘈杂标签的存在下培训深层神经网络（DNNS）是一项重要且具有挑战性的任务。由分类器和过渡矩阵组成的概率建模描绘了从真实标签到嘈杂标签的转换，并且是一种有希望的方法。然而，最近的概率方法直接将过渡矩阵应用于DNN，忽略了DNN对过度拟合的敏感性，并实现了不令人满意的性能，尤其是在均匀的噪声下。在本文中，受标签平滑的启发，我们提出了一种新颖的方法，其中使用平滑的过渡矩阵来更新DNN，以限制DNN在概率建模中的过度拟合。我们的方法称为矩阵平滑。我们还从经验上证明，我们的方法不仅可以显着提高概率建模的鲁棒性，而且还可以更好地估计过渡矩阵。
Training deep neural networks (DNNs) in the presence of noisy labels is an important and challenging task. Probabilistic modeling, which consists of a classifier and a transition matrix, depicts the transformation from true labels to noisy labels and is a promising approach. However, recent probabilistic methods directly apply transition matrix to DNN, neglect DNN's susceptibility to overfitting, and achieve unsatisfactory performance, especially under the uniform noise. In this paper, inspired by label smoothing, we proposed a novel method, in which a smoothed transition matrix is used for updating DNN, to restrict the overfitting of DNN in probabilistic modeling. Our method is termed Matrix Smoothing. We also empirically demonstrate that our method not only improves the robustness of probabilistic modeling significantly, but also even obtains a better estimation of the transition matrix.
