论文标题
KNN增强的深度学习与嘈杂的标签
KNN-enhanced Deep Learning Against Noisy Labels
论文作者
论文摘要
深度神经网络（DNN）的监督学习是饥饿的数据。在存在嘈杂标签的情况下，优化DNN的性能变得至关重要，因为收集大型数据集通常会引入嘈杂的标签。受到K-Nearest邻居（KNN）对数据噪声的鲁棒性的启发，我们建议将深knn应用于标签清理。我们的方法利用DNN进行特征提取和KNN进行地面真实标签推断。我们对神经网络进行迭代训练，并更新标签以同时进行更高的标签恢复率和更好的分类性能。实验结果表明，在相同的设置下，我们的方法的表现优于现有标签校正方法，并且在多个数据集上获得了更好的准确性，例如，服装1M数据集的76.78％。
Supervised learning on Deep Neural Networks (DNNs) is data hungry. Optimizing performance of DNN in the presence of noisy labels has become of paramount importance since collecting a large dataset will usually bring in noisy labels. Inspired by the robustness of K-Nearest Neighbors (KNN) against data noise, in this work, we propose to apply deep KNN for label cleanup. Our approach leverages DNNs for feature extraction and KNN for ground-truth label inference. We iteratively train the neural network and update labels to simultaneously proceed towards higher label recovery rate and better classification performance. Experiment results show that under the same setting, our approach outperforms existing label correction methods and achieves better accuracy on multiple datasets, e.g.,76.78% on Clothing1M dataset.
