arXiv (Cornell University) · 2018 · 105 citations · 4 references
ComputingComputer ScienceComputer VisionMachine LearningMachine VisionDeep LearningNeural NetworksOptimization For Machine LearningOverfittingPattern RecognitionRepresentation LearningDeep Learning-based ClassificationJoint OptimizationLabel Noise MitigationLabel Noise ModelingLabel Noise RobustnessLabelsPerformance Degradation
Deep neural networks (DNNs) trained on large-scale datasets have exhibited significant performance in image classification. Many large-scale datasets are collected from websites, however they tend to contain inaccurate labels that are termed as noisy labels. Training on such noisy labeled datasets causes performance degradation because DNNs easily overfit to noisy labels. To overcome this problem, we propose a joint optimization framework of learning DNN parameters and estimating true labels. Our framework can correct labels during training by alternating update of network parameters and labels. We conduct experiments on the noisy CIFAR-10 datasets and the Clothing1M dataset. The results indicate that our approach significantly outperforms other state-of-the-art methods.
4
Semi-Supervised Learning Literature Survey
Xiaojin Zhu · Minds at UW (University of Wisconsin) · 2005 · 3.9K citations · Full text
Making Deep Neural Networks Robust to Label Noise: A Loss Correction Approach
Giorgio Patrini, Alessandro Rozza, Aditya Krishna Menon et al. · 2017 · 1.4K citations