Joint Optimization Framework for Learning with Noisy Labels
This addresses the issue of noisy labels in image classification for researchers and practitioners using web-collected datasets, representing a strong specific gain rather than a foundational advancement.
The paper tackles the problem of performance degradation in deep neural networks due to noisy labels in large-scale datasets by proposing a joint optimization framework that learns network parameters and estimates true labels, achieving significant outperformance over state-of-the-art methods on noisy CIFAR-10 and Clothing1M datasets.
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.