期刊
ARTIFICIAL INTELLIGENCEAND SOFT COMPUTING, PT I
卷 11508, 期 -, 页码 215-222出版社
SPRINGER INTERNATIONAL PUBLISHING AG
DOI: 10.1007/978-3-030-20912-4_21
关键词
Neural networks; Deep learning; Robust learning; Label noise; Categorical cross-entropy
Deep neural networks obtain nowadays outstanding results on many vision, speech recognition and natural language processing-related tasks. Such deep structures need to be trained on very large datasets, what makes annotating the data for supervised learning, particularly difficult and time-consuming task. In the supervised datasets label noise may occur, which makes the whole training process less reliable. In this paper we present a novel robust loss function based on categorical cross-entropy. We demonstrate its robustness for several amounts of noisy labels, on popular MNIST and CIFAR-10 datasets.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据