arXiv 1804.06872

Co-teaching: Robust Training of Deep Neural Networks with Extremely Noisy Labels

By Bo Han, Quanming Yao, et al.

Published 2018-04-18

Wiki summary

Explore the paper's summary, context, and related research on Papiers.

Deep learning with noisy labels is practically challenging, as the capacity of deep models is so high that they can totally memorize these noisy labels sooner or later during training. Nonetheless, recent studies on the memorization effects of deep neural networks show that they would first memorize training data of clean labels and then those of noisy labels. Therefore in this paper, we propose a new deep learning…

View the original paper on arXiv