Timezone: »

Co-teaching: Robust training of deep neural networks with extremely noisy labels
Bo Han · Quanming Yao · Xingrui Yu · Gang Niu · Miao Xu · Weihua Hu · Ivor Tsang · Masashi Sugiyama

Tue Dec 04 07:45 AM -- 09:45 AM (PST) @ Room 517 AB #110

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 paradigm called ''Co-teaching'' for combating with noisy labels. Namely, we train two deep neural networks simultaneously, and let them teach each other given every mini-batch: firstly, each network feeds forward all data and selects some data of possibly clean labels; secondly, two networks communicate with each other what data in this mini-batch should be used for training; finally, each network back propagates the data selected by its peer network and updates itself. Empirical results on noisy versions of MNIST, CIFAR-10 and CIFAR-100 demonstrate that Co-teaching is much superior to the state-of-the-art methods in the robustness of trained deep models.

Author Information

Bo Han (RIKEN & UTS)
Quanming Yao (4Paradigm)
Xingrui Yu (University of Technology Sydney)

Ph.D Candidate at University of Technology Sydney.

Gang Niu (RIKEN)
Weihua Hu (The University of Tokyo)
Ivor Tsang (University of Technology, Sydney)
Masashi Sugiyama (RIKEN / University of Tokyo)

More from the Same Authors