O2U-Net: A Simple Noisy Label Detection Approach for Deep Neural Networks

By Jinchi Huang, Lie Qu, Rongfei Jia, Binqiang Zhao. 2019

283 citations recorded by the source corpus.

This paper proposes a novel noisy label detection approach, named O2U-net, for deep neural networks without human annotations. Different from prior work which requires specifically designed noise-robust loss functions or networks, O2U-net is easy to implement but effective. It only requires adjusting the hyper-parameters of the deep network to make its status transfer from overfitting to underfitting (O2U) cyclically. The losses of each sample are recorded during iterations. The higher the normalized average loss of a sample, the higher the probability of being noisy labels. O2U-net is naturally compatible with active learning and other human annotation approaches. This introduces extra flexibility for learning with noisy labels. We conduct sufficient experiments on multiple datasets in various settings. The experimental results prove the state-of-the-art of O2S-net.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Optimization & Training

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Optimization & Training

Citations in the corpus

Continual Learning on Noisy Data Streams via Self-Purified Replay (2021, background): “Recently, filtering methods based on training dynamics [32, 64, 58] have gained much popularity, based on the observation that models tend to learn clean data first and memorize the noisy labeled data later.”
Robust Learning by Self-Transition for Handling Noisy Labels (2020, background): “” One of the most common approaches is sample selection, which involves training a DNN for a possibly clean subset of noisy training data [5, 9, 11, 12, 31, 34].”
Image Classification with Deep Learning in the Presence of Noisy Labels: A Survey (2019, methodology): “ataset from a similar domain is ﬁne-tuned on the noisy dataset for relabeling. Afterward, the network is again trained on relabeled data to re-sample the dataset to construct a ﬁnal clean dataset. In [68], learning rate is adjusted cyclicly to change network status between underﬁtting and overﬁtting”
Learning From Noisy Labels With Deep Neural Networks: A Survey (2020, methodology): “We introduce publicly available image datasets and then describe widely used evaluation metrics.”