Paper Session - CVPR Main Conference Invited Speakers : Paper #1 (10:36 - 10:40)

Paper Title: Distilling Effective Supervision From Severe Label Noise

Authors: Zizhao Zhang, Han Zhang, Sercan Ö. Arık, Honglak Lee, Tomas Pfister

Email: zizhaoz@google.com, zhanghan@google.com, soarik@google.com, honglak@google.com, tpfister@google.com

Short Description: We estimate Data Coefficients with a generalized meta learning framework and set new state of the arts on noisy label benchmarks. Code is available.

Keywords: Robust training, Meta learning, Noise labels.

Abstract : Collecting large-scale data with clean labels for supervised training of neural networks is practically challenging. Although noisy labels are usually cheap to acquire, existing methods suffer a lot from label noise. This paper targets at the challenge of robust training at high label noise regimes. The key insight to achieve this goal is to wisely leverage a small trusted set to estimate exemplar weights and pseudo labels for noisy data in order to reuse them for supervised training. We present a holistic framework to train deep neural networks in a way that is highly invulnerable to label noise. Our method sets the new state of the art on various types of label noise and achieves excellent performance on large-scale datasets with real-world label noise. For instance, on CIFAR100 with a 40% uniform noise ratio and only 10 trusted labeled data per class, our method achieves 80.2±0.3% classification accuracy, where the error rate is only 1.4% higher than a neural network trained without label noise. Moreover, increasing the noise ratio to 80%, our method still maintains a high accuracy of 75.5±0.2%, compared to the previous best accuracy 48.2% 1 .