Making Deep Neural Networks Robust to Label Noise: A Loss Correction Approach

• Published in: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) pp. 2233 - 2241
• Main Authors: Patrini, Giorgio, Rozza, Alessandro, Menon, Aditya Krishna, Nock, Richard, Lizhen Qu
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.07.2017
• Subjects: Computer architecture; Computer vision; Neural networks; Noise measurement; Robustness; Training
• ISSN: 1063-6919; 1063-6919
• DOI: 10.1109/CVPR.2017.240

We present a theoretically grounded approach to train deep neural networks, including recurrent networks, subject to class-dependent label noise. We propose two procedures for loss correction that are agnostic to both application domain and network architecture. They simply amount to at most a matrix inversion and multiplication, provided that we know the probability of each class being corrupted into another. We further show how one can estimate these probabilities, adapting a recent technique for noise estimation to the multi-class setting, and thus providing an end-to-end framework. Extensive experiments on MNIST, IMDB, CIFAR-10, CIFAR-100 and a large scale dataset of clothing images employing a diversity of architectures - stacking dense, convolutional, pooling, dropout, batch normalization, word embedding, LSTM and residual layers - demonstrate the noise robustness of our proposals. Incidentally, we also prove that, when ReLU is the only non-linearity, the loss curvature is immune to class-dependent label noise.