We introduce Interpolation Consistency Training (ICT), a simple and computation efficient algorithm for training Deep Neural Networks in the semi-supervised learning paradigm. ICT encourages the prediction at an interpolation of unlabeled points to be consistent with the interpolation of the predictions at those points. In classification problems, ICT moves the decision boundary to low-density regions of the data distribution. Our experiments show that ICT achieves state-of-the-art performance when applied to standard neural network architectures on the CIFAR-10 and SVHN benchmark datasets. Our theoretical analysis shows that ICT corresponds to a certain type of data-adaptive regularization with unlabeled points which reduces overfitting to labeled points under high confidence values.
@article{arxiv.1903.03825,
title = {Interpolation Consistency Training for Semi-Supervised Learning},
author = {Vikas Verma and Kenji Kawaguchi and Alex Lamb and Juho Kannala and Arno Solin and Yoshua Bengio and David Lopez-Paz},
journal= {arXiv preprint arXiv:1903.03825},
year = {2022}
}
Comments
This is the latest version, which is published in the Journal, "Neural Networks", in 2022. All the previous results are unchanged. Keyword: Deep Learning, Semi-supervised Learning, Mixup