papers AI Learner
The Github is limit! Click to go to the new site.

Interpolation Consistency Training for Semi-Supervised Learning

2019-03-09
Vikas Verma, Alex Lamb, Juho Kannala, Yoshua Bengio, David Lopez-Paz

Abstract

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.

Abstract (translated by Google)
URL

http://arxiv.org/abs/1903.03825

PDF

http://arxiv.org/pdf/1903.03825


Similar Posts

Comments