Deep unsupervised learning through spatial contrasting

Elad Hoffer, Itay Hubara, Nir Ailon

Abstract:

Convolutional networks have marked their place over the last few years as the
best performing model for various visual tasks. They are, however, most suited
for supervised learning from large amounts of labeled data. Previous attempts
have been made to use unlabeled data to improve model performance by applying
unsupervised techniques. These attempts require different architectures and training methods.
In this work we present a novel approach for unsupervised training
of Convolutional networks that is based on contrasting between spatial regions
within images. This criterion can be employed within conventional neural net-
works and trained using standard techniques such as SGD and back-propagation,
thus complementing supervised methods.

Conflicts: technion.ac.il, intel.com

Keywords: Unsupervised Learning, Deep learning, Computer vision