Online co-regularized algorithms

Abstract

We propose an online co-regularized learning algorithm for classification and regression tasks. We demonstrate that by sequentially co-regularizing prediction functions on unlabeled data points, our algorithm provides improved performance in comparison to supervised methods on several UCI benchmarks and a real world natural language processing dataset. The presented algorithm is particularly applicable to learning tasks where large amounts of (unlabeled) data are available for training. We also provide an easy to set-up and use Python implementation of our algorithm. © 2012 Springer-Verlag Berlin Heidelberg.