We present an approach to semi-supervised learning based on an exponential family characterization. Our approach generalizes previous work on coupled priors for hybrid generative/discriminative models. Our model is more flexible and natural than previous approaches. Experimental results on several data sets show that our approach also performs better in practice.
Cite
@article{arxiv.1003.0696,
title = {Exponential Family Hybrid Semi-Supervised Learning},
author = {Arvind Agarwal and Hal Daume},
journal= {arXiv preprint arXiv:1003.0696},
year = {2010}
}