Semi-supervised Learning via Gaussian Processes
Neil David Lawrence, Michael I. Jordan · Neural Information Processing Systems · 2004
We present a probabilistic approach to learning a Gaussian Process classifier in the presence of unlabeled data. Our approach involves a null category noise (NCNM) inspired by ordered categorical noise models. The noise model reflects an assumption that the data density is lower between the class-conditional densities. We illustrate our approach on a toy problem and present comparative results for the semi-supervised classification of handwritten digits.