Outlier Robust Learning Algorithm for Gaussian Process Classification
김현철, Zoubin Ghahramani · 한국정보과학회 학술발표논문집 · 2007
Gaussian process classifiers (GPCs) are fully statistical kernel classification models which have a latent function with Gaussian process prior. Recently, EP approximation method has been proposed to infer the posterior over the latent function. It can have a special hyperparameter which can treat outliers potentially. In this paper, we propose the outlier robust algorithm which alternates EP and the hyperparameter updating until convergence. We also show its usefulness with the simulation results.