In the form of the support vector machine and Gaussian processes, kernel-ba
sed systems are currently very popular approaches to supervised learning. U
nfortunately, the computational load for training kernel-based systems incr
eases drastically with the size of the training data set, such that these s
ystems are not ideal candidates for applications with large data sets. Neve
rtheless, research in this direction is very active. In this paper, I revie
w some of the current approaches toward scaling kernel-based systems to lar
ge data sets.