Practical experience has shown that in order to obtain the best possible pe
rformance, prior knowledge about invariances of a classification problem at
hand ought to be incorporated into the training procedure. We describe and
review all known methods for doing so in support vector machines, provide
experimental results, and discuss their respective merits. One of the signi
ficant new results reported in this work is our recent achievement of the l
owest reported test error on the well-known MNIST digit recognition benchma
rk task, with SVM training times that are also significantly faster than pr
evious SVM methods.