Learning Vector Fields with Spectral Filtering

We present a class of regularized kernel methods for vector valued learning, which are based on filtering the spectrum of the kernel matrix. The considered methods include Tikhonov regularization as a