We investigate the use of second order optimization approaches for solving the multiple kernel learning (MKL) problem. We show that the hessian of the MKL can be computed efficiently and this informat
