Simplify specification of kernels for Gaussian processes
|Reported by:||gkronber||Owned by:|
|Priority:||lowest||Milestone:||HeuristicLab 3.3.x Backlog|
It is currently quite cumbersome to specify the kernel function for Gaussian process regression. The current implementation is closely related to the GPML implementation.
Instead we could also use our existing framework for symbolic expressions and AutoDiff to parse the expression for the kernel.