GPFlowSampling in Robotics

[AdrianPrados]

Good afternoon, I am applying GPFlow to a robotic process based on Gaussian Process for manipulation. I have a couple of questions about the algorithm.

Firstly, the first function generated in the examples above is randomly generated with random_fourier. I was wondering if it would be possible to initialise the prior function with information from a previously generated GP solution.

Secondly, I've been testing the codes and I haven't seen how the covariance of the final solution is treated. Are we working with it for simplicity in the examples, or are we just not looking to update the covariance for the new solution?

Thank you very much for your work and I hope it will be useful for me.