(Optimization of mean hyperparameters) Assume the mean function of a Gaussian process is

parametrized by a parameter ??? , ??(x;?)?

2. (Optimization of mean hyperparameters) Assume the mean function of a Gaussian process is parametrized by a parameter 6', ,u,(x; 6). (a) (10 points) Given a xed covariance function K , noise variance 02, and data set D = (X, y), compute the derivative of the negative log marginal likelihood with respect to 9: d 2 logpcy i new )- Express your answer in terms of the derivative of the mean function w.r.t. 9, (iii/d9