Utils

OBJECTIVE module-attribute

OBJECTIVE = 'OBJECTIVE'

Tag for the objective dataset/function in standard utility functions.

FunctionEvaluator module-attribute

FunctionEvaluator = Callable[[Float[Array, 'N D']], Dict[str, Dataset]]

Type alias for function evaluators, which take an array of points of shape \([N, D]\) and evaluate a set of functions at each point, returning a mapping from function tags to datasets of the evaluated points. This is the same as the Observer in Trieste: https://github.com/secondmind-labs/trieste/blob/develop/trieste/observer.py

build_function_evaluator

build_function_evaluator(functions)

Takes a dictionary of functions and returns a FunctionEvaluator which can be used to evaluate each of the functions at a supplied set of points and return a dictionary of datasets storing the evaluated points.

get_best_latent_observation_val

get_best_latent_observation_val(posterior, dataset)

Takes a posterior and dataset and returns the best (latent) function value in the dataset, corresponding to the minimum of the posterior mean value evaluated at locations in the dataset. In the noiseless case, this corresponds to the minimum value in the dataset.