Score-Based Active Learning Proposer

Scores a finite candidate set once with an acquisition function and returns the top candidates.

Creates a score-based proposer.

Arguments

acq_id

(character(1))
Acquisition-function registry id.

surrogate_id

(character(1))
Surrogate registry id.

candidate_sampler

(NULL | SpaceSampler)
Optional sampler used to create the candidate set to score. NULL means exhaustive scoring of the remaining finite pool.

n_candidates

(integer(1) | Inf)
Number of candidates requested from candidate_sampler. Ignored when candidate_sampler is NULL.

acq_fit_scope

(character(1))
"global" fits acquisition functions on the run's full finite pool; "candidate" fits on the current candidate set; "search_space" fits on the search space bounds.