Struct MetaLearner 

pub struct MetaLearner { /* private fields */ }

Meta-learner that estimates good initial GP hyperparameters from prior tasks.

Uses a simple similarity-weighted average of per-task features.

Implementations

impl MetaLearner

pub fn from_runs(runs: &[PriorRun]) -> Self

Create a meta-learner from a set of prior runs.

pub fn suggest_hyperparams( &self, target_bounds: &[(f64, f64)], ) -> (f64, f64, f64)

Suggest initial GP hyperparameters for a target task.

Returns (length_scale, signal_variance, noise_variance).

pub fn task_similarity(prior: &PriorRun, target_bounds: &[(f64, f64)]) -> f64

Compute similarity between a prior task and a new target task.

Trait Implementations

impl Clone for MetaLearner

fn clone(&self) -> MetaLearner

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for MetaLearner

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.