Package index • celecx

Package

celecx celecx-package: celecx: Computer Experiment LEarning Curve eXtrapolation

High-level entry points that construct and run active-learning optimizers.

optimize_active(): Run Active Learning
optimizer_al(): Active Learning Optimizer Factory
optimizer_pool_al(): Convenience Constructor for Pool-Based Active Learning Optimizers

Instance, callback context, and termination condition of a search run.

Optimizer base classes and concrete optimizers over search spaces or candidate pools.

OptimizerSearchAbstract: Abstract Base Class for Search-Compatible Optimizers
OptimizerPoolAbstract: Abstract Base Class for Pool-Aware Optimizers
OptimizerAL: Proposer-Based Active Learning Optimizer
mlr_optimizers_pool_sample OptimizerPoolSample: Sampler-Based Search on Pool-Restricted or Discrete Objectives

Proposers that build active-learning batches inside OptimizerAL, and the context they operate on.

ALContext: Active Learning Proposal Context
ALProposer: Active Learning Proposer
ALProposerScoreAbstract: Abstract Base Class for Score-Based Active Learning Proposers
ALProposerScore: Score-Based Active Learning Proposer
ALProposerSequentialScore: Sequential Score-Based Active Learning Proposer
ALProposerSequentialReference: Sequential Reference Active Learning Proposer
ALProposerPseudoLabel: Pseudo-Label Active Learning Proposer
ALProposerPortfolio: Portfolio Active Learning Proposer

Modifiers that adjust acquisition scores during sequential batch construction.

ALScoreModifier: Active Learning Score Modifier
ALScoreModifierDiversity: Diversity Active Learning Score Modifier
ALScoreModifierLocalPenalization: Local-Penalization Active Learning Score Modifier
ALScoreModifierNone: No-Op Active Learning Score Modifier

Active-learning acquisition functions, registered in mlr3mbo’s dictionary.

Distance objects, their dictionary, and sugar constructors.

ALDistance: Active Learning Distance Base Class
ALDistanceGeometry: Geometry-Based Active Learning Distance
mlr_al_distances: Dictionary of Active Learning Distances
mlr_al_distances_affine ALDistanceAffine: Affine Active Learning Distance
mlr_al_distances_gower ALDistanceGower: Gower Active Learning Distance
mlr_al_distances_standardize ALDistanceStandardize: Standardized Active Learning Distance
clx_ald(): Syntactic Sugar Active Learning Distance Construction
clx_alds(): Syntactic Sugar Active Learning Distances Construction

Space-filling and pool-based samplers, their dictionary, and sugar constructors.

SpaceSampler: Space Sampler Base Class
SpaceSamplerDistance: Distance-Based Space Sampler
mlr_space_samplers: Dictionary of Space Samplers
mlr_space_samplers_chain SpaceSamplerChain: Chained Space Sampler
mlr_space_samplers_conditional SpaceSamplerConditional: Conditional Space Sampler
mlr_space_samplers_gsx SpaceSamplerGSx: GSx Space Sampler
mlr_space_samplers_kmeans SpaceSamplerKMeans: K-Means Space Sampler
mlr_space_samplers_kmedoids SpaceSamplerKMedoids: K-Medoids Space Sampler
mlr_space_samplers_lhs SpaceSamplerLhs: LHS Space Sampler
mlr_space_samplers_relational_kmeans SpaceSamplerRelationalKMeans: Relational K-Means Space Sampler
mlr_space_samplers_sobol SpaceSamplerSobol: Sobol Space Sampler
mlr_space_samplers_uniform SpaceSamplerUniform: Uniform Space Sampler
clx_sps(): Syntactic Sugar Space Sampler Construction
clx_spss(): Syntactic Sugar Space Samplers Construction

Objectives evaluated by table lookup or fitted learners, and pool-restricted objective wrappers.

ObjectiveDataset: Objective Based on Pre-evaluated Dataset
ObjectiveLearner: Objective Function Based on a Fitted Learner
ObjectivePoolAbstract: Abstract Base Class for Pool-backed Objectives
ObjectivePoolRFun: Objective Function Based on a Candidate Pool and R Function
ObjectivePoolWrapper: Objective Function Wrapping Another Objective on a Candidate Pool

Regression learners and wrappers with uncertainty quantification.

mlr_learners_regr.bootstrap_se LearnerRegrBootstrapSE: Bootstrap Ensemble Learner with SE Prediction
mlr_learners_regr.deepgp LearnerRegrDeepGP: Deep GP Regression Learner
mlr_learners_regr.gpfit LearnerRegrGPfit: GPfit Regression Learner
mlr_learners_regr.hetgp LearnerRegrHetGP: hetGP Regression Learner
mlr_learners_regr.quantile_se LearnerRegrQuantileSE: Quantile Regression Learner with SE Prediction
mlr_learners_regr.tgp LearnerRegrTGP: tgp Regression Learner

Turning a run into a learning-curve task, online via callback or offline by replay.

celecx.surrogate_performance CallbackSurrogatePerformance: Surrogate Performance Callback
replay_surrogate_performance(): Offline Surrogate-Performance Replay
replay_surrogate_perf_table(): Offline Surrogate-Performance Table
task_lce_from_perf(): Build an LCE Task from an Archive and a Performance Table
task_lce_best_so_far(): Best-So-Far Optimization Trace as an LCE Task
TaskLCE: Learning Curve Extrapolation Task
lce_link() lce_link_from_range(): Predictive Link Functions for LCE Forecasts

Learners that extrapolate learning curves, and their prediction object.

LearnerLCE: Learning Curve Extrapolation Learner
mlr_learners_lce.bootstrap LearnerLCEBootstrap: Residual Bootstrap LCE Learner Wrapper
mlr_learners_lce.conformal LearnerLCEConformal: Split-Conformal LCE Learner Wrapper
mlr_learners_lce.featureless LearnerLCEFeatureless: Featureless LCE Learner
mlr_learners_lce.gam LearnerLCEGam: GAM LCE Learner
mlr_learners_lce.isotonic LearnerLCEIsotonic: Isotonic LCE Learner
mlr_learners_lce.parametric_exponential LearnerLCEParametricExponential: Parametric Exponential LCE Learner
mlr_learners_lce.parametric_log LearnerLCEParametricLog: Parametric Log LCE Learner
mlr_learners_lce.parametric_logistic LearnerLCEParametricLogistic: Parametric Logistic LCE Learner
mlr_learners_lce.parametric_power_law LearnerLCEParametricPowerLaw: Parametric Power-Law LCE Learner
mlr_learners_lce.rolling_slope LearnerLCERollingSlope: Rolling Slope LCE Learner
mlr_learners_lce.simulate LearnerLCESimulate: Forward-Simulation LCE Learner
mlr_learners_lce.spline_monotone LearnerLCESplineMonotone: Monotone Spline LCE Learner
PredictionLCE: LCE Prediction Object

Measures and resampling schemes for learning-curve tasks.

MeasureLCE: Learning Curve Extrapolation Measure
mlr_measures_lce: Per-Batch LCE Loss Measures
mlr_measures_lce_distributional: Distributional LCE Measures
mlr_resamplings_lce.expanding_cv ResamplingLCE: Expanding-Window LCE Cross-Validation

Shared abstractions.