Enum DifficultyEstimationType

Namespace

AiDotNet.CurriculumLearning.Interfaces

Assembly

AiDotNet.dll

Types of difficulty estimation methods.

public enum DifficultyEstimationType

Fields

ComplexityBased = 5

Uses sample complexity metrics (e.g., input magnitude).

ConfidenceBased = 1

Uses model confidence/margin as difficulty measure. Lower confidence = harder sample.

Ensemble = 6

Combines multiple estimators.

EnsembleBased = 3

Uses prediction variance across model ensemble. Higher variance = harder sample.

ExpertDefined = 4

Uses domain expert-defined difficulty scores.

LossBased = 0

Uses training loss as difficulty measure. Higher loss = harder sample.

TransferBased = 2

Uses gap between simple and complex model performance. Larger gap = harder sample.