ActiveLearningContext<T>

Models & Types · AiDotNet.ActiveLearning.Interfaces

Context information for stopping criterion evaluation.

For Beginners

This class provides all the information a stopping criterion needs to make its decision, including history of metrics and current state.

Constructors

Constructor Summary
ActiveLearningContext()

Properties

Property Summary
List<T> AccuracyHistory { get; set; } Gets or sets the history of training accuracy.
List<T> LossHistory { get; set; } Gets or sets the history of training loss.
List<T> QueryScoreHistory { get; set; } Gets or sets the history of query informativeness scores.
List<T> UncertaintyHistory { get; set; } Gets or sets the history of average uncertainty scores.
List<T> ValidationAccuracyHistory { get; set; } Gets or sets the history of validation accuracy.
TimeSpan ElapsedTime { get; set; } Gets or sets the time elapsed since learning started.
TimeSpan? MaxTime { get; set; } Gets or sets the maximum allowed time for learning.
Vector<T> CurrentPredictions { get; set; } Gets or sets the current predictions for stability checking.
Vector<T> PreviousPredictions { get; set; } Gets or sets the predictions from the previous iteration for stability checking.
int CurrentIteration { get; set; } Gets or sets the current iteration number.
int MaxBudget { get; set; } Gets or sets the maximum labeling budget.
int TotalLabeled { get; set; } Gets or sets the total number of labeled samples.
int UnlabeledRemaining { get; set; } Gets or sets the number of unlabeled samples remaining.