ContinualLearningResult<T>

Models & Types · AiDotNet.ContinualLearning.Results

Result from training on a single task in continual learning.

For Beginners

This class captures everything that happened during training on one task - how well the model learned, how long it took, and whether it forgot previous knowledge.

How It Works

Key Metrics Explained:

  • Training Loss: How far predictions were from targets (lower is better)
  • Training Accuracy: Percentage of correct predictions on training data
  • Average Previous Task Accuracy: How well the model still performs on old tasks
  • Forgetting: How much accuracy was lost on previous tasks (lower is better)

Reference: Parisi et al. “Continual Lifelong Learning with Neural Networks: A Review” (2019)

Constructors

Constructor Summary
ContinualLearningResult(int taskId, T trainingLoss, T trainingAccuracy, T averagePreviousTaskAccuracy, TimeSpan trainingTime, Vector<T> lossHistory, Vector<T> regularizationLossHistory) Initializes a new instance of the ContinualLearningResult class.

Properties

Property Summary
IReadOnlyDictionary<string, object> StrategyMetrics { get; set; } Gets additional strategy-specific metrics.
T AveragePreviousTaskAccuracy { get; } Gets the average accuracy on all previously learned tasks after training on this task.
T EffectiveLearningRate { get; set; } Gets the effective learning rate used (may vary with schedulers).
T Forgetting { get; set; } Gets the amount of forgetting on previous tasks (accuracy drop).
T ForwardTransfer { get; set; } Gets the forward transfer metric.
T TrainingAccuracy { get; } Gets the final training accuracy on this task.
T TrainingLoss { get; } Gets the final training loss on this task.
T ValidationAccuracy { get; set; } Gets the validation accuracy on this task, if validation was performed.
T ValidationLoss { get; set; } Gets the validation loss on this task, if validation was performed.
TimeSpan TrainingTime { get; } Gets the time taken to train on this task.
Vector<T> LossHistory { get; } Gets the loss history across training epochs.
Vector<T> RegularizationLossHistory { get; } Gets the regularization loss history (e.g., EWC penalty term).
int TaskId { get; } Gets the task identifier (0-indexed).
int? GradientUpdates { get; set; } Gets the number of gradient updates performed.
int? SampleCount { get; set; } Gets the number of samples used for training.
long? PeakMemoryBytes { get; set; } Gets the peak memory usage during training in bytes.

Methods

Method Summary
string ToString() Returns a string representation of the training result.