ACLOptions<T, TInput, TOutput>
Options & Configuration ยท AiDotNet.MetaLearning.Options
Configuration options for the ACL (Adaptive Continual Learning) algorithm.
Inheritance: ModelOptions โ ACLOptions<T, TInput, TOutput>
Implements: IMetaLearnerOptions<T>
How It Works
ACL learns task-specific parameter importance masks that protect critical weights from catastrophic forgetting. The importance is estimated from gradient magnitudes accumulated via exponential moving average across adaptation steps.
Constructors
| Constructor | Summary |
|---|---|
ACLOptions(IFullModel<T, TInput, TOutput> metaModel) |
Creates options for meta-learning metaModel, with the default settings. |
Properties
| Property | Summary |
|---|---|
IEpisodicDataLoader<T, TInput, TOutput> DataLoader { get; set; } |
Episodic data loader that supplies the meta-training tasks; null when tasks are passed to training directly. |
IFullModel<T, TInput, TOutput> MetaModel { get; set; } |
The model whose parameters the meta-learner trains to adapt quickly to new tasks. |
IGradientBasedOptimizer<T, TInput, TOutput> InnerOptimizer { get; set; } |
Must stay null: this algorithm takes no plain gradient inner step (it adapts without one, or by its own update rule), so a supplied optimizer is rejected when the learner is built. |
IGradientBasedOptimizer<T, TInput, TOutput> MetaOptimizer { get; set; } |
Optimizer for the meta-update. |
ILossFunction<T> LossFunction { get; set; } |
Loss function for meta-training; null uses the default multi-class classification loss. |
bool EnableCheckpointing { get; set; } |
|
bool UseFirstOrder { get; set; } |
Whether to use the first-order approximation of the meta-gradient. |
double ElasticRegWeight { get; set; } |
Regularization weight pulling adapted parameters toward the initial (pre-task) values for protected parameters. |
double ImportanceDecay { get; set; } |
EMA decay rate for accumulating parameter importance across tasks. |
double InnerLearningRate { get; set; } |
|
double MaskSparsityPenalty { get; set; } |
L1 penalty coefficient on the parameter importance masks to encourage sparsity. |
double OuterLearningRate { get; set; } |
|
double ProtectionStrength { get; set; } |
How much to scale down the learning rate for parameters deemed important. |
double? GradientClipThreshold { get; set; } |
|
int AdaptationSteps { get; set; } |
|
int CheckpointFrequency { get; set; } |
|
int EvaluationFrequency { get; set; } |
|
int EvaluationTasks { get; set; } |
|
int MetaBatchSize { get; set; } |
|
int NumMetaIterations { get; set; } |
|
int? RandomSeed { get; set; } |
Methods
| Method | Summary |
|---|---|
IMetaLearnerOptions<T> Clone() |
|
bool IsValid() |