MetaPACOHAlgorithm<T, TInput, TOutput>
Models & Types · AiDotNet.MetaLearning.Algorithms
Implementation of Meta-PACOH: Hierarchical PAC-Bayesian Meta-Learning with per-group prior variances.
Inheritance: ModelBase<T, TInput, TOutput> → MetaLearnerBase<T, TInput, TOutput> → MetaPACOHAlgorithm<T, TInput, TOutput>
Implements: IGeneratedParameterRegistrar<T>
How It Works
Meta-PACOH extends PACOH by partitioning the parameter space into G groups and learning independent prior log-variances {log(σ²_g)} for each group. This allows different parts of the network (e.g., early vs late layers) to have different levels of flexibility during task adaptation.
Algorithm:
Constructors
| Constructor | Summary |
|---|---|
MetaPACOHAlgorithm(MetaPACOHOptions<T, TInput, TOutput> options) |
Properties
| Property | Summary |
|---|---|
MetaLearningAlgorithmType AlgorithmType { get; } |
Methods
| Method | Summary |
|---|---|
IModel<TInput, TOutput, ModelMetadata<T>> Adapt(IMetaLearningTask<T, TInput, TOutput> task) |
|
T MetaTrain(TaskBatch<T, TInput, TOutput> taskBatch) |