ContextMetaRLAlgorithm<T, TInput, TOutput>
Models & Types · AiDotNet.MetaLearning.Algorithms
Implementation of Context Meta-RL: context-conditioned meta-reinforcement learning with multi-head attention-based aggregation.
Inheritance: ModelBase<T, TInput, TOutput> → MetaLearnerBase<T, TInput, TOutput> → ContextMetaRLAlgorithm<T, TInput, TOutput>
Implements: IGeneratedParameterRegistrar<T>
How It Works
Context Meta-RL uses multi-head attention to aggregate task context from the support set gradient history. A learned query vector attends over encoded gradient entries to produce a context vector. This context vector multiplicatively modulates the model parameters, providing a smooth, deterministic task conditioning mechanism.
Algorithm:
Constructors
| Constructor | Summary |
|---|---|
ContextMetaRLAlgorithm(ContextMetaRLOptions<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) |