HyperNetMetaRLAlgorithm<T, TInput, TOutput>
Models & Types · AiDotNet.MetaLearning.Algorithms
Implementation of HyperNet Meta-RL: hypernetwork-based task-specific parameter generation.
Inheritance: ModelBase<T, TInput, TOutput> → MetaLearnerBase<T, TInput, TOutput> → HyperNetMetaRLAlgorithm<T, TInput, TOutput>
Implements: IGeneratedParameterRegistrar<T>
How It Works
A hypernetwork takes a task embedding (computed from the support gradient) and generates the full policy parameter vector in a single forward pass. The task embedding is computed by compressing the initial support gradient through a learned encoder. The hypernetwork is a 2-layer MLP that maps the embedding to a parameter delta vector.
Algorithm:
Constructors
| Constructor | Summary |
|---|---|
HyperNetMetaRLAlgorithm(HyperNetMetaRLOptions<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) |