HyperCLIPAlgorithm<T, TInput, TOutput>
Models & Types · AiDotNet.MetaLearning.Algorithms
Implementation of HyperCLIP: Contrastive Learning for Hypernetwork-based Meta-Learning.
Inheritance: ModelBase<T, TInput, TOutput> → MetaLearnerBase<T, TInput, TOutput> → HyperCLIPAlgorithm<T, TInput, TOutput>
Implements: IGeneratedParameterRegistrar<T>
How It Works
HyperCLIP uses contrastive alignment between task embeddings (from support gradients) and parameter embeddings (from adapted parameter deltas). An InfoNCE contrastive loss aligns each task’s gradient signature with its adapted parameter fingerprint, learning a shared projection space. This cross-modal alignment improves adaptation quality.
Algorithm:
Constructors
| Constructor | Summary |
|---|---|
HyperCLIPAlgorithm(HyperCLIPOptions<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) |