MetaCollaborativeAlgorithm<T, TInput, TOutput>
Models & Types · AiDotNet.MetaLearning.Algorithms
Implementation of Meta-Collaborative Learning for cross-domain few-shot learning.
Inheritance: ModelBase<T, TInput, TOutput> → MetaLearnerBase<T, TInput, TOutput> → MetaCollaborativeAlgorithm<T, TInput, TOutput>
Implements: IGeneratedParameterRegistrar<T>
How It Works
Meta-Collaborative Learning maintains a set of domain-specific gradient momentum buffers and uses gradient alignment (cosine similarity) between tasks to modulate cross-task knowledge transfer. Tasks with well-aligned gradients reinforce each other’s updates; conflicting gradients are dampened via a PCGrad-inspired projection.
Algorithm:
Constructors
| Constructor | Summary |
|---|---|
MetaCollaborativeAlgorithm(MetaCollaborativeOptions<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) |