BayTransProtoAlgorithm<T, TInput, TOutput>

Models & Types · AiDotNet.MetaLearning.Algorithms

Implementation of BayTransProto: Bayesian Transductive Prototypical Networks.

Inheritance: ModelBase<T, TInput, TOutput> → MetaLearnerBase<T, TInput, TOutput> → BayTransProtoAlgorithm<T, TInput, TOutput>

Implements: IGeneratedParameterRegistrar<T>

How It Works

BayTransProto combines Bayesian posterior sampling with transductive refinement. After standard inner-loop adaptation, the algorithm samples multiple parameter vectors from a Gaussian posterior centered on the adapted params. Transductive refinement then uses query-set gradients to iteratively update the posterior mean, leveraging unlabeled query data to improve adaptation.

Algorithm:

Constructors

Constructor Summary
BayTransProtoAlgorithm(BayTransProtoOptions<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)