BayProNetAlgorithm<T, TInput, TOutput>

Models & Types · AiDotNet.MetaLearning.Algorithms

Implementation of BayProNet: Bayesian Prototypical Networks with uncertainty-aware prototype distributions for few-shot learning.

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

Implements: IGeneratedParameterRegistrar<T>

How It Works

BayProNet extends Prototypical Networks by modeling class prototypes as Gaussian distributions rather than point estimates. In the parameter-space formulation, adaptation produces a posterior distribution N(θ_adapted, σ²) over parameters, and predictions are made by ensembling over parameter samples. The variance σ² is meta-learned per parameter dimension, enabling uncertainty-aware few-shot learning.

Algorithm:

Constructors

Constructor Summary
BayProNetAlgorithm(BayProNetOptions<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)