PACOHAlgorithm<T, TInput, TOutput>
Models & Types · AiDotNet.MetaLearning.Algorithms
Implementation of PACOH: PAC-Bayesian Meta-Learning with Optimal Hyperparameters (Rothfuss et al., ICLR 2021).
Inheritance: ModelBase<T, TInput, TOutput> → MetaLearnerBase<T, TInput, TOutput> → PACOHAlgorithm<T, TInput, TOutput>
Implements: IGeneratedParameterRegistrar<T>
How It Works
PACOH meta-learns a Gaussian prior distribution N(μ_prior, diag(σ²_prior)) over model parameters such that the PAC-Bayesian generalization bound is minimized. The key insight is that optimizing the prior (not just the posterior) yields tighter generalization guarantees for meta-learning.
Algorithm:
Constructors
| Constructor | Summary |
|---|---|
PACOHAlgorithm(PACOHOptions<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) |