ICMFusionOptions<T, TInput, TOutput>

Options & Configuration · AiDotNet.MetaLearning.Options

Configuration options for ICM-Fusion (In-Context Meta-Optimized LoRA Fusion, 2025).

Inheritance: ModelOptions → ICMFusionOptions<T, TInput, TOutput>

Implements: IMetaLearnerOptions<T>

How It Works

ICM-Fusion fuses multiple task-specific parameter deltas by encoding them into a latent space via a Fusion-VAE, then reconstructing the fused adapter. The VAE is meta-learned so that task vector arithmetic in latent space resolves inter-weight conflicts.

Constructors

Constructor Summary
ICMFusionOptions(IFullModel<T, TInput, TOutput> metaModel) Creates options for meta-learning metaModel, with the default settings.

Properties

Property Summary
IEpisodicDataLoader<T, TInput, TOutput> DataLoader { get; set; } Episodic data loader that supplies the meta-training tasks; null when tasks are passed to training directly.
IFullModel<T, TInput, TOutput> MetaModel { get; set; } The model whose parameters the meta-learner trains to adapt quickly to new tasks.
IGradientBasedOptimizer<T, TInput, TOutput> InnerOptimizer { get; set; } Optimizer for task adaptation.
IGradientBasedOptimizer<T, TInput, TOutput> MetaOptimizer { get; set; } Must stay null: this algorithm’s meta-update is part of the method rather than a plain gradient step, so a supplied optimizer is rejected when the learner is built.
ILossFunction<T> LossFunction { get; set; } Loss function for meta-training; null uses the default multi-class classification loss.
bool EnableCheckpointing { get; set; }
bool UseFirstOrder { get; set; } Whether to use the first-order approximation of the meta-gradient.
double FusionDecay { get; set; } Exponential decay for older fusion components.
double InnerLearningRate { get; set; }
double KLWeight { get; set; } KL divergence weight in the VAE loss: L = L_recon + KLWeight * L_KL.
double OuterLearningRate { get; set; }
double? GradientClipThreshold { get; set; }
int AdaptationSteps { get; set; }
int CheckpointFrequency { get; set; }
int EvaluationFrequency { get; set; }
int EvaluationTasks { get; set; }
int LatentDim { get; set; } Dimensionality of the VAE latent space.
int MetaBatchSize { get; set; }
int NumFusionComponents { get; set; } Number of task-specific components to maintain for fusion.
int NumMetaIterations { get; set; }
int? RandomSeed { get; set; }

Methods

Method Summary
IMetaLearnerOptions<T> Clone()
bool IsValid()