GatedFeatureLearningUnitLayer<T>
Layers · AiDotNet.NeuralNetworks.Layers
Gated Feature Learning Unit (GFLU) for GANDALF architecture.
Inheritance: LayerBase<T> → GatedFeatureLearningUnitLayer<T>
Implements: IShapeContract
For Beginners
GFLU works like a smart filter:
- Look at all features and decide which ones matter (gating)
- Transform the selected features
- Combine them for the next layer
The “gate” is like a dimmer switch that can turn features on/off or anywhere in between.
How It Works
The GFLU is the core building block of GANDALF that performs feature selection and transformation through a gating mechanism. It learns which features are important and how to transform them.
Constructors
| Constructor | Summary |
|---|---|
GatedFeatureLearningUnitLayer(int outputDim) |
Initializes a Gated Feature Learning Unit. |
Properties
| Property | Summary |
|---|---|
bool SupportsTraining { get; } |
Methods
| Method | Summary |
|---|---|
IReadOnlyList<ILayer<T>> GetSubLayers() |
|
IReadOnlyList<OutputAxisContract> OutputAxesFor(int inputRank) |
|
Tensor<T> GetGateValues() |
Gets the current gate values (for interpretability). |
Vector<T> GetFeatureImportance() |
Gets feature importance based on gate activation magnitudes. |
void ResetState() |
|
void UpdateParameters(T learningRate) |