GuidedBackpropExplanation<T>

Models & Types · AiDotNet.Interpretability.Explainers

Result of Guided Backpropagation explanation.

Constructors

Constructor Summary
GuidedBackpropExplanation(Vector<T> input, Vector<T> guidedGradients, int targetClass, T prediction, int[] inputShape, Tensor<T> gradientTensor) Initializes a new explanation.

Properties

Property Summary
T Prediction { get; } Gets the prediction for the target class.
Tensor<T> GradientTensor { get; } Gets the gradient tensor (for tensor inputs).
Vector<T> GuidedGradients { get; } Gets the guided gradients (attributions).
Vector<T> Input { get; } Gets the original input.
int TargetClass { get; } Gets the target class that was explained.
int[] InputShape { get; } Gets the input shape (for tensor inputs).

Methods

Method Summary
IEnumerable<ValueTuple<int, T>> GetTopFeatures(int k) Gets the most important input features.
Vector<T> GetNormalizedGradients() Normalizes gradients to [0, 1] range.
string ToString() Returns string representation.