InverseProblemOptions<T>
Options & Configuration · AiDotNet.PhysicsInformed.Interfaces
Configuration options for inverse problem PINN training.
Inheritance: ModelOptions → InverseProblemOptions<T>
Constructors
| Constructor | Summary |
|---|---|
InverseProblemOptions() |
Properties
| Property | Summary |
|---|---|
InverseProblemRegularization Regularization { get; set; } |
The type of regularization to apply. |
T DataWeight { get; set; } |
Weight for the observation data loss relative to physics loss. |
T RegularizationStrength { get; set; } |
Regularization strength (λ in the formulas above). |
T[] PriorMeans { get; set; } |
Prior means for Bayesian regularization. |
T[] PriorStandardDeviations { get; set; } |
Prior standard deviations for Bayesian regularization. |
bool EstimateUncertainty { get; set; } |
Whether to estimate parameter uncertainty. |
bool LogParameterHistory { get; set; } |
Whether to log parameter estimates during training. |
bool UseSeparateLearningRates { get; set; } |
Whether to use separate learning rates for solution and parameters. |
double ParameterLearningRate { get; set; } |
Learning rate for the unknown parameters (if separate rates are used). |
int NumCollocationPoints { get; set; } |
Number of collocation points at which the PDE residual is enforced. |
int UncertaintySamples { get; set; } |
Number of samples for uncertainty estimation (if enabled). |