LiquidStateMachineOptions

Options & Configuration · AiDotNet.NeuralNetworks.Options

Configuration options for the LiquidStateMachine.

Inheritance: ModelOptions → NeuralNetworkOptions → ModelHyperparameterOptions → ReservoirOptions → LiquidStateMachineOptions

For Beginners

A Liquid State Machine is a reservoir model whose pool behaves like ripples in water — an input disturbs the surface, and the pattern of ripples carries a memory of it. Only the readout is trained. The values here are the ones it ships with.

Constructors

Constructor Summary
LiquidStateMachineOptions() Initializes a new instance carrying the model’s published defaults.
LiquidStateMachineOptions(LiquidStateMachineOptions other) Initializes a new instance by copying another instance.

Properties

Property Summary
double ConnectionProbability { get; set; } Gets or sets the probability that any two reservoir neurons are connected.
double InputScaling { get; set; } Gets or sets how strongly the input is scaled before entering the reservoir.
double ReadoutLearningRate { get; set; } Learning rate for the readout layer.

Methods

Method Summary
void Validate() Throws if a value this model requires has been left unset or is not positive.