Class WorldModelsOptions<T>

Namespace

AiDotNet.Models.Options

Assembly

AiDotNet.dll

Configuration options for World Models agents.

public class WorldModelsOptions<T> : ReinforcementLearningOptions<T>

Type Parameters

T

The numeric type used for calculations.

Inheritance

object

ReinforcementLearningOptions<T>

WorldModelsOptions<T>

Inherited Members

ReinforcementLearningOptions<T>.LearningRate

ReinforcementLearningOptions<T>.DiscountFactor

ReinforcementLearningOptions<T>.LossFunction

ReinforcementLearningOptions<T>.Seed

ReinforcementLearningOptions<T>.BatchSize

ReinforcementLearningOptions<T>.ReplayBufferSize

ReinforcementLearningOptions<T>.TargetUpdateFrequency

ReinforcementLearningOptions<T>.UsePrioritizedReplay

ReinforcementLearningOptions<T>.EpsilonStart

ReinforcementLearningOptions<T>.EpsilonEnd

ReinforcementLearningOptions<T>.EpsilonDecay

ReinforcementLearningOptions<T>.WarmupSteps

ReinforcementLearningOptions<T>.MaxGradientNorm

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Remarks

World Models learns compact spatial and temporal representations using VAE and RNN. The agent learns entirely within the "dream" of its learned world model.

For Beginners: World Models is inspired by how humans learn: we build mental models of the world, then make decisions based on those models rather than raw sensory input.

Key components:

• VAE (V): Compresses visual observations into compact latent codes
• MDN-RNN (M): Learns temporal dynamics (what happens next)
• Controller (C): Simple linear/neural policy acting in latent space
• Learning in Dreams: Agent trains entirely in imagined rollouts

Think of it like: First, learn to compress images (VAE). Then, learn how compressed images change over time (RNN). Finally, learn to act based on compressed predictions (controller).

Famous for: Car racing from pixels, learning with limited real environment samples

Properties

ActionSize

public int ActionSize { get; init; }

Property Value

int

ControllerGenerations

public int ControllerGenerations { get; init; }

Property Value

int

ControllerLayers

public List<int> ControllerLayers { get; init; }

Property Value

List<int>

ControllerPopulationSize

public int ControllerPopulationSize { get; init; }

Property Value

int

LatentSize

public int LatentSize { get; init; }

Property Value

int

NumMixtures

public int NumMixtures { get; init; }

Property Value

int

ObservationChannels

public int ObservationChannels { get; init; }

Property Value

int

ObservationHeight

public int ObservationHeight { get; init; }

Property Value

int

ObservationWidth

public int ObservationWidth { get; init; }

Property Value

int

Optimizer

The optimizer used for updating network parameters. If null, Adam optimizer will be used by default.

public IOptimizer<T, Vector<T>, Vector<T>>? Optimizer { get; init; }

Property Value

IOptimizer<T, Vector<T>, Vector<T>>

RNNEpochs

public int RNNEpochs { get; init; }

Property Value

int

RNNHiddenSize

public int RNNHiddenSize { get; init; }

Property Value

int

RNNLayers

public int RNNLayers { get; init; }

Property Value

int

RolloutLength

public int RolloutLength { get; init; }

Property Value

int

Temperature

public double Temperature { get; init; }

Property Value

double

VAEBeta

public double VAEBeta { get; init; }

Property Value

double

VAEEncoderChannels

public List<int> VAEEncoderChannels { get; init; }

Property Value

List<int>

VAEEpochs

public int VAEEpochs { get; init; }

Property Value

int