TVAEOptions<T>

Options & Configuration · AiDotNet.Models.Options

Configuration options for TVAE (Tabular Variational Autoencoder), a VAE-based model for generating realistic synthetic tabular data.

Inheritance: ModelOptions → NeuralNetworkOptions → ModelHyperparameterOptions → FinancialNeuralNetworkOptions → RiskModelOptions<T> → SyntheticDataGeneratorOptions<T> → TVAEOptions<T>

For Beginners

TVAE learns to compress your data into a small “summary” (latent space) and then reconstruct it back. Think of it like learning a recipe:

  1. Encoder: Looks at a data row and writes a compact “recipe” (latent code)
  2. Decoder: Reads the recipe and recreates the data row
  3. Training: The model learns to write good recipes that can recreate realistic data
  4. Generation: Sample random recipes from the latent space and decode them into new rows

TVAE is often faster to train than CTGAN and works well for moderate-sized datasets.

Example:

How It Works

TVAE uses a variational autoencoder architecture adapted for tabular data:

  • Same VGM normalization and one-hot encoding as CTGAN for preprocessing
  • Encoder compresses data into a latent Gaussian distribution (mean, logvar)
  • Decoder reconstructs data from sampled latent codes via reparameterization trick
  • ELBO loss balances reconstruction quality with KL divergence regularization

Reference: “Modeling Tabular Data using Conditional GAN” (Xu et al., NeurIPS 2019)

Constructors

Constructor Summary
TVAEOptions() Initializes a new instance with the paper-default TVAE settings.
TVAEOptions(TVAEOptions<T> other) Initializes a new instance by copying all user-configurable settings.

Properties

Property Summary
double LearningRate { get; set; } Gets or sets the learning rate for the optimizer.
double LossWeight { get; set; } Gets or sets the weight for reconstruction loss relative to KL divergence.
int BatchSize { get; set; } Gets or sets the training batch size.
int Epochs { get; set; } Gets or sets the number of training epochs.
int LatentDimension { get; set; } Gets or sets the dimension of the latent space.
int VGMModes { get; set; } Gets or sets the number of Gaussian mixture components for VGM normalization.
int[] DecoderDimensions { get; set; } Gets or sets the hidden layer sizes for the decoder network.
int[] EncoderDimensions { get; set; } Gets or sets the hidden layer sizes for the encoder network.