TabSynGenerator<T>

Models & Types · AiDotNet.NeuralNetworks.SyntheticData

TabSyn generator combining VAE pretraining with latent diffusion for state-of-the-art synthetic tabular data generation.

Inheritance: NeuralNetworkBase<T> → NeuralSyntheticTabularGeneratorBase<T> → TabSynGenerator<T>

Implements: ISyntheticTabularGenerator<T>

For Beginners

TabSyn is a two-step generator:

**Step 1 - VAE Training (learning to compress):****Step 2 - Diffusion Training (learning the latent distribution):**Generation:

If you provide custom layers in the architecture, those will be used directly for the VAE encoder. If not, the network creates industry-standard TabSyn encoder layers based on the original research paper specifications.

Example usage:

How It Works

TabSyn trains in two phases:

  1. VAE Phase: Encoder-decoder learns latent representation of tabular data
  2. Diffusion Phase: Gaussian diffusion model learns the distribution of latent codes

Generation: z ~ DiffusionModel -> decoded = VAEDecoder(z) -> inverse transform

This implementation follows the standard neural network architecture pattern with:

  • Proper inheritance from NeuralNetworkBase
  • Layer-based architecture using ILayer components
  • Engine-based tensor operations for CPU/GPU acceleration
  • Full autodiff and JIT compilation support

Reference: “TabSyn: Bridging the Gap” (Zhang et al., NeurIPS 2023)

Constructors

Constructor Summary
TabSynGenerator() Initializes a new TabSyn generator with the specified architecture.
TabSynGenerator(NeuralNetworkArchitecture<T> architecture, TabSynOptions<T> options, IGradientBasedOptimizer<T, Tensor<T>, Tensor<T>> optimizer, ILossFunction<T> lossFunction)

Properties

Property Summary
IReadOnlyList<ColumnMetadata> Columns { get; }
TabSynOptions<T> Options { get; } Gets the TabSyn-specific options.
bool IsFitted { get; set; }

Methods

Method Summary
Dictionary<string, T> GetFeatureImportance()
Matrix<T> Generate(int numSamples, Vector<T> conditionColumn, Vector<T> conditionValue) Generates new synthetic tabular data rows.
ModelMetadata<T> GetModelMetadata()
Task FitAsync(Matrix<T> data, IReadOnlyList<ColumnMetadata> columns, int? epochs, CancellationToken ct)
void Fit(Matrix<T> data, IReadOnlyList<ColumnMetadata> columns, int? epochs) Fits the TabSyn generator to the provided real tabular data.