TabTransformerGenOptions<T>

Options & Configuration · AiDotNet.Models.Options

Configuration options for TabTransformer-Gen, a generative model that uses contextual embeddings from multi-head self-attention over categorical columns to generate realistic tabular data.

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

For Beginners

TabTransformer-Gen works like a fill-in-the-blank game:

  1. Each categorical column gets its own “word” (embedding)
  2. These words “talk to each other” through attention (e.g., “Occupation”

pays attention to “Education” because they’re related)

  1. During training, we randomly mask some columns and ask the model to guess them
  2. During generation, we start with everything masked and iteratively fill in columns

Example:

How It Works

TabTransformer-Gen adapts the TabTransformer architecture for data generation:

  • Categorical columns get learned embeddings that attend to each other
  • Continuous columns pass through with optional normalization
  • A masked prediction objective trains the model to reconstruct missing columns
  • Generation uses iterative masked prediction (like masked language modeling)

Reference: “TabTransformer: Tabular Data Modeling Using Contextual Embeddings” (Huang et al., 2020) — adapted for generation with masked prediction

Constructors

Constructor Summary
TabTransformerGenOptions()

Properties

Property Summary
double DropoutRate { get; set; } Gets or sets the dropout rate.
double LearningRate { get; set; } Gets or sets the learning rate.
double MaskRatio { get; set; } Gets or sets the fraction of columns to mask during training.
int BatchSize { get; set; } Gets or sets the training batch size.
int EmbeddingDimension { get; set; } Gets or sets the embedding dimension for each column.
int Epochs { get; set; } Gets or sets the number of training epochs.
int FeedForwardDimension { get; set; } Gets or sets the feed-forward dimension in each transformer block.
int GenerationSteps { get; set; } Gets or sets the number of iterative refinement steps during generation.
int NumHeads { get; set; } Gets or sets the number of attention heads per layer.
int NumLayers { get; set; } Gets or sets the number of transformer layers.
int VGMModes { get; set; } Gets or sets the number of VGM modes for continuous column transformation.