TabFlowOptions<T>

Options & Configuration · AiDotNet.Models.Options

Configuration options for TabFlow, a flow matching model for generating synthetic tabular data using continuous normalizing flows with optimal transport paths.

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

For Beginners

TabFlow learns to gradually transform random noise into realistic data through a smooth, deterministic path (no randomness during generation):

  1. Training: Learn a velocity field that moves noise toward data along straight paths
  2. Generation: Start from noise and follow the velocity field using an ODE solver

Advantages over diffusion models:

  • Faster generation (fewer steps needed since paths are straighter)
  • Deterministic sampling (same noise → same output)
  • Often higher quality for tabular data

Example:

How It Works

TabFlow uses flow matching (a continuous normalizing flow approach) to learn a deterministic mapping from noise to data. Unlike diffusion models that use stochastic differential equations, TabFlow uses ordinary differential equations (ODEs) with optimal transport conditional paths.

Reference: “Flow Matching for Tabular Data” (2024)

Constructors

Constructor Summary
TabFlowOptions()

Properties

Property Summary
double DropoutRate { get; set; } Gets or sets the dropout rate for the MLP.
double LearningRate { get; set; } Gets or sets the learning rate.
double Sigma { get; set; } Gets or sets the sigma for the optimal transport conditional flow.
int BatchSize { get; set; } Gets or sets the training batch size.
int Epochs { get; set; } Gets or sets the number of training epochs.
int NumSteps { get; set; } Gets or sets the number of ODE solver steps for generation.
int TimeEmbeddingDimension { get; set; } Gets or sets the dimension of the time embedding.
int VGMModes { get; set; } Gets or sets the number of VGM modes for continuous columns.
int[] MLPDimensions { get; set; } Gets or sets the hidden layer sizes for the velocity field MLP.
string Solver { get; set; } Gets or sets the ODE solver type.