Structural Equation-VAE: Disentangled Latent Representations for Tabular Data

TL;DR

SE-VAE is a new deep generative model that embeds measurement structure into its architecture to produce interpretable, disentangled latent representations for tabular data, improving factor recovery and robustness.

Contribution

The paper introduces SE-VAE, a novel architecture that incorporates measurement structure directly into a variational autoencoder for better disentanglement of tabular data.

Findings

SE-VAE outperforms baseline models in disentanglement metrics.

Architectural design, not regularization, drives performance.

SE-VAE enhances interpretability and robustness to nuisance variation.

Abstract

Learning interpretable latent representations from tabular data remains a challenge in deep generative modeling. We introduce SE-VAE (Structural Equation-Variational Autoencoder), a novel architecture that embeds measurement structure directly into the design of a variational autoencoder. Inspired by structural equation modeling, SE-VAE aligns latent subspaces with known indicator groupings and introduces a global nuisance latent to isolate construct-specific confounding variation. This modular architecture enables disentanglement through design rather than through statistical regularizers alone. We evaluate SE-VAE on a suite of simulated tabular datasets and benchmark its performance against a series of leading baselines using standard disentanglement metrics. SE-VAE consistently outperforms alternatives in factor recovery, interpretability, and robustness to nuisance variation. Ablation results reveal that architectural structure, rather than regularization strength, is the key driver of performance. SE-VAE offers a principled framework for white-box generative modeling in scientific and social domains where latent constructs are theory-driven and measurement validity is essential.