CardiCat: a Variational Autoencoder for High-Cardinality Tabular Data

TL;DR

CardiCat tackles the challenge of high-cardinality categorical features in mixed-type tabular data by adopting a variational autoencoder with regularized dual encoder-decoder embedding layers that are learned jointly. This design yields covariate-dependent embeddings and a compact parameterization, enabling scalable learning on large datasets. The approach demonstrates competitive synthetic data quality across real and simulated datasets and provides a public implementation for evaluation. Overall, CardiCat advances high-fidelity, scalable tabular data synthesis with improved handling of imbalanced and high-cardinality features, benefiting privacy-preserving analytics and data sharing.

Abstract

High-cardinality categorical features are a common characteristic of mixed-type tabular datasets. Existing generative model architectures struggle to learn the complexities of such data at scale, primarily due to the difficulty of parameterizing the categorical features. In this paper, we present a general variational autoencoder model, CardiCat, that can accurately fit imbalanced high-cardinality and heterogeneous tabular data. Our method substitutes one-hot encoding with regularized dual encoder-decoder embedding layers, which are jointly learned. This approach enables us to use embeddings that depend also on the other covariates, leading to a compact and homogenized parameterization of categorical features. Our model employs a considerably smaller trainable parameter space than competing methods, enabling learning at a large scale. CardiCat generates high-quality synthetic data that better represent high-cardinality and imbalanced features compared to competing VAE models for multiple real and simulated datasets.

Figures (1)

• Figure 1: illustration and description of CardiCat's network architecture.