CardiCat: a Variational Autoencoder for High-Cardinality Tabular Data

TL;DR

CardiCat is a novel variational autoencoder designed to effectively model and generate high-cardinality, imbalanced tabular data using joint embedding layers, outperforming existing models in quality and scalability.

Contribution

It introduces a regularized dual encoder-decoder embedding approach that reduces parameters and improves modeling of complex categorical features in tabular data.

Findings

Generates high-quality synthetic data for complex categorical features.

Uses fewer parameters than competing models, enabling large-scale learning.

Outperforms existing VAEs in representing imbalanced high-cardinality data.

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.