RecTable: Fast Modeling Tabular Data with Rectified Flow
Masane Fuchi, Tomohiro Takagi
TL;DR
RecTable tackles the expensive training costs of diffusion and LLM-based tabular data generation by employing rectified flow with a lightweight GLU-based network. It introduces a mixed-type noise model for numerical and categorical features and uses a logit-normal timestep distribution, while avoiding the reflow step to speed up generation. Across six real-world datasets, RecTable achieves competitive fidelity and superior machine-learning efficiency, often with substantially faster training than state-of-the-art baselines. The results indicate rectified flow, combined with targeted architectural and training choices, as a viable path toward high-quality, time-efficient tabular data synthesis that could surpass diffusion-based approaches with further improvements.
Abstract
Score-based or diffusion models generate high-quality tabular data, surpassing GAN-based and VAE-based models. However, these methods require substantial training time. In this paper, we introduce RecTable, which uses the rectified flow modeling, applied in such as text-to-image generation and text-to-video generation. RecTable features a simple architecture consisting of a few stacked gated linear unit blocks. Additionally, our training strategies are also simple, incorporating a mixed-type noise distribution and a logit-normal timestep distribution. Our experiments demonstrate that RecTable achieves competitive performance compared to the several state-of-the-art diffusion and score-based models while reducing the required training time. Our code is available at https://github.com/fmp453/rectable.