REaLTabFormer: Generating Realistic Relational and Tabular Data using Transformers

TL;DR

REaLTabFormer is a novel transformer-based model that generates realistic relational and tabular data, effectively capturing structure and improving prediction tasks without fine-tuning.

Contribution

It introduces a new generative framework combining autoregressive and sequence-to-sequence transformers for relational data synthesis.

Findings

Outperforms baseline models in capturing relational structure

Achieves state-of-the-art results on prediction tasks

Effective in generating large non-relational datasets

Abstract

Tabular data is a common form of organizing data. Multiple models are available to generate synthetic tabular datasets where observations are independent, but few have the ability to produce relational datasets. Modeling relational data is challenging as it requires modeling both a "parent" table and its relationships across tables. We introduce REaLTabFormer (Realistic Relational and Tabular Transformer), a tabular and relational synthetic data generation model. It first creates a parent table using an autoregressive GPT-2 model, then generates the relational dataset conditioned on the parent table using a sequence-to-sequence (Seq2Seq) model. We implement target masking to prevent data copying and propose the $Q_{\delta}$ statistic and statistical bootstrapping to detect overfitting. Experiments using real-world datasets show that REaLTabFormer captures the relational structure better than a baseline model. REaLTabFormer also achieves state-of-the-art results on prediction tasks, "out-of-the-box", for large non-relational datasets without needing fine-tuning.