Universal Embeddings of Tabular Data

2025-07-08

AI Summary

Industrial relational databases contain rich tabular data, yet downstream tasks—such as regression, classification, and anomaly detection—are often undefined during database construction, necessitating task-agnostic universal table embedding methods. To address this, we propose a two-stage graph autoencoder framework: first, modeling tables as heterogeneous graphs to capture semantic and structural relationships among fields; second, learning entity-level embeddings via a graph autoencoder and aggregating them into row-level representations. Our approach enables zero-shot generalization—adapting to novel downstream tasks without task-specific fine-tuning. Extensive experiments on multiple real-world datasets demonstrate that our method significantly outperforms existing universal table embedding baselines across diverse downstream tasks, exhibiting superior generalizability and robustness.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Reasoning under Uncertainty: Relational Probabilistic Models
• Knowledge Representation and Reasoning: Automated Reasoning and Theorem Proving

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Tabular data in relational databases represents a significant portion of industrial data. Hence, analyzing and interpreting tabular data is of utmost importance. Application tasks on tabular data are manifold and are often not specified when setting up an industrial database. To address this, we present a novel framework for generating universal, i.e., task-independent embeddings of tabular data for performing downstream tasks without predefined targets. Our method transforms tabular data into a graph structure, leverages Graph Auto-Encoders to create entity embeddings, which are subsequently aggregated to obtain embeddings for each table row, i.e., each data sample. This two-step approach has the advantage that unseen samples, consisting of similar entities, can be embedded without additional training. Downstream tasks such as regression, classification or outlier detection, can then be performed by applying a distance-based similarity measure in the embedding space. Experiments on real-world datasets demonstrate that our method achieves superior performance compared to existing universal tabular data embedding techniques.

Problem

• Generating task-independent embeddings for tabular data
• Transforming tabular data into graph structures for embeddings
• Enabling downstream tasks without predefined targets

Innovation

• Transforms tabular data into graph structure
• Uses Graph Auto-Encoders for entity embeddings
• Aggregates embeddings for downstream tasks