Relational In-Context Learning via Synthetic Pre-training with Structural Prior

2026-03-04

AI Summary

Relational databases lack foundational models due to the privacy, scarcity, and structural heterogeneity of real-world data. This work proposes the first relational database foundation model trained entirely on synthetic data. It introduces a structural causal model–driven relational prior generator to construct large-scale, diverse synthetic databases from scratch, combined with DFS-based linearization and a lightweight neural architecture to pretrain the RDB-PFN model. The approach enables genuine in-context learning and significantly outperforms graph-based and single-table foundation model baselines across 19 real-world relational prediction tasks, while maintaining efficient inference and a compact model structure.

Technology Category

• Reasoning under Uncertainty: Relational Probabilistic Models
• Machine Learning: Deep Neural Architectures and Foundation Models
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Relational Databases (RDBs) are the backbone of modern business, yet they lack foundation models comparable to those in text or vision. A key obstacle is that high-quality RDBs are private, scarce and structurally heterogeneous, making internet-scale pre-training infeasible. To overcome this data scarcity, We introduce $\textbf{RDB-PFN}$, the first relational foundation model trained purely via $\textbf{synthetic data}$. Inspired by Prior-Data Fitted Networks (PFNs) where synthetic data generated from Structural Causal Models (SCMs) enables reasoning on single tables, we design a $\textbf{Relational Prior Generator}$ to create an infinite stream of diverse RDBs from scratch. Pre-training on $\textbf{over 2 million}$ synthetic single-table and relational tasks, RDB-PFN learns to adapt to any new database instantly via genuine $\textbf{in-context learning}$. Experiments verify RDB-PFN achieves strong few-shot performance on 19 real-world relational prediction tasks, outperforming graph-based and single-table foundation-model baselines (given the same DFS-linearized inputs), while using a lightweight architecture and fast inference. The code is available at https://github.com/MuLabPKU/RDBPFN

Problem

• Relational Databases
• Foundation Models
• Data Scarcity
• Structural Heterogeneity
• In-Context Learning

Innovation

• Relational Foundation Model
• Synthetic Data Pre-training
• In-Context Learning
• Structural Prior
• Relational Database