RelICL: Training-free Relational Learning with Tabular Foundation Models

2026-10-01

AI Summary

This study addresses the feature explosion and interaction blind spots induced by deep feature synthesis in relational learning. To overcome these limitations, this work proposes RelICL, a method that departs from the conventional paradigm of generating aggregated features. Instead, it designs an iterative information propagation and fusion mechanism based on schema graphs, leveraging tabular foundation models to achieve training-free relational learning. Evaluated on the RelBench benchmark, RelICL attains performance comparable to state-of-the-art deep feature synthesis methods while effectively overcoming scalability bottlenecks. By doing so, it achieves a favorable balance between computational scalability and expressive feature interaction capability, offering a robust alternative for automated relational representation learning without requiring task-specific model training.

Abstract

Tabular foundation models achieve state-of-the-art performance on single-table tasks without any training. Recent work suggests that they are also well-suited for relational learning via deep feature synthesis (DFS), which flattens a relational schema into a single table by adding aggregates of the other tables' columns as features. This approach is appealing because it directly benefits from improvements to or customization of the underlying tabular foundation model. In this paper, we identify two key problems with DFS: feature explosion and interaction blindness. The first problem arises because the number of DFS features grows quickly as the schema becomes more complex, limiting scalability and performance. The second problem arises because column-wise aggregates do not account for feature interactions, limiting performance. We propose and explore an alternative method termed RelICL, which keeps the benefits of DFS but alleviates these two problems. At its heart, RelICL propagates and fuses information step by step through the schema graph, using the same tabular foundation model that is eventually used for prediction to do so. In our experimental study using RelBench tasks, RelICL was on par with the strongest approach based on deep feature synthesis.

Problem

• Relational Learning
• Tabular Foundation Models
• Deep Feature Synthesis
• Feature Explosion
• Interaction Blindness

Innovation

• Tabular Foundation Models
• Relational Learning
• Training-free
• Schema Graph
• Deep Feature Synthesis