A Locality Radius Framework for Understanding Relational Inductive Bias in Database Learning

TL;DR

This paper introduces the locality radius, a measure of the minimal neighborhood needed for predictions in relational schemas, and demonstrates its importance in aligning model architecture with task complexity for improved performance.

Contribution

It formalizes the concept of locality radius and empirically shows its critical role in model performance across various relational learning tasks.

Findings

Performance depends on alignment between task locality radius and model aggregation depth.

Bias-radius alignment significantly affects predictive accuracy.

Empirical results confirm the importance of locality considerations in relational learning.

Abstract

Foreign key discovery and related schema-level prediction tasks are often modeled using graph neural networks (GNNs), implicitly assuming that relational inductive bias improves performance. However, it remains unclear when multi-hop structural reasoning is actually necessary. In this work, we introduce locality radius, a formal measure of the minimum structural neighborhood required to determine a prediction in relational schemas. We hypothesize that model performance depends critically on alignment between task locality radius and architectural aggregation depth. We conduct a controlled empirical study across foreign key prediction, join cost estimation, blast radius regression, cascade impact classification, and additional graph-derived schema tasks. Our evaluation includes multi-seed experiments, capacity-matched comparisons, statistical significance testing, scaling analysis, and synthetic radius-controlled benchmarks. Results reveal a consistent bias-radius alignment effect.