From Features to Structure: Task-Aware Graph Construction for Relational and Tabular Learning with GNNs

2025-06-02

AI Summary

Real-world tabular and relational data are ubiquitous, yet mainstream deep learning methods struggle to effectively model their structured characteristics; existing GNN approaches rely on fixed-schema graphs (e.g., primary-foreign key joins), overlooking discriminative signals in non-key attributes. This paper proposes auGraph—a task-aware dynamic graph construction framework that, for the first time, promotes non-key attributes to graph nodes based on their downstream-task relevance, enabling joint optimization of graph structure generation and predictive objectives. Our method comprises a learnable scoring function for graph augmentation, end-to-end GNN training, and joint feature-structure embedding. Across diverse tabular and relational prediction tasks, auGraph significantly outperforms both schema-based and heuristic graph construction baselines, achieving superior accuracy, generalizability, and interpretability—while preserving schema consistency.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Tabular and relational data remain the most ubiquitous formats in real-world machine learning applications, spanning domains from finance to healthcare. Although both formats offer structured representations, they pose distinct challenges for modern deep learning methods, which typically assume flat, feature-aligned inputs. Graph Neural Networks (GNNs) have emerged as a promising solution by capturing structural dependencies within and between tables. However, existing GNN-based approaches often rely on rigid, schema-derived graphs -- such as those based on primary-foreign key links -- thereby underutilizing rich, predictive signals in non key attributes. In this work, we introduce auGraph, a unified framework for task-aware graph augmentation that applies to both tabular and relational data. auGraph enhances base graph structures by selectively promoting attributes into nodes, guided by scoring functions that quantify their relevance to the downstream prediction task. This augmentation preserves the original data schema while injecting task-relevant structural signal. Empirically, auGraph outperforms schema-based and heuristic graph construction methods by producing graphs that better support learning for relational and tabular prediction tasks.

Problem

• Enhancing graph structures for tabular and relational data learning
• Addressing underutilization of non-key attributes in GNNs
• Improving task-aware graph construction for better prediction performance

Innovation

• Task-aware graph augmentation for tabular data
• Selective attribute promotion guided by scoring
• Preserves schema while enhancing structural signals