Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Graph Neural Networks for Tabular Data Learning: A Survey with Taxonomy and Directions

Cheng-Te Li, Yu-Che Tsai, Chih-Yao Chen, Jay Chiehen Liao

TL;DR

This survey systematically reviews Graph Neural Networks for Tabular Data Learning (GNN4TDL), arguing that modeling latent instance and feature correlations via graphs can surpass traditional tabular methods. It introduces a four-axis taxonomy—graph formulation, graph construction, representation learning, and training plans—and maps a wide range of methods (homogeneous, heterogeneous, and hypergraphs) to these axes. The paper covers practical graph-construction strategies, specialized GNN designs for tabular data, and training regimens, illustrating applications in fraud detection, CTR prediction, medical prediction, missing data imputation, FinTech, and relational databases. It also discusses limitations and future directions, including graph-transformer hybrids, scalability, SSL, robustness, and non-homogeneous graph learning, to spur further research in GNN-enabled tabular data analysis.

Abstract

In this survey, we dive into Tabular Data Learning (TDL) using Graph Neural Networks (GNNs), a domain where deep learning-based approaches have increasingly shown superior performance in both classification and regression tasks compared to traditional methods. The survey highlights a critical gap in deep neural TDL methods: the underrepresentation of latent correlations among data instances and feature values. GNNs, with their innate capability to model intricate relationships and interactions between diverse elements of tabular data, have garnered significant interest and application across various TDL domains. Our survey provides a systematic review of the methods involved in designing and implementing GNNs for TDL (GNN4TDL). It encompasses a detailed investigation into the foundational aspects and an overview of GNN-based TDL methods, offering insights into their evolving landscape. We present a comprehensive taxonomy focused on constructing graph structures and representation learning within GNN-based TDL methods. In addition, the survey examines various training plans, emphasizing the integration of auxiliary tasks to enhance the effectiveness of instance representations. A critical part of our discussion is dedicated to the practical application of GNNs across a spectrum of GNN4TDL scenarios, demonstrating their versatility and impact. Lastly, we discuss the limitations and propose future research directions, aiming to spur advancements in GNN4TDL. This survey serves as a resource for researchers and practitioners, offering a thorough understanding of GNNs' role in revolutionizing TDL and pointing towards future innovations in this promising area.

Graph Neural Networks for Tabular Data Learning: A Survey with Taxonomy and Directions

TL;DR

This survey systematically reviews Graph Neural Networks for Tabular Data Learning (GNN4TDL), arguing that modeling latent instance and feature correlations via graphs can surpass traditional tabular methods. It introduces a four-axis taxonomy—graph formulation, graph construction, representation learning, and training plans—and maps a wide range of methods (homogeneous, heterogeneous, and hypergraphs) to these axes. The paper covers practical graph-construction strategies, specialized GNN designs for tabular data, and training regimens, illustrating applications in fraud detection, CTR prediction, medical prediction, missing data imputation, FinTech, and relational databases. It also discusses limitations and future directions, including graph-transformer hybrids, scalability, SSL, robustness, and non-homogeneous graph learning, to spur further research in GNN-enabled tabular data analysis.

Abstract

In this survey, we dive into Tabular Data Learning (TDL) using Graph Neural Networks (GNNs), a domain where deep learning-based approaches have increasingly shown superior performance in both classification and regression tasks compared to traditional methods. The survey highlights a critical gap in deep neural TDL methods: the underrepresentation of latent correlations among data instances and feature values. GNNs, with their innate capability to model intricate relationships and interactions between diverse elements of tabular data, have garnered significant interest and application across various TDL domains. Our survey provides a systematic review of the methods involved in designing and implementing GNNs for TDL (GNN4TDL). It encompasses a detailed investigation into the foundational aspects and an overview of GNN-based TDL methods, offering insights into their evolving landscape. We present a comprehensive taxonomy focused on constructing graph structures and representation learning within GNN-based TDL methods. In addition, the survey examines various training plans, emphasizing the integration of auxiliary tasks to enhance the effectiveness of instance representations. A critical part of our discussion is dedicated to the practical application of GNNs across a spectrum of GNN4TDL scenarios, demonstrating their versatility and impact. Lastly, we discuss the limitations and propose future research directions, aiming to spur advancements in GNN4TDL. This survey serves as a resource for researchers and practitioners, offering a thorough understanding of GNNs' role in revolutionizing TDL and pointing towards future innovations in this promising area.
Paper Structure (34 sections, 2 equations, 2 figures, 9 tables)

This paper contains 34 sections, 2 equations, 2 figures, 9 tables.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. Tabular Data Learning
  4. Notations on Graphs
  5. Graph Neural Networks (GNNs)
  6. Downstream Tasks
  7. Why are GNNs required for TDL?
  8. Pipeline & Categorization
  9. Methods
  10. Graph Formulation
  11. Homogeneous Graphs
  12. Heterogeneous Graphs
  13. Hypergraphs
  14. Graph Construction
  15. Intrinsic Structure
  16. ...and 19 more sections

Figures (2)

  • Figure 1: The general pipeline of a GNN-based tabular data learning model. Given the tabular data as the model input, we first formulate the graph structure considering data instances and features, construct the graph structure, learn feature representations using GNNs, and train to generate the prediction outcomes, along with some auxiliary tasks.
  • Figure 2: The taxonomy systematically organizes GNNs for tabular data learning along four axes: graph formulation, graph construction, graph representation learning, and tabular training.