Skip to content

Latest commit

 

History

History
148 lines (123 loc) · 16 KB

README.md

File metadata and controls

148 lines (123 loc) · 16 KB
Raw

Graph Neural Networks for Tabular Data Learning (GNN4TDL)

Table of contents

👉 If you notice any errors or have suggestions, feel free to share them with us!

⭐️ If you find this resource helpful, please consider to star this repository and cite our survey paper:

@misc{li2024graph,
      title={Graph Neural Networks for Tabular Data Learning: A Survey with Taxonomy and Directions}, 
      author={Cheng-Te Li and Yu-Che Tsai and Chih-Yao Chen and Jay Chiehen Liao},
      year={2024},
      eprint={2401.02143},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}

✨ News

  • [2023-12-16] We have released this repository that collects the resources related to GNNs for tabular data learning (GNN4TDL).

Intro: Graph Neural Networks for Tabular Data Learning

The deep learning-based approaches to Tabular Data Learning (TDL), classification and regression, have shown competing performance, compared to their conventional counterparts. However, the latent correlation among data instances and feature values is less modeled in deep neural TDL. Recently, graph neural networks (GNNs), which can enable modeling relations and interactions between different tabular data elements, has received tremendous attention across application domains including TDL. It turns out creating proper graph structures from the input tabular dataset, along with GNN learning, can improve the TDL performance. In this survey, we systematically review the methodologies of designing and applying GNNs for TDL (GNN4TDL). The topics to be covered include: (1) foundations and overview of GNN-based TDL methods; (2) a comprehensive taxonomy of constructing graph structures and representation learning in GNN-based TDL methods; (3) how to apply GNN to various TDL application scenarios and tasks; (4) limitations in current research and future directions.

taxnomy2-1

This survey presents an in-depth exploration into the application of GNNs in tabular data learning. It starts by establishing the fundamental problem statement and introduces various graph types used to represent tabular data. The survey is structured around a detailed GNN-based learning pipeline, encompassing phases like Graph Formulation, where tabular elements are converted into graph nodes; Graph Construction, focusing on establishing connections within these elements; Representation Learning, highlighting how GNNs process these structures to learn data instance features; and Training Plans, discussing the integration of auxiliary tasks for enhanced predictive outcomes.

Homogeneous Instance Graphs

Year Title Venue Paper Code
2023 EGG-GAE: Scalable Graph Neural Networks for Tabular Data Imputation ICML 2023 Link Link
2023 Graph Neural Networks for Missing Value Classification in a Task-driven Metric Space TKDE 2023 Link
2023 Look Around! A Neighbor Relation Graph Learning Framework for Real Estate Appraisal AAAI 2023 Link
2023 Homophily-Enhanced Self-Supervision for Graph Structure Learning: Insights and Directions TNNLS 2023 Link Link
2023 TabGSL: Graph Structure Learning for Tabular Data Prediction arXiv 2023 Link
2023 Oversmoothing Relief Graph Convolutional Network-Based Fault Diagnosis Method with Application to the Rectifier of High-Speed Trains IEEE TII 2023 Link
2022 Self-Adaptation Graph Attention Network via Meta-Learning for Machinery Fault Diagnosis with Few Labeled Data IEEE TIM 2022 Link
2023 A Graph Neural Network-based Bearing Fault Detection Method Scientific Reports 2023 Link
2023 Latent Graphs for Semi-Supervised Learning on Biomedical Tabular Data IJCLR 2023 Link
2023 Interpretable Graph Neural Networks for Tabular Data arXiv 2023 Link
2022 Differentiable graph module (dgm) for graph convolutional networks IEEE TPAMI 2022 Link Link
2022 Towards Unsupervised Deep Graph Structure Learning WWW 2022 Link Link
2022 LUNAR: Unifying Local Outlier Detection Methods via Graph Neural Networks AAAI 2022 Link Link
2022 Graph Regularized Autoencoder and its Application in Unsupervised Anomaly Detection IEEE TPAMI 2022 Link
2022 Deep Unsupervised Active Learning on Learnable Graphs IEEE TNNLS 2022 Link
2022 Graph autoencoder-based unsupervised outlier detection Information Sciences 2022 Link Link
2022 Structure-Aware Siamese Graph Neural Networks for Encounter-Level Patient Similarity Learning Journal of Biomedical Informatics 2022 Link
2022 GEDI: A Graph-based End-to-end Data Imputation Framework arXiv 2022 Link
2021 Self-supervision improves structure learning for graph neural networks. NeurIPS 2021 Link Link
2021 Predicting Patient Outcomes with Graph Representation Learning DLG-AAAI 2021 Link Link
2021 A Weighted Patient Network-based Framework for Predicting Chronic Diseases Using Graph Neural Networks Scientific Reports 2021 Link
2021 k-Nearest Neighbor Learning with Graph Neural Networks Mathematics 2021 Link
2020 Iterative Deep Graph Learning for Graph Neural Networks: Better and Robust Node Embeddings NeurIPS 2020 Link Link
2020 Missing Data Imputation with Adversarially-trained Graph Convolutional Networks Neural Networks 2020 Link Link
2019 Learning Discrete Structures for Graph Neural Networks ICML 2019 Link Link

Homogeneous Feature Graphs

Year Title Venue Paper Code
2023 T2G-Former: Organizing Tabular Features into Relation Graphs Promotes Heterogeneous Feature Interaction AAAI 2023 Link Link
2023 Deep Tabular Data Modeling With Dual-Route Structure-Adaptive Graph Networks IEEE TKDE 2023 Link
2023 Causality-based CTR Prediction using Graph Neural Networks Information Processing & Management 2023 Link
2023 Graph Neural Network Contextual Embedding for Deep Learning on Tabular Data Arxiv 2023 Link Link
2023 FT-GAT: Graph Neural Network for Predicting Spontaneous Breathing Trial Success in Patients with Mechanical Ventilation Computer Methods and Programs in Biomedicine 2023 Link
2022 Table2Graph: Transforming Tabular Data to Unifed Weighted Graph IJCAI 2022 Link
2022 Local Contrastive Feature Learning for Tabular Data CIKM 2022 Link
2021 FIVES: Feature Interaction Via Edge Search for Large-Scale Tabular Data KDD 2021 Link
2021 TabularNet: A Neural Network Architecture for Understanding Semantic Structures of Tabular Data KDD 2021 Link
2021 GCN-Int: A Click-Through Rate Prediction Model Based on Graph Convolutional Network Interaction IEEE Access 2021 Link
2019 Fi-GNN: Modeling Feature Interactions via Graph Neural Networks for CTR Prediction CIKM 2019 Link Link

Bipartite Graphs

Year Title Venue Paper Code
2023 Data Imputation with Iterative Graph Reconstruction AAAI 2023 Link
2022 Learning Enhanced Representations for Tabular Data via Neighborhood Propagation NeurIPS 2022 Link Link
2022 Relational Multi-Task Learning: Modeling Relations between Data and Tasks ICLR 2022 Link Link
2022 Predicting the Survival of Cancer Patients with Multimodal Graph Neural Network IEEE TCBB 2022 Link
2021 Towards Open-World Feature Extrapolation: An Inductive Graph Learning Approach NeurIPS 2021 Link Link
2021 MedGraph: Structural and Temporal Representation Learning of Electronic Medical Records ECAI 2021 Link Link
2021 Disease Prediction via Graph Neural Networks IEEE JBHI 2021 Link Link
2020 Handling Missing Data with Graph Representation Learning NeurIPS 2020 Link Link
2019 Large-Scale Heterogeneous Feature Embedding AAAI 2019 Link Link

Heterogeneous Graphs

Year Title Venue Paper Code
2023 Relational Deep Learning: Graph Representation Learning on Relational Databases arXiv 2023 Link Link
2023 GCondNet: A Novel Method for Improving Neural Networks on Small High-Dimensional Tabular Data NeurIPS 2023 TRL Workshop Link
2023 Lifelong Property Price Prediction: A Case Study for the Toronto Real Estate Market IEEE TKDE 2023 Link Link
2023 GraphFC: Customs Fraud Detection with Label Scarcity CIKM 2023 Link Link
2023 GFS: Graph-based Feature Synthesis for Prediction over Relational Databases arXiv 2023 Link
2022 GCF-RD: A Graph-based Contrastive Framework for Semi-Supervised Learning on Relational Databases ACM CIKM 2022 Link Link
2022 Reinforced Neighborhood Selection Guided Multi-Relational Graph Neural Networks ACM TOIS 2022 Link Link
2022 H2-Fdetector: A GNN-based Fraud Detector with Homophilic and Heterophilic Connections WWW 2022 Link
2022 xFraud: Explainable Fraud Transaction Detection VLDB 2022 Link Link
2022 AUC-oriented Graph Neural Network for Fraud Detection WWW 2022 Link
2022 Hierarchical Multi-Modal Fusion on Dynamic Heterogeneous Graph for Health Insurance Fraud Detection IEEE ICME 2022 Link
2021 Pick and Choose: A GNN-based Imbalanced Learning Approach for Fraud Detection WWW 2021 Link Link
2021 Learning Graph Meta Embeddings for Cold-Start Ads in Click-Through Rate Prediction SIGIR 2021 Link Link
2021 Modeling Heterogeneous Graph Network on Fraud Detection: A Community-based Framework with Attention Mechanism CIKM 2021 Link
2021 Towards Consumer Loan Fraud Detection: Graph Neural Networks with Role-Constrained Conditional Random Field AAAI 2021 Link
2021 Intention-Aware Heterogeneous Graph Attention Networks for Fraud Transactions Detection KDD 2021 Link
2021 TabGNN: Multiplex Graph Neural Network for Tabular Data Prediction DLP-KDD 2021 Link Link
2020 Enhancing Graph Neural Network-based Fraud Detectors against Camouflaged Fraudsters CIKM 2020 Link Link
2020 Loan Default Analysis with Multiplex Graph Learning ACM CIKM 2020 Link
2020 Financial Risk Analysis for SMEs with Graph-based Supply Chain Mining IJCAI 2020 Link
2020 Heterogeneous Similarity Graph Neural Network on Electronic Health Records IEEE BigData 2020 Link
2020 Supervised Learning on Relational Databases with Graph Neural Networks ICLR 2020 Link Link
2020 Learning the Graphical Structure of Electronic Health Records with Graph Convolutional Transformer AAAI 2020 Link Link

Hypergraphs

Year Title Venue Paper Code
2023 HYTREL: Hypergraph-enhanced Tabular Data Representation Learning NeurIPS 2023 Link Link
2022 Learning Enhanced Representations for Tabular Data via Neighborhood Propagation NeurIPS 2022 Link Link
2022 Hypergraph Contrastive Learning for Electronic Health Records SDM 2022 Link

Knowledge Graphs

Year Title Venue Paper Code
2023 High dimensional, tabular deep learning with an auxiliary knowledge graph NeurIPS 2023 Link link