Self-supervised learning for graph-structured data in healthcare applications: A comprehensive review

Area of Science:

• Graph-structured data analysis
• Machine learning in healthcare
• Biomedical informatics

Background:

• Healthcare data is increasingly complex and interconnected.
• Graph data structures excel at modeling these relationships.
• Limited labeled data necessitates advanced learning techniques like self-supervised learning (SSL).

Purpose of the Study:

• To provide a comprehensive review of SSL methods for graph-structured data in healthcare.
• To explore challenges and opportunities in applying SSL to healthcare data.
• To assess the effectiveness of SSL in various healthcare applications.

Main Methods:

• Systematic review of existing literature on SSL for graph data in healthcare.
• Analysis of SSL techniques applied to disease prediction, medical image analysis, and drug discovery.
• Critical evaluation of SSL method performance, strengths, and limitations.

Main Results:

• SSL offers a powerful paradigm for learning representations from unlabeled graph-structured healthcare data.
• SSL techniques show promise across diverse healthcare applications, including prediction, diagnosis, and discovery.
• The review identifies key challenges and future research directions for SSL in this domain.

Conclusions:

• Self-supervised learning is a crucial tool for unlocking the potential of graph-structured healthcare data.
• This paper serves as a foundational resource for researchers and practitioners in the field.
• SSL application in healthcare graph data is poised to significantly enhance patient outcomes and accelerate medical advancements.