This repository is a curated list of resources for self-supervised learning (SSL) on graphs, targeting researchers and practitioners in graph representation learning. It aims to consolidate key papers, surveys, blog posts, and talks, categorizing methods into generative/predictive and contrastive approaches to facilitate understanding and adoption of SSL techniques in the graph domain.

How It Works

The list categorizes self-supervised graph representation learning methods into two primary paradigms: generative/predictive approaches that optimize in the output space, and contrastive methods that optimize in the latent space. This division helps users navigate the rapidly evolving field by understanding the core mechanisms and trade-offs between different SSL strategies for graphs.

Highlighted Details

• Comprehensive categorization of papers into generative/predictive and contrastive methods.
• Includes links to official papers (PDFs) and associated code where available.
• Covers a wide range of applications, including node representation learning, graph classification, recommendation systems, and anomaly detection.
• Features relevant surveys, blog posts, and talks for a holistic understanding of the field.

Maintenance & Community

This is a community-driven "awesome" list, encouraging contributions via pull requests to add new resources. Specific maintainers or community channels are not detailed in the README.

Licensing & Compatibility

The repository itself is not software but a curated list; therefore, it does not have a software license. The linked papers and code repositories will have their own respective licenses.

Limitations & Caveats

As a curated list, it relies on community contributions for updates and may not be exhaustive or perfectly up-to-date with the very latest research. The quality and maintenance of linked code repositories vary.