This repository serves as a curated collection of essential research papers on Graph Foundation Models (GFMs), targeting researchers and practitioners in graph machine learning and natural language processing. It aims to consolidate the rapidly evolving landscape of GFMs, providing a structured overview of key advancements and enabling efficient literature review.

How It Works

The repository categorizes papers into distinct areas: Survey Papers, GNN-based Papers, LLM-based Papers, GNN+LLM-based Papers, Benchmarks and Datasets, and Other relevant works. Papers are tagged with keywords and organized by publication venue and year, facilitating targeted exploration of specific sub-fields within GFM research.

Quick Start & Requirements

• Access to the paper list is via the GitHub repository.
• No software installation is required; it's a curated list of academic papers.
• Links to PDFs and associated code/paperlists are provided for each entry.

Highlighted Details

• Comprehensive coverage of survey papers on GFMs, LLMs on graphs, and GNN+LLM integration.
• Includes papers on GNN-based pre-training and prompt tuning techniques.
• Features research on LLM applications and evaluations for graph-related tasks.
• Lists benchmarks and datasets specifically designed for GFM evaluation.

Maintenance & Community

• Maintained by members of BUPT GAMMA Lab.
• Open to contributions from the community.
• A tutorial on graph foundation models was held at WebConf 2024, with parts available online.

Licensing & Compatibility

• The repository itself is likely under a permissive license (e.g., MIT, Apache 2.0) given its nature as a curated list.
• Individual papers retain their original publication licenses.

Limitations & Caveats

• This is a literature compilation, not a software library; it does not provide runnable code or models directly.
• The field is rapidly evolving, so the list may require frequent updates to remain fully comprehensive.