This repository contains code and data used in the paper "MetaGL: Evaluation-Free Selection of Graph Learning Models via Meta-Learning" (ICLR 2023).
Running install.sh will set up the conda environment for MetaGL and install required packages.
You can run MetaGL by executing python main.py.
A comprehensive benchmark environment for evaluation-free selection of graph learning models is available in the GLEMOS repository, which provides a suite of model selection algorithms including MetaGL, evaluation testbeds, and meta-graph features, among others.
If you use code or data in this repository, please cite our paper.
@inproceedings{park2023metagl,
title={Meta{GL}: Evaluation-Free Selection of Graph Learning Models via Meta-Learning},
author={Namyong Park and Ryan A. Rossi and Nesreen Ahmed and Christos Faloutsos},
booktitle={The Eleventh International Conference on Learning Representations},
year={2023},
url={https://openreview.net/forum?id=C1ns08q9jZ}
}
