JGNN

Resource efficient machine learning and graph neural networks in native Java.

Graph Neural Networks (GNNs) are getting more and more popular, for example to make predictions based on relational information, or to perform inference on small datasets. JGNN provides native Java implementations of this machine learning paradigm, and does not require dedicated hardware or firmware. Follow the Jitpack badge for Gradle or Maven integration.

• Cross-platform
• Lightweight
• Optimized: data views, automatic datatypes, SIMD, parallelized batching
• Neuralang scripting language for model definitions

Feel free to contribute in any way, for example through the issue tracker. In addition to bug reports, requests for features and clarifications are welcome.

:rocket: Guidebook

:dart: Javadoc

:computer: Tutorials

:notebook: Citation

@article{krasanakis2023101459,
	title = {JGNN: Graph Neural Networks on native Java},
	journal = {SoftwareX},
	volume = {23},
	pages = {101459},
	year = {2023},
	issn = {2352-7110},
	doi = {https://doi.org/10.1016/j.softx.2023.101459},
	url = {https://www.sciencedirect.com/science/article/pii/S2352711023001553},
	author = {Emmanouil Krasanakis and Symeon Papadopoulos and Ioannis Kompatsiaris}
}

Apache license 2.0. Copyright © 2024, Emmanouil Krasanakis (maniospas@hotmail.com).