Skip to content

official repository for CollabKG [Accepted by COLING2024]

License

Notifications You must be signed in to change notification settings

cocacola-lab/CollabKG

Repository files navigation

CollabKG: A Learnable Human-Machine-Cooperative Information Extraction Toolkit for (Event) Knowledge Graph Construction

  • CollabKG is an open-source IE annotation toolkit that unifies NER, RE, and EE tasks, integrates KG and EKG, and supports both English and Chinese languages.

  • CollabKG combines automatic and manual labeling to build a learnable human-machine cooperative system. In particular, humans benefit from machines and meanwhile, manual labeling provides a reference for machines to update during annotation.

  • Additionally, CollabKG is designed with many other appealing features (customization, training-free, propagation, etc) that enhance productivity, power, and user-friendliness. We holistically compare our toolkit with other existing tools on these features.

  • CollabKG Extensive human studies suggest that CollabKG can significantly improve the effectiveness and efficiency of manual annotation, as well as reduce variance.

    🖥 Try out CollabKG online

    🖹 CollabKG paper

    🎥 CollabKG systems demonstration video

    📌 Overview of how to use CollabKG

    📌 Frequently Asked Questions (FAQ)

    📨 Feel free to reach out if you have any questions by emailing 22120436@bjtu.edu.cn

Getting started

CollabKG can be built using Docker. Before doing so please add a secure token to the TOKEN_SECRET field in /server/.env for user password hashing and salting. After this, in the repository root directory, execute:

$ make run

or alternatively:

$ docker-compose -f docker-compose.yml up

Issues, Bugs and Feedback

If you come across any issues, bugs or have any general feedback please feel free to reach out (email: 22120436@bjtu.edu.cn). Alternatively, feel free to raise an issue, or better yet, make a pull request 🙂.

Known Issues/Bugs

Future features

Acknowledges

Thanks to the QuickGraph team for their support.

Attribution

Please cite our [paper] if you find it useful in your research:


About

official repository for CollabKG [Accepted by COLING2024]

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published