LinkingPark is an open-sourced system for automatic semantic table interpretation. Given a table of data, the system can match tabular elements to existing knowledge graphs (in this release, primarily Wikidata). Such semantic information from tabular data can then empower various downstream applications, e.g., intelligent spreadsheet programs, knowledge base population, data integration and cleaning, etc. LinkinkingPark is unsupervised, stand-alone, and also provides flexibility for multilingual support out of the box. This initial public release builds on top of the previous internal version that won 2nd prize in the SemTab 2020 competition and provides much improved efficiency and stability while maintaining or improving correctness over different datasets.

LinkingPark is available on GitHub: https://github.com/microsoft/vert-papers/tree/master/papers/LinkingPark. Contributions and collaboration are welcome.