Kairntech participated with presentation about automatic relation extraction for Knowledge Graph construction
For the 20th anniversary of the B-IT at University of Bonn, Prof Martin Hofmann-Apitius and Prof Holger Fröhlich had invited alumnis of the programs Life Science Informatics and Computational Life Sciences as well as participants from industry and politics for a two event on the state and perspectives of the use of AI in the Life Sciences. Kairntech had the honor to be part of the event and present work we’ve been doing with Fraunhofer SCAI on the creation of Knowledge Graphs with automatic relation extraction. (more information on the event)
Relation extraction is a recent addition to the portfolio of components in the Kairntech software: deep learning models extracting relations can be integrated in larger NLP pipelines that in this case allow to analyse large amounts of scientific publications and output the resulting network of entities (proteins, genes, drugable substances and diseases) as well as their relations. Results are rendered in the BEL format and allow large scale creation of cause-effect networks for therapeutic areas such. Joint publication of this work with the teams from SCAI and Kairntech include this and this.
The setup is running on the HPC cluster at SCAI and currently serves to cover analyses and more and more therapeutic areas.