4TU FAIR Data Fund 2021

Artur Schweidtmann has won a FAIR Data Fund 2021 granted by the the 4TU.ResearchData Community. His proposal is about FAIR scientific publications. The project obtained 73 out of 81 points from three reviewers.

The goal of this project is to make figures from peer-reviewed open-access scientific publications FAIR. We focus on an image dataset that we mined and annotated from chemical engineering and chemistry publications. The dataset includes metadata about the content and type of the figures. Furthermore, we aim to link the images to our new domain-specific knowledge graph “ChemEngKG” (currently under development) as well as the Open Research Knowledge Graph. This project will enable the reuse of scientific figures. As this image data can be processed by machine learning algorithms, we envision that the classification algorithm will be trained in an active learning setup for the classification of scientific figures. These algorithms can later be integrated into the publishing process enabling FAIR data generation at a publishing stage. Moreover, we envision that the FAIR database will be the basis for future works on automated information extraction from figures. Last but not least, the searchable figures can be finable and reusable by humans for various applications (Wikipedia, education, research, etc. ).