Modular pipeline system for text extraction from nanoscience publications
We propose an interactive machine learning system to extract textual facts from nanomaterial publications titled WISDOM. The system is based on a modular pipeline design, allowing future research teams to implement various algorithms and perform extraction on arbitrary facts.
Current systems only focus on extracting specific facts and do not support the annotation process. In contrast, we designed a modular architecture for textual fact extraction from journal articles. We deploy state-of-the-art architecture principles to enable generalizability. With this pipeline, we, for the first time, bring interactive machine learning into nanoinformatics. We are contributing to the current research in nanoinformatics by a) eliciting the status quo of existing systems and b) by proposing a modular, adaptable pipeline system which can be leveraged by researchers in the future.
WISDOM - Modular pipeline system for text extraction from nanoscience publications
Technical University of Munich
85748 Garching b. München