ws22 23_chemicals

Information Extraction from Scientific Literature and Patents

Motivation

• Information is published in scientific paper or patents
• The information is not machine-understandable (“hidden” in PDF documents)
• Data is not FAIR (Findability, Accessibility, Interoperability, and Reuse)

Task

• We aim to extract and structure the “hidden” information from scientific literature and patents
• We focus on figures (images) in patents and publications

Approach