Assessing Impact Patterns in Research Texts Applying Corpus Driven Methods (Text Transfer)

A continuation of a previous pilot study, this project aims to detect transferrable knowledge across several domains. In the pilot study, the researchers demonstrated the feasibility of this idea by using inductive and deductive methods to identify relevant impact categories and patterns for the domain of electromobility. These methods included interviews, information extraction, natural language processing, and machine learning. With the current grant, they extend this work to the domains of artificial intelligence, music, and linguistics, among others.

“The main goal of this project is to advance, validate, and apply methods to study the impact of research on society at large across different dimensions of possible impact of science on people. We are also developing and validating new taxonomies and methods for this purpose," said Diesner. Read more