Jana Diesner, assistant professor, has received a start-up allocation award from the Extreme Science and Engineering Discovery Environment (XSEDE). This award provides Diesner and her co-PI Brent Fegley, a doctoral student in the Informatics program, with time on XSEDE’s high-performance computing resources and help from XSEDE staff members with parallelizing their code. The overall goal with this project is to build prediction models for an entity extractor that covers and goes beyond the set of entity classes that are typically considered by entity extractors such that the resulting technology becomes particularly useful for applications in the social sciences and humanities. Once these entities have been extracted from (large-scale) corpora, they can be linked into network data based on additional characteristics of the texts. The team uses natural language processing techniques and supervised machine learning for their work.
Diesner and Fegley will make the resulting prediction models publicly available in form of an end-user technology; providing a tool that can help others, especially scholars from the social sciences and humanities, to collect and construct network data that allow for asking meaningful and substantive questions about socio-technical sytems. The outcome of this work is a crucial component for constructing better models and theories about the relationship between information and social structure.