Invited Talk: Knowledge Discovery in Large Datasets using LLMs
I was invited to give a talk on the topic “Knowledge Discovery in Large Datasets using LLMs” at the 2nd Conference on Impact and Social Work. In the presentation, I made the case that …
This project brings computational methods into social-work research and explores how large literature collections can be made accessible without obscuring the limitations and ethical questions of automated analysis.
SWORM §
The Social Work Research Map structures a collection of almost 25,000 international journal articles using topic modelling, visualization, search, and recommendations. Its purpose is to provide a lower-threshold route into international social-work literature.
extra-mining §
This line of work applies text mining and network visualization to a large collection of German social-work literature reviews. It examines what computational analysis can reveal about disciplinary discourse—and where the available data limits those conclusions.
Language models for data analysis §
More recent work considers how language models can make exploratory data analysis accessible to people with limited programming experience, as well as the new risks created when users cannot independently verify generated analyses.
