17.05.2025 – 09:30-10:15 (Africa/Abidjan), Fab8.B 206
Sprache: English
Although there are several language models accessible via chat interfaces (e.g., ChatGPT) they have to be used carefully because of data privacy, untransparent data filtering, bias, and other issues. In this session, you will have the opportunity to test a locally deployed prototype of a conversational agent called BeeChat, deployed at Bern University of Applied Sciences. The tool is based on open-source software components and deployed on customer-grade architecture. We will briefly explain the system architecture and design decisions, point out the limitations and – following your hands-on experience - collaboratively discuss open challenges and opportunities.
- open source software applications
IT managers, Researchers, Tertiary education (universities, adult education, ...)
Daniel Reichenpfader is a research associate at Bern University of Applied Sciences and PhD candidate at the university of Geneva. His research is focused on the development and application of natural language processing-based methods in medicine, as well as for research and higher education.
