#Publication

Presentation of Study on the Effectiveness of Large Langue Models at the EbM

Published on: 2025

S. Lengauer, F. Proprentner, M. Tytarenko, C. Krenn, K. Jeitler, T. Schreck: Automatische Beantwortung von Gesundheitsfragen durch generative KI mittels Retrieval Augmented Generation — Vergleich und Potentiale von Großen Sprachmodellen anhand des Beispiels Diabetes mellitus Typ 2. 26. Jahrestagung des Deutschen Netzwerks Evidenzbasierte Medizin, Freiburg, 2025.

Researchers from the A+CHIS research group attend the anniversary of the EbM-Netzwerk from March 26 to 28, 2025 in Freiburg, Germany. In the course of the conference, a poster on intermediate results of the applicability of Large Langue Models for answering diabetes-related questions using a Retrieval Augmented Generation (RAG) approach was presented by Stefan Lengauer. The contribution was well received and lead to vivid discussions with domain researchers from the field of medicine.

Diagram illustrating risk factors and complications associated with Type 2 Diabetes. On the left, risk factors include genetic predisposition, age, unhealthy diet, physical inactivity, high blood pressure, overweight/waist circumference, and smoking. On the right, complications include diabetic eye disease, kidney disease, nerve damage, foot syndrome, coronary heart disease, arterial occlusive disease, strokes, and depression. 'Diabetes Typ 2' is highlighted in the center as the connection between the factors.