This is a page that curates AI-related papers published worldwide. All content here is summarized using Google Gemini and operated on a non-profit basis. Copyright for each paper belongs to the authors and their institutions; please make sure to credit the source when sharing.
Performance and Practical Considerations of Large and Small Language Models in Clinical Decision Support in Rheumatology
Created by
Haebom
Author
Sabine Felde, R udiger Buchkremer, Gamal Chehab, Christian Thielscher, J org HW Distler, Matthias Schneider, Jutta G. Richter
Outline
This paper presents the potential of large-scale language models (LLMs) to support clinical decision making in complex medical fields such as rheumatology. Our results show that small-scale language models (SLMs) combined with search-augmented generation (RAG) outperform large-scale models in diagnostic and therapeutic performance, while consuming significantly less energy and enabling cost-effective local deployment. These features make them attractive for resource-constrained healthcare settings. However, no model has consistently achieved rheumatologist-level accuracy, so expert supervision is essential.
Takeaways, Limitations
•
Takeaways:
◦
A combination of small-scale language models (SLMs) and RAGs shows better diagnostic and therapeutic performance in rheumatology.
◦
SLMs are more energy efficient than LLMs and can be deployed locally for cost-effective, localized deployments, making them ideal for resource-limited healthcare environments.
•
Limitations:
◦
No model consistently achieves expert-level accuracy.