Retrieval-Augmented Generation with Large Language Models in Radiology: From Theory to Practice
Anna Maria Fink, Alexander Rau, Marco Reisert, Fabian Bamberg, Maximilian Frederik Russe
Abstract
Retrieval-augmented generation (RAG) integrates task-specific expert knowledge into large language model (LLM) queries, offering potential to improve diagnostic accuracy, transparency, and decision support. This approach enables LLMs to function as complementary tools for radiologists, helping optimize patient care and meet growing health care demands.
Topics & Concepts
WorkflowComputer scienceWorkloadData scienceOperating systemDatabaseTopic ModelingArtificial Intelligence in Healthcare and EducationRadiomics and Machine Learning in Medical Imaging