1 paper
Kim-Celine Kahl, Selen Erkan, Jeremias Traub +4
Vision-Language Models (VLMs) have great potential in medical tasks, like Visual Question Answering (VQA), where they could act as interactive assistants for both patients and clin…