Fig. 2: Reasoning RAG agentic workflow. | Communications Medicine

Fig. 2: Reasoning RAG agentic workflow.

From: Simulated patient systems powered by large language model-based AI agents offer potential for transforming medical education

Fig. 2: Reasoning RAG agentic workflow.

AIPatient system’s processing backbone comprises three key stages: retrieval, reasoning, and generation. It first retrieves relevant information from the knowledge graph (Retrieval Agent and KG Query Generation Agent), then applies contextual reasoning to reduce hallucinations (Abstraction Agent and Checker Agent), and finally generates natural language responses based on conversation continuity and tailored to the perceived patient personality (Rewrite Agent and Summarization Agent). Created by the authors using built-in Microsoft PowerPoint icons and shapes.

Back to article page