CliniChat addresses the critical gap in clinical interview training by enabling LLMs to simulate authentic doctor-patient conversations using a two-module framework.

• Clini-Recon module reconstructs realistic clinical dialogues by integrating multiple knowledge sources
• Clini-Eval module provides comprehensive evaluation metrics for assessing dialogue quality
• Tackles the fundamental problem of limited high-quality clinical dialogue datasets
• Creates opportunities for medical education, training, and potentially improving doctor-patient communication

This innovation has significant implications for the medical field, offering a scalable approach to train and evaluate AI systems for healthcare conversations without compromising patient privacy.

CliniChat: A Multi-Source Knowledge-Driven Framework for Clinical Interview Dialogue Reconstruction and Evaluation