MIND: Unified Inquiry and Diagnosis RL with Criteria Grounded Clinical Supports for Psychiatric Consultation

Researchers propose MIND, a reinforcement learning framework that improves AI-powered psychiatric consultation by addressing key challenges in diagnostic accuracy and clinical reasoning. The system uses a Criteria-Grounded Psychiatric Reasoning Bank to provide better clinical support and reduce inquiry drift during multi-turn patient interactions.

• →MIND framework addresses fundamental challenges in AI psychiatric consultation including unsupported clinical assertions and inquiry drift.
• →The system incorporates a Criteria-Grounded Psychiatric Reasoning Bank that retrieves similar reference consultations for better clinical support.
• →Reinforcement learning approach uses rubric-based process rewards to improve diagnostic decision-making across conversation turns.
• →Experiments show MIND outperforms existing baselines in diagnostic accuracy, empathetic interaction quality, and interpretability.
• →The research advances AI applications in healthcare by tackling complex psychiatric diagnosis challenges that require nuanced reasoning.