WiseMind: a knowledge-guided multi-agent framework for accurate and empathetic psychiatric diagnosis

2025-02-28 · Source: cs.AI updates on arXiv.org · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Data Science & Analytics · Depth: Expert, quick

Summary

WiseMind is a novel multi-agent framework designed to enhance psychiatric assessment by integrating both diagnostic accuracy and empathetic communication, inspired by Dialectical Behavior Therapy. It features a "Reasonable Mind" Agent for evidence-based logic and an "Emotional Mind" Agent for empathetic interaction. The framework utilizes a DSM-5-guided Structured Knowledge Graph to direct diagnostic inquiries, which significantly reduces hallucinations compared to standard prompting methods. Evaluated across three common psychiatric conditions using virtual standard patients, simulated interactions (1206 conversations), and real human interaction datasets (180 sessions), WiseMind achieved an 85.6% top-1 diagnostic accuracy. This performance approaches that of board-certified psychiatrists and surpasses knowledge-enhanced single-agent baselines by 15-54 percentage points, demonstrating its capability for reliable and supportive AI-driven psychiatric assessments under human oversight.

Key takeaway

For AI Scientists developing healthcare applications, WiseMind demonstrates that combining specialized agents with structured knowledge can significantly improve diagnostic accuracy and patient interaction. You should consider multi-agent architectures and knowledge graphs to mitigate LLM limitations in sensitive domains, ensuring both clinical soundness and psychological support in your AI systems, always with appropriate human oversight.

Key insights

WiseMind is a multi-agent LLM framework for psychiatric diagnosis, balancing accuracy with empathetic communication.

Principles

• Integrate logic and empathy for holistic care.
• Knowledge graphs reduce LLM hallucinations.
• Multi-agent systems enhance diagnostic reliability.

Method

WiseMind employs a "Reasonable Mind" agent for logic and an "Emotional Mind" agent for empathy, guided by a DSM-5 Structured Knowledge Graph to steer diagnostic inquiries.

In practice

• Use multi-agent LLMs for complex diagnostic tasks.
• Incorporate structured knowledge graphs to improve accuracy.
• Design agents for specific communication roles.

Topics

• WiseMind Framework
• Psychiatric Diagnosis
• Multi-Agent Systems
• Empathetic AI
• Knowledge Graphs

Best for: AI Scientist, Research Scientist, Domain Expert

Related on AIssential

• Mental-R1: Aligning LLM Reasoning for Mental Health Assessment — Aligning LLM reasoning with human cognitive processes via specialized reinforcement learning improves mental health assessment accuracy.
• ArogyaSutra: A Multi-Agent Framework for Multimodal Medical Reasoning in Indic Languages — Multimodal medical reasoning in Indic languages is enhanced by an actor-critic framework with tool grounding and dual-memory.
• MIND: AI Co-Scientist for Material Research — MIND is an LLM-driven multi-agent framework for automated, experimentally verified hypothesis validation in materials science.
• MedPanel AI - Multi-Agent Clinical AI — Multi-agent systems, particularly with adversarial agents, enhance diagnostic safety by simulating human clinical reasoning and challenging initial conclusions.
• SAGE: A Strategy-Aware Graph-Enhanced Generation Framework For Online Counseling — SAGE integrates clinical knowledge graphs with LLMs to enhance strategy-aware, theory-driven online mental health counseling.
• Can AI Chatbots Reason Like Doctors? — Large language models demonstrate potential in clinical reasoning, but require careful integration and standardized evaluation in medical workflows.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.AI updates on arXiv.org.